Bitcoin Forum

I'm opening a thread to discuss my work on the the NanoFury project and general design questions regarding the Bitfury ASIC.

The project home is at: https://github.com/nanofury/NanoFury

Project Summary:
Interface: USB 2.0
Power Requirements: 2.5W when being operated within USB2.0 specs, up to 3.5W when using a powered hub
Performance: Approx 2GH/s when within the USB2.0 specs, up to about 2.5GH with passive cooling
Dimensions: 25x42.5mm
ASIC: 1x Bitfury chip

PCB/CAD software: KiCAD

This is how everything looked while designing it :
https://i.imgur.com/KtyQzFql.png (http://imgur.com/KtyQzFq)

And how the real ones ended up:
https://i.imgur.com/WuQV0Qjl.jpg (http://imgur.com/WuQV0Qj)

There are more images and galleries at: http://nanofury.imgur.com/

Project Status

Hardware:
As of version 0.6 of the PCB everything seems to work and be pretty stable. If you're looking to try out the design - start from that version.

Software:
Even though we started with cgminer that one needs a lot more work and has not been checked in into github yet.

Luke-Jr however did wonders and his bfgminer works beautifully! Luke - thanks one more time!!!

As part of the bfgminer's requirements each device needs to report a specific "Product String". I've put up a small test program that can exercise each device and also fix that string. Alternatively, you can tell bfgminer the individual serial numbers and in that case it won't care about that string. (although I think it's much easier to just set the product string :)). You can find the test program at https://github.com/nanofury/NanoFury_Init

Overclocking
I guess everyone's favorite topic :)

The board does support some overclocking as well as voltage tweaks (e.g. "pencil mod").

In order to meet the USB2.0 requirements bfgminer defaults the bitfury chip to speed of 50 bits which results in about 2.5W and around 2GH or just below.

Most computers would actually tolerate slightly over-spec devices and won't shut the port down. For such cases you can experiment by overriding the number of speed bits and find out where the limits of your hardware are.

Also, keep in mind that no two bitfury chips are the same - each board may perform slightly differently.

And I guess it comes without saying that some cooling will be required. Start with at least a small heatsink to the back of the board.
If you want to go a bit more aggressive - add a heatsink to the bitfury chip and the inductor.
And you can't go wrong by adding some active cooling (like a small fan).

WARNING: THOSE BOARDS WILL RUN HOT! AND I MEAN VERY HOT - 80-85C! HANDLE THEM WITH CARE!!!

If your boards overheat you may see the hashrate dropping significantly (like 0.3-0.6GH instead of 2.4). Sometimes the USB port will start dropping the voltage and instead of 5V it may go down to 4V or even 3.5V. Once it goes below 3.6V this may interfere with the USB communication, so you'll likely lose communication with that board. In such case you will have to unplug the board and plug it back again.


Future plans:

- Converting the design to a two-chip one would be a very easy task. It just needs more power and an extra bitfury chip. On the other hand, power consumption would most likely exceed USB3 specs ... that's unless you have a powered USB hub providing 1-1.2A per port. It will also produce a lot of heat. So a heatsink will be a must.

- Use the design for a zero-chip board to serve as an adapter to be used with the existing 8-chip and 16-chip boards (such as punin's and buzzdave's ) for those that would prefer to run the mining software on a PC (and those that for whatever reason don't have or don't want to use an M-board).
It would be a USB-to-HCard adapter.
There are several designs of H-card clones that use 8 chips (including buzzdave and tytus's new 8-chip boards) so that may be a good fit for pre-starter kits.

bfgminer already has the necessary support and will discover the more than one-chip-per-port situation.

Some history:

I got initially the idea of the NanoFury as I didn't like having to buy a Raspberry Pi in order to use Bitfury's chips. I wanted something that I can just plug in my laptop. I even joined some of the Avalon and BFL preorders (what a disaster after a disaster!). Then I discovered bitfury's thread ([ANN] Bitfury is looking for alpha-testers of first chips! FREE MONEY HERE! (https://bitcointalk.org/index.php?topic=228677.0)) and things started getting interesting ... but no chips though...

And what happened next is all tytus' fault! I got a message that he'll be sending me few sample chips! And a few days later I got them! And I promptly lost sleep :)

Unfortunately to this day there is no documentation for bifury's chip. I started digging trough the forums and collecting various bits from here and putting the pieces together. I've put all of those notes in a wiki article - which uses the "docuwiki" format, and github wants the "mediawiki" format, so on my to-do list is to start reworking that. If there are any volunteers to help with that or know how to easily convert it - please PM me.
Thanks to z3phyreo who is porting that documentation we already have some of it available here: https://github.com/nanofury/NanoFury/wiki/The-missing-bitfury-chip-documentation .

My very first idea was to make it an Arduino shield. But that would've been an overkill for most people.

So, for the pre- v0 of the board I picked Microchip's MCP2210 chip - it seemed like a good candidate. Except that after wasting a week combing through the documentation I could not find a way to bitbang the SPI pins, which is a must for bitfury's RESET sequence. Which lead me to abandoning that chip.

So, version 0 of the board was made around the FT232R chip. I picked that one specifically as it claimed that it has a built-in 3.3V regulator, and that it operates at down to 1.8V - just perfect for interfacing with bitfury's chip which also works at 1.8V. It is also a very popular chip - especially among hobbyists. I probably have a dozen boards with it around - various rs232/serial and custom adapters.
The FT232R chip however is really an asynchronous interface, so the SPI required quite a lot of work. And worst of all - it is very slow when doing bitbang kind of operations - I couldn't reliably squeeze out more than 1-2kHz out of it. I had almost given up when I noticed that it can work as a pseudo-parallel interface, in which case it would use the /RD and /WR buses and do a send, then receive sequentially. That way I could make the individual pins of the SPI interface be simulated via the 8 data bus pins. I got 50-100kHz that way, and it also took care of the RESET sequence. The CPU however would have to serialize and de-serialize all the communication.
This was a bit of a hack, and a rather ugly solution, and I had spotted another (better!) chip which seemed like a more appropriate candidate!

I started working on v0.1 of the board using Cypress' CY7C65211 chip. Luckily I wasn't too far into that when both Cypress and Digikey (and Avnet a bit later) said that those are on back order...

And while discussing the various solutions and options one of them was:

So that became v0.1 of the board, back with MCP2210. At that time it was unclear how much power does the bitfury chip draw from the 1.8V, so I picked a rather bulky LDO. Meanwhile it got clarified that it is a lot less.

Which became v0.2 of the board. But I had forgotten about those 0402 components - they are tiny!!!! And it was just too much pain to deal with them, and on top of everything else there was plenty of room on the board, so not really necessary to go that small.

In v0.3 of the board I changed everything to 0805 - nice and big! :) Actually so big that for some components it was getting too cramped ...

And v0.4 was born with most components back to 0603. While I was waiting for the v0.4 to arrive I noticed that I had overlooked the MCP2210 specs - it didn't like 1.8V signals ... so I had to add a voltage converter. Some of the ideas that were floated used a cheap bipolar transistor, but I wasn't quite sure how the extra currents floating through it would affect the bitfury chip, so I picked the MOSFET option. I had a bunch of them laying around, so I just grabbed the first one and started testing ... and it did not work. I had almost gone back to a plan NPN one when I figured out what was wrong with mine - 1.8V wasn't enough for it to switch - it needed over 2.5V. I dug out one with a gate-source threshold of 1.4V and - bingo!

Which lead to v0.5 of the board. I also decided to ditch the printed USB-a connector for a number of reasons - PCB had to be thicker (read: more expensive), then that connector needed to be gold-plated (read: more money), then the boards needed some special cuts for that connector (read: even more money) also my sample boards seemed a bit flaky in the USB slots and contact wasn't always as good, etc.
So I put a proper USB-A connector which was also a relatively cheap one. Except that I messed up the pin numbers - they were in reverse order.

Which resulted in PCB version 0.6.

Software in the mean time was quite a pain too. Microchip only had a .NET version of that DLL .. so any linux-support hopes were a bit foggy. Luckily I found Kerry Wong's library. Which was meant mostly for linux usage. (Did I mention that I'm more of a Windows fan? :))

Eventually, after many many hours of fiddling with it (thanks Val!) we finally got something working on Windows and I started debugging. As it turned out that there was a bug in the library, which I fixed in the nanofury fork (https://github.com/nanofury/MCP2210-Library).

And as they say - the rest is history :)



Acknowledgments and other information:

* This project was inspired to a great extent by BKKCoins' Klondike K1 Nano Open Source Board (https://bitcointalk.org/index.php?topic=190731) and I'm very grateful for all his work and I'll continue to support that project down the road as I can!

* I've reused the same physical dimensions to as much as possible (including heatsink holes size and position) with the goal that if necessary both projects may share the same plastic hosing and/or heatsink.

* Several people contributed to the development of this project and I'd like to say Thank You one more time for the dedicated work! You can find more details on the github page.


Usage
I guess for most people that's the most important question :)

First - start by downloading KiCad: http://www.kicad-pcb.org

Get the project files, and open the project from within KiCAD.

When you open either the schematic or PCB documents KiCAD may want to start with its own libraries. Remove them from the list - the NanoFury project includes a lib folder with all libraries that you may need. Just add all files from the NanoFury/NF1/kicad/lib folder. (this also would prevent the issue when KiCAD wants to use its own library for a given component instead of the one from the project which may result in bad schematic or PCB)

From within KiCAD then you can print the schematic documents and generate all files needed by the PCB and factories.

And that's all - it is really that simple! :)

I guess many of you will have questions, so feel free to post on this thread and I'll try to get back to you as soon as possible.


Donations

If you found this project helpful and wanted to donate some bitcoins please send them to 1AHvViTD5wohPywNYHi8NZ8uPhbY9dZXCY. All donations are much appreciated! :)

If you end up using the NanoFury designs for manufacturing of any boards for sale/resale or otherwise for profit -  I'd like to ask that you consider adding a few bucks "design donation" to the first few hundred boards (or as much or many as you feel appropriate) to help recoup some of the several thousand dollars that went into the development of this project (not mentioning the countless man-hours of work).


Finally:

And most importantly - HAVE FUN! And Happy Mining! :)

reserved

Nice work.

How much would you like us to kick back to you if we fab out your NF's... or should I PM you?

Well, you probably won't be the first or only one asking that question, so I guess it is appropriate to be here.


The answer is - As much as you want or feel appropriate :)

Ideally a few bucks on the first several hundred boards would be fine, and in the bitcoin reality that may shrink even further with time.

After all I didn't want to bog any manufacturer with fees - that's why I left it open source. If the designs worked for you - then feel free to cut a bit or as much as you like for us :)

And of course - feel free to PM me and we can chat a bit more.

Let the god bless you with the kids....

Will do thanks for the information.

If I understand correctly inside Bitfury chips is comparator. CMQ is the output. Unused opampas should be connected differently. Minus connected with output and plus to the ground. Shorting output to ground may cause unnecessary current flow.

Great design with great potential.

Dexter - bitfury posted some stuff here:
https://bitcointalk.org/index.php?topic=183368.msg2460608#msg2460608

but I haven't had time to translate it (and google translate is a bit useless with technical stuff).

A few posts later however there is a sample schematic that sort of explains the story. Basically the CMQ/CMMINUS/CMPLUS pins are for chips when designed to be in series - like Christmas lights. I remember reading somewhere a reply from bitfury that this stuff is probably not going to be reliable and that's why he recommends those pins to be connected to GND.

I guess tytus & buzzdave decided to give it a shot anyways - and the results are more than obvious : :)

Great work. Thank you so much.
I've contacted with vs3 before. He's a very nice guy. :D

Thanks for the link. I was learning russian in school 25 years ago ;) I can read but don't understand to much ;)
I think that simply connecting chips in series is a big mistake, pictures above. I've made experiment with 3 uC's connected in series with some additional electronics and it worked beautifully. Unfortunately I don't have Bitfury chips to verify this concept with them. I have concept to supply 40 chips with one DC/DC converter 3,5V 20A.

+1 on everything. Congrats vs3 on the successful project! Open-sourcing this is so good for the community.

Hope many good tings will come for this project in the future, and many production batches.

Amazing work! I would love to be able to get my hands on one of these. Looks like a lot of energy and care went into the design, am excited to see how it all pans out!

Just wanted to throw in my encouragement!

Thank you guys for the kind words! It took quite some work indeed! Hopefully this will help others and save them from having to go through the same hassles I went trough.

And to be honest - I think bitfury should've released something like that to begin with.

Very nice, simple and effective, one question though could more chips be chained directly from the test points you have on the board? I have been working on a simple raspi to 4 chip but after I saw you USB interface I like that better but looking at the software comments it looks to be a 1:1 setup w/o more software work (or not I hope). At any rate pretty awesome and thanks for sharing !

Yup - that's exactly what those test points were meant for :)
Presuming you've already figured how to provide 0.8V to the other chips all that remains is to hook them on the chain - via those test points.
You'll basically need hook 5 wires: SCK, MOSI, MISO, 1.8V and GND.

Also, among my chats with Luke-JR was the question to add support to bfgminer for "chaining" of such chips - and he's done that too (as this is being used by almost all bitfury designs anyways).
So from that point on all you'll need is just some hardware - no additional software development is necessary. (well... hopefully ... you never know :))

As I mentioned early - my goal was to do the ultimate K.I.S.S. design for those chips :)

Thanks I figured that was pretty much a given anymore on chaining chips and such, good job and thanks again for the hard work!

I am new to kicad and somewhat lost.  I live next to a pcb fab and assembly facility. I am tryimg to get the Gerber file and bom parts list so they can get me a quote for putting a proto together.  Also if this is financially viable what would be a good donation?

Stan - on the several questions:

- KiCAD and gerbers: Start by downloading and installing KiCAD. Then open the pcb file - it will probably complain about libraries and stuff but you can ignore that. Then from the menu select Plot - that's how you generate the gerber files. On that window there is also a button for the drill file.
- BOM : Open the schematics and from the menu -> tools -> generate bill of materials. I usually pick a tab-delimited text file which then Excel can easily read.
If you get stuck with those feel free to PM me and I can get those for you.

As for the donation part - a few bucks per board on the first several hundred or as much as you feel appropriate.

Vs3 thank you. ----- I do have a quote from the facility.  We are going to do a test run of 20.  I will be ordering chips when I get home from the job. 

Thank u for your open source project

Will be starting making these ones soon (if i got all things i need) and hope to get a working project

:-) Keep up the good work and will support when its all is done




When you open either the schematic or PCB documents KiCAD may want to start with its own libraries. Remove them from the list - the NanoFury project includes a lib folder with all libraries that you may need. Just add all files from the NanoFury/NF1/kicad/lib folder. (this also would prevent the issue when KiCAD wants to use its own library for a given component instead of the one from the project which may result in bad schematic or PCB)

I tried this method but when i want to open the project  NanoFury/NF1/kicad/lib folder its seems empty

I can open NF1-MCP2210 and i get a layout but i do not know for sure if thats the complete one to sent to the fab or make ourself

hope you can help me out with this

Nice!

p.s. - btw they could probably hand-assemble 20 units in a few hours ;)

The PCB and SCH files will be in NanoFury/NF1/kicad/NF1-MCP-V06/
When you open either of them then you may get some messages about libraries - then you go to Preferences->Libraries and use the ones in NanoFury/NF1/kicad/lib

Aside from that the PCB file will contain everything needed for exporting of the gerber and other manufacturing files.

I will have a long wait at this facility.  I am looking for quotes for others that can deliver sooner.  If I was able to place a larger order they would start sooner but this is not the case.  I made pre orders with Hasfast etc and have a limited amount of funds .............. it is what it is. 

Sweet :-) you are a great help !! :-)

when I open CvPCB to open the netlist I get this error message.




If I open the netlist directly in Pcbnew the components do not appear.

I'm not sure what's the story about "NanoFury.mod" file - it is certainly there:
https://github.com/nanofury/NanoFury/blob/master/NF1/kicad/lib/NanoFury.mod

I noticed that KiCAD tends to embed full paths - which is most likely why you're getting those errors (as yours won't be where mine were). When you open either the schema or pcb go to libraries and remove all references (KiCAD tends to insert its own libraries there too) and add just the files from the lib/ folder.

Which version of KiCAD are you using? Mine is probably old by now - 2013-07-07 BZR 4022 stable on Win32.

It works for me i can see all components but i get the same error some files could not be found nanofury.mod but i still do see them there are 48 parts wright ?

is there a easy way to look them up and buy them online ?

Sorry, I was using a version of the Ubuntu repositories 2011   ::)

Probably not everyone's idea of a Saturday evening drink ... but some steady hands and a glass of single malt (for the steady hands part! :)) and a few hours of boredom killed:

Let's start with some SMT paste:
https://i.imgur.com/SdlO3K5l.jpg (http://imgur.com/SdlO3K5)

Paste looking good:
https://i.imgur.com/eOouIatl.jpg (http://imgur.com/eOouIat)

a few hours later ...

Finally something ready to be baked :
https://i.imgur.com/Iwcd0pSl.jpg (http://imgur.com/Iwcd0pS)

NanoFury Cake anyone?
https://i.imgur.com/S3QlIbXl.jpg (http://imgur.com/S3QlIbX)

count down to liftoff ...
https://i.imgur.com/NMyojyZl.jpg (http://imgur.com/NMyojyZ)

Houston, we have ignition! :) Middle 2 boards are a few degrees ahead of the pack!
https://i.imgur.com/1fnHFMel.jpg (http://imgur.com/1fnHFMe)

Melt solder! Melt!
https://i.imgur.com/XavGnWel.jpg (http://imgur.com/XavGnWe)

Finally cooled down and ready for a close inspection:
https://i.imgur.com/upclCINl.jpg (http://imgur.com/upclCIN)

And we got some really nicely baked NanoFury miners :)

Full album is here: http://imgur.com/a/5hXU4

Hmmm, cookies!

Edit:
How do you pick-and-place them?

Steady hands! :)

btw the trickiest part was placing the bitfury chip ... on one of the boards it is off by half a pin length. Luckily it was shifted along the power pins, so all data ones were good.

Damn, I always wanted the excuse to make one of these:
http://www.youtube.com/watch?v=bkYNBFXlJYQ (http://www.youtube.com/watch?v=bkYNBFXlJYQ)
http://vpapanik.blogspot.gr/2012/11/low-budget-manual-pick-place.html (http://vpapanik.blogspot.gr/2012/11/low-budget-manual-pick-place.html)

I guess I'll have to get few full kits with ASICs just for the fun of assembling them.

Yeah, I've looked at something like that .. So far I haven't found anything that beats steady hands and good tweezers though.
I mean - down to 0603 everything is big enough for handling it that way. 0402 is probably as tiny as I'd ever do by hand (and given an option - read that as "never").

Want to make 110?

Sure, although I think you may not like my pricing :)

Like.
 :)

I will be looking at these designs later today but am interested in adding in more bitfury chips.  I'm not an electrical engineer but am pretty good with pcb design so was wondering what changes would need to be made to the schematics themselves to make this happen.  

Also, I have a local pcb builder that has a 5 day turnaround and a local pcb assembly company that takes on small-medium production runs.  I know I do not have much time on this forum but I've been a member for a while.

I'm located in the states.

Have a guy a idea to run this devices under the raspberry?
I bring them not to work:-(

vs3, can you tell us about the heat of the chips, without any dissipation layer, how much temperature rises?
I happened to make a big hole in the GND pad and once the bitfury soldier is padded on the backside with tin or thermal compound to insert a rear heatsink.

I had the sample boards that I made over the weekend run for a while without any heatsink on the back - just using the entire copper backside as a heatsink. They worked just fine at 50 bits speed (and I also let them run for a while at 51 and they worked fine too but I didn't let them run for long enough to be certain).

Also - keep in mind that the chips are not the only heat source. Chips produce about 80-85% of all the heat, but also the coil and regulator produce another 15-20% too. So heatsink on the back side helps them as well.

What we know for sure is that the bitfury chips can take a lot of beating - and I mean A LOT :

Well, I finished a modification of nanofury project to be printed on a single layer PCB, also brings built-in USB connector and holes to install standard chipset heatsinks.

https://bitcointalk.org/index.php?topic=343366.msg3678665#msg3678665

Anyone know the best place to buy the bitfury chips in bulk?

BuzzDave at Megabigpower.com. Got a small order in 2 days. I think he has like 1500 on hand.

I was hoping for a better price with an order of 50 or so.  I pm'd me but he didn't really give any specifics.

No you have to be in the stratosphere before there is a discount.  Like 3000 or more.

Seems to be sold out. Any other source?

I am willing to pay a decent fair market price for any.  Please PM me if you need some extra cash or coin.  Ideally I would like to get as many as possible.



------Dave is out.  I believe I took his last one. Waiting for FedEx now. 

I'm waiting for FedEx as well. :)

He said in an e-mail that he's restocking in February.  Preorders available at $5/chip, but in 3000-chip batches.  Unknown as of yet if we can order individuals or small batches from the website again when he finally restocks.  Hope so...

I wonder why so long between runs?

manufacturing of the chips is like waiting in line at a gov't office.....long wait

Is it possible to do a USB miner with a BFL chip?  I am guessing no since no one has done it and they are much more readily available than BF.

Might be possible, but with the BGA packaging they used makes them a lot harder to work with than the more hand-solder friendly packages like TQFP (with the leads sticking out the sides).  Plus, BFL lists the chips as out of stock, so I imagine you'd have to buy them from someone who has them stocked up just like BitFury.

The BFL chips uses 3.2 W/GH/s so that's 12.8 W/chip for the 4GH/s chip. That doesn't fall within current USB specs but may be possible in the future.

yup - that pretty much sums it up.

As people already mentioned it - there are 2 major issues when considering it for a USB miner:
#1 - package is too big and cooling requirements are just too much of a pain for a USB miner (BF produces 2-3W of heat, BFL is well over 10W)
#2 - most importantly BFL chip consumes MUCH MUCH MORE power than the 2.5W that USB2 can provide.

Although it is probably possible to make an externally-powered big device that has fans for dealing with the extra cooling, etc, etc - that would sort of defeat the purpose of the simplicity that you get with a USB miner.

BitMain?

May be an interesting option. Could conceivably fit 3 bitmain chips on a stick

The external power is already available(plenty of 12v & 5v on one of these)
https://www.dropbox.com/s/dpjtcouau6ah0kj/2013-11-02%2012.41.13.jpg
and connected:-
https://www.dropbox.com/s/x8ho5djfthq0a7u/2013-11-02%2010.26.41.jpg

I got a few of these, they run great!

Can you provide detailed instructions on overlocking?

I can't find no info online on the pencil mod for the nano fury. How do I got about doing this? Can you provide pictures with annotations?

I built a few ones for me. A friend of mine reorganised the layout. My fork: https://github.com/rgr-rgr/NanoFury. Pull request sent.

By the way: I noticed a new version v0.7, not described in the Release_Notes ... ?

Sorry I have a silly question. Does the microchip PIC need NOT firmware ?  I could not find it from github.

The device interacts directly with bfgminer so no it doesn't have any firmware.

HI,
here is our cgminer support

technobit.eu/0_1_3.rar (http://technobit.eu/0_1_3.rar)

0.1.3 Milestone release - Nanfury support is added with native libusb api. No hid api's are required!
* cgminer ./autegen.sh --enable-hexmineru to add Nanos' manufactured by TechnoBIT known as HEXu
* cgminer --hexmineru-frequency command line to set chip frequency - range 1-62
* cgminer Added to default /etc/config/cgminer  Factory default or web Save+Apply is required for changes to be applied!
* cgminer updated to 3.8.5 rev afe7710858e4ce28bb60f6ae6e167a18d687634f
* cgminer patch to cgminer 3.8.5 rev afe7710858e4ce28bb60f6ae6e167a18d687634f.patch - various code cleanup and optimization needs to be done
* hotplugd added hotplug support for HEXu in /udev/rules.d/01-cgminer.rules. Factory default is required for changes to be applied!
* openwrt updated to r39151

Marto - you guys rock! :)

Thanks for the awesome news!

I'd like to help with the documentation. Please let me know what I could do.

Bitly - please PM me your email address and I'll set you up with access to that wiki page.

I saw your pull - very good job! (I also have a few minor questions - I sent you a PM)

The v0.7 had a minor change mostly addressing shortages of the MCP2210 chip in I/SS footprint. In 0.7 both the I/SO and I/SS footprints are present, so you can use whichever chip you manage to find in stock.

I think I pushed all files already - if you find any missing pieces let me know though and I'll make sure to add them.

I've seen documentation for the bitfury chip, bi*fury (not sure why it's mentioned here though..), and the mcp2210 chip in nanofury.

Thanks. Good idea. Customers need not worry about firmware updating.

Ckolivas
DrHaribo
MineForeMan
Nwoolls

Have all been sent devices so they can test and incorporate support for the NanoFury design.

Luke - the guys that made the bi*fury design are the same that have been working very closely with bitfury and making his boards since the very beginning. I won't be surprised if they actually do have some documentation. However I did not see any such thing published while I was designing the NanoFury (and I may have missed it if that happened in the last 2 months since I already know more than enough and I've just stopped looking for such documentation). If you do have a public link for such documentation - please share it :)

As for the1 MCP2210 documentation - that's freely available from Microchip's web site and just in case I included a copy in the NanoFury github repository.

There is a test programm flying around called "nf1_init.exe". I want to run it on linux. Does anyone know the source for source?

+1

Is this what you're looking for:
https://github.com/nanofury/NanoFury_Init

Yes, thanks for the wake up :-)
I knew I have seen somewhere the source - but did not see it in front of me :-)

Well it is written Visual C for Windows - I would need a linux port. Has somebody done it?

Basically I just need to set the product string:

You can always use a windows VM and patch the rgrfury into the vm

Upon review of the source for nf1_init. It should only be a matter of changing some windows specific functions to Linux counterparts and compiling

I'm Working on a few ideas on increasing the hashing speed.

But I note on the IOREFF for Ver.7 that it is shown as 0V8(fine!!), but that it appears to be derived from the BUCK convertor (301F) rather than via a voltage divider & cap.
'Bitfury' seems to have stated that the IOREFF pin should be tied to 0V8, but if people start playing about with R2/R3 and the 'stupid' pencil mod, will this not impact the voltage that IOREFF sees potentially making it as high as 0v95?

Also I notice that VUSB does not appear to be decoupled correctly (0uf22/0uf47)?


Next question…
Prior to the nano 50 miners meant 50  embedded SOC's

Since all the function of the SOC has now been offloaded to the miner software, how does that impact the % processor power required to support multiple miners?
I.E can a PI or other SOC (A10/A20)still carry the utilization 'load'?

RF

they all use a 2 layer pcb correct? none use a 4 layer?

v0.7 is two layer. Can't state for fact that v0.6 was, but I'm pretty sure.

That's actually a good catch! Thanks!
I just updated the Release Notes to cover the v0.7 PCB.

There are several questions and I'll try to address them all -

First - huge thanks to z3phyreo for porting my collection of quotes and links regarding the bitfury chip (a.k.a. The Documentation :))! It is now on the project's wiki page:
https://github.com/nanofury/NanoFury/wiki/The-missing-bitfury-chip-documentation

---

Quoting my translated docs: https://github.com/nanofury/NanoFury/wiki/The-missing-bitfury-chip-documentation#performance-testing-and-results
(source (RU) (https://bitcointalk.org/index.php?topic=242745.msg3234190#msg3234190))

---

Well, bitfury's exact quote is: "IOREF - feed it with 0.9 V for standard signalling (better not take VDD but put resistive divider between GND and IOVDD) and some cap to remove pulsations." (I'm copy-pasting from my wiki: https://github.com/nanofury/NanoFury/wiki/The-missing-bitfury-chip-documentation#pinout-and-usage). In the first excel spreadsheet that he published he also said that "LOGIC 0 INPUT : INPUT < IOREF + 50 mV (+- 50%)" (there was actually a typo in the ">" sign) and vice versa.
As a result as long as the input voltage levels are 50mV above/below that IOREF reference voltage one you're good. Actually from what I've seen almost everyone uses the exact same trick - since VDD is about half of IOVDD anyways people just connect it straight there.
In my case - due to using resistor dividers for the SCK pin voltage on the SCK may drop to 1.25V. So - following the +50mV rule -  as long as your VDD is below 1.2V you should be fine.
Practically speaking - going over 1V will probably be pointless. From the above quote - at such high speeds the chip uses quite a lot of power - approx 6W which is over twice the USB2.0 specs. Also, the voltage regulator is rated for up to 3A and that will be your second limitation for going that fast. Your main limitation however is due to the way how those LDOs work - unless you use a very expensive multi-phase regulator you will always have some minor voltage fluctuations, and for this one it is normal to have 20-50mV (and no matter how many filtering capacitors you add - you can never get <1mV fluctuations).

So from an academic point - can you do over 3GH with those chips - yes. From a practical point - it's pointless. Achieving that gain of 0.5-0.7GH will cost you as much as several other miners, so it's cheaper to just put one more chip/miner.

---

Can you clarify? There are 4 decoupling capacitors - C1 and C3 (100nF) and C2 and CF1 (22uF) (source: schematic (https://github.com/nanofury/NanoFury/blob/master/NF1/kicad/NF1-MCP-V07/NF1-MCP2210-v07-SCH.pdf))

---

The amount of traffic per chip is very small - less than a kilobyte per second. That's certainly not a challenge for any desktop PC. I've had at some point 30+ USB miners running simultaneously on my PC and BFGMINER is using less than 1% CPU.

Raspberry PIs have already been show to work fine and are being used in Dave & Punin's standard mining solution and can easily handle 16 boards with 16 chips each (256 chips total).

Yup - all are 2 layer (from version 0.0 till 0.7 inclusive). Actually 0.5 and later have been optimized and don't have any tracks on he back side - which serves both as a GND plane and also a heatsink.
As a matter of heatsink - I was actually surprised how efficient that was - I ran a bunch of miners last night without any cooling whatsoever - no heatsink, no fan, etc - and they performed pretty nicely at 48 bits (1.7GH avg) at room temperature of around 28C.

great thread and props on not puttign fees to use it - just 'give me what u want'. doesnt happen in this world too often!

Hey I really love the Nanofury Design and thank you for putting it out there for us to use. I just got a small batch of Nanofury .7 "Icefury". It runs really great on one of my desktops, but on my laptop and other desktop it runs for about 10 minutes then fails. Any idea what has happened?

I am using BFGMiner 3.8.1 and have also tried 3.9.

If it works for a while and then starts misbehaving I'd say most likely it overheats. Get a small fan blowing at it and see if that will make any difference.
Also, when most electronics get hot they start using a bit more power - and at that point it might be getting beyond the limits of your laptop's power supply (but most desktops can tolerate higher power usage and that's why it works there). Try it also with a powered USB hub.
If you don't have an infrared thermometer handy you can start by lowering the speed by one or two bits and see how that changes the situation - at lower speed there will be less heat (and lower power usage).

There's also problem with maturing parts (mostly capacitors). Old laptops rather have standard electrolytic aluminium caps. Those are drying  fast and power on USB ports become very noisy. I had few problems with that on few laptops that was older than 3 years... Extreme case was laser mouse that was consumig ~50mA and that was too much, and mouse was working in random pattern ;)

Cant get this working :/

Just wait a little while ckolivas has received the nano fury i sent him and is working to incorporate support for it.

Why?
Grab a tplink and try it in case you are not able to compile. I can tell you it works perfect 2.5Gh per usb stick - marto's ones

I use it on debian and ubuntu.

I have used bfgminer for a while and I noticed a lot of messages similar "drop of frequency detected - restart". I have seen this on all my devices. Tried different things like dropping bits down to 30, active cooling, different powered hubs - no difference.

Did anyone else see something like that?

I switched to cgminer, so no problem, I am just curious.

Hey,
Pls clarify did you switched to cgminer - technobit.eu/0_1_3.rar
10X

I switched to  "patch to cgminer 3.8.5 rev afe7710858e4ce28bb60f6ae6e167a18d687634f" as mentioned in https://bitcointalk.org/index.php?topic=321287.msg4086475#msg4086475

or in short:

Yes :-)

The rar you mentioned contains a patch and the firmware for openwrt.

I downloaded cgminer in correct version and compiled it on ubuntu and debian using the patch and ignoring the openwrt files. Snippet of my history:

That is what i wanted to hear ;)
Thank you

Those messages are used for the dynamic clock/timeout detection. If I remember correctly, it uses received nonces per time period, figures in hash speed based on osc6. Or something similar

I'm seeing considerably higher hashing rates on cgminer with patch, than I was seeing with bfgminer.

With bfgminer, I was seeing 2-2.45GHs and I had to vary the oscillator between 50 and 53 to reduce the number of 'frequency drop, resetting' messages.

With cgminer, I haven't had to adjust the oscillator and I'm seeing between 2.5 and 4.3GHs. 4.3 is a rarity, but it does tend to average around 3.2GHs. I've seen 5 second averages hitting 4.5GHs!

I'm quite impressed by the performance increase. I also haven't seen any of those 'frequency drop' messages.

For reference, I'm running cgminer in an Ubuntu 12.04 Server VM, running on VMware ESXi 5.5 and the Nanofury is connected to an Orico 10 Port USB 3 Powered hub (12V 4A power supply). It has a heatsink and I have a USB powered fan blowing over it at all times. It's cool to the touch; so I don't think I'm seeing heat problems with this setup.

Thanks to the producer of the patch! - Nice work :)

-T

the 5 second avg means nada. its based on returned nonces over 5 seconds. hell you could get a work item where all 2^32 hashes returned a diff1 share.
but you need to compare bfgminers utility hashrate 3rd column after several hours running


edit: now that has me wondering what the (unrecorded) record is for most diff1+ shares in a single work item

I was finding that I saw larger numbers of 'frequency drop' messages the longer that BFG ran; reducing the oscillations seemed to reduce the number, but consequently reduced the hash rate.

Using cgminer, I'm seeing a higher reported hashrate both on cgminer and my worker on Slush's pool. I'm seeing a reported 3558MH/s on the worker; but that's a combined rate for the Nanofury and two USB Block Eruptors. That figure is a definite increase over what I was seeing before; which was closer to 2800MH/s

I think I probably need to do a lot more reading on the subject, but I'm definitely observing an increased hash-rate on the worker.

 

hmm. interesting.  same amount of runtime?

try each at the same osc6 for a day each and report back if u dont mind plz  :)

I'll see what I can do; but I'm not currently certain what osc6 cgminer is running at. I can see a '54' listed next to the name; so I'm assuming 54.

With CGminer running on Slush, my average over 10 rounds was 3243MH/s

With BFGminer running on Slush, my average over 10 rounds was about 2600MH/s; however I'll run this again with the osc6 set at 54. If I see a big drop in my average hash rate, I'll switch back; as the Nanofury is currently my heaviest lifter. At least till my order from Hashrate store turns up...

Cgminer is not changing osc6 dynamically. On startup you can set it with --hexmineru-frequency
It is making 2.5gh+ stable per stick assuming everything else is ok osc6 is at 54 by default. Your score 3+ GH is a lucky one ;)  you need at least 24 hours of mining to compare results .
If you are in doubt about hash rate you can always use accepted shares formula as suggested by Kano and double check accepted shares cgminer stats and your pool stats

There's the point I was trying to get at. Le sigh. I got too wordy :/

I never compared but I also think that the sticks are running faster with cgminer. But I am careful because cgminer does not report hardware errors.

And be careful two: Yesterday a psu of one of my hubs melted. It was a cheap one. It was rated to 2A and I used 4 Nanofuries with 54 Bits -> too much. Yes expectable, because with 54 Bits they are running outside of USB2.0 specs.

These beasts are really hungry for power and suck everything out.

Take care using good hubs.

That is why I always use Psu red and black :D
First thing to do is to garbage hub adapter but before that I am cutting the plug

if you have any 2A+ @5V i would take em

Unfortunately all are thrown away. But I will start to save them :)

No need to cut the cable...........http://www.ebay.co.uk/itm/4pcs-lot-Male-12V-DC-Power-Jack-Connectors-Cable-Adapter-CCTV-Camera-CAM-DVR-/221288401417?pt=UK_Sound_Vision_Other&hash=item3385d10a09

Then you will need to cut psu red and black :D
Anyway I am the last owner of my psu's so I am cutting them all then I spike and solder the wires quick easy and secure. Nothing melts anymore
In the beginning I used various extenders and crimp type connectors
My math showed me that I was spending for this as much as psu cost itself ;) and it was less secure

on v0.7 how do you choose the footprint for the MCP2210 I/SO or I/SS when making gerber files.

I would like to use the MCP2210 I/SO


thanks

The gerber files will have both footprints. Later during the assembly process you will solder it at the footprint that matches the chip (e.g. if you got I/SO chips you would put them at that place on the PCB).

I see


the holes that are under the bitfury chip is there solder coming throught them or does thermal paste go through them for heat

most likely a small amount of wicked solder. helps with thermal conductivity

Yes and Yes :)
Small amount does go through and that's specifically to help with conductivity (in addition to the copper coating of the holes). Unfortunately there is no way to control the amount of solder that goes through - sometimes very little, sometimes quite a lot .. and in the second case that makes a little bump on the back side, so you either need to process it a bit further if you want an extra-smooth surface, or just use a thicker stick-on thermal tape (that can tolerate some minor roughness).

hello to all. sorry but you can ' know where you can find the components to produce the product ? thanks

should be a BOM somewhere in the git repo: https://github.com/nanofury/NanoFury (https://github.com/nanofury/NanoFury)

I would like to achieve at least 10, but me and ' hard to find parts , even if you know some store on ebay that sells them , thank you

I would like to know if there 'to plan some component , or you just need to solder them ? ? Thank you.

P.S. sorry if I do so many questions, but in the Italian translation many words are unclear .

Ok, reporting back with my findings.

After over 24 hours running on BFGminer against Slush pool, my average hashrate over 10 rounds was 2779MH/s; which is a combined figure including two Block Eruptors ~335Mh/s each.

Compared to my 24 hours+ on cgminer at 3243Mh/s (including the two Block Eruptors) I have to say it definitely appears that cgminer is faster.

In both instances, this the reported hash rate at Slush's end; not the hash rate reported in the miner.

The VM was rebooted between the uses of CG and BFG; but the host was not. Aside from changing the miner, no other settings were changed and no patches installed. Both were run at 54 osc6. Currently there's nothing else running on the host and the USB hub has a 12V 4A power supply; so I don't think it's a power issue.

I do realise that there are other factors that can influence the hashrate, but it looks a lot like the Nanofury is under-performing while running under BFG. Can anyone corroborate my findings?

I didn't notice a difference between the two as far as hashing, but I did notice that cgminer was less stable for me over long periods than bfgminer.

I settled with the members , I think I find them all but the doubt remains , you must program some chips or welding and forth? ? thanks

Where is the BOM?

You can generate the BOM from KiCAD. There is no separate (extracted) file with that as the data will change with schematic changes.

Can you export the KiCAD to eagle schematics file?

does the nano fury need a heatsink or what hashrate will work with no heatseak

it can run w/o a heatsink but not at full hashing power (and not affect the life of components)

recomendation: run it with a heatsink

I couldn't find any option to export KiCAD files into another format ... Probably Eagle (being a commercial product) would have some import tools...

Yes, it will run without a heatsink - but in that case you are heavily dependent on the temperature in the room and any natural air flow. In my experiments in a room at 28C they ran smoothly for several days at 49 bits with no heatsink and no air circulation. With just a tiny fan nearby they were stable at 52 bits (still no heatsink).

So - yes, you could run them without a heatsink, but they're so much better with even the cheapest one that you can find around ($0.2-0.5) that it's just pointless not to add a heatsink.

I rehashed the whole blockchain a while back (up to block 262837) and found 23 had 8 nonces (though some of those 23 could have had more than 8)
Also, the stats output of cgminer when using a BFL asic shows the distribution of results ... up to 8 nonces (they return a max of 8)

interesting

I'm sorry, but I have a new question. when I open the file NF1 - MCP2210 - v07.net the list you see on the left and ' composed of 53 elements , while on the right and ' of 40 elements , which you have to follow ? ? any of you who made the project would pass me the list that he used ? ? thanks

It looks like official Nanofury support has been added to CGminer 3.10.0

-T

Can someone share the hash rate output from 3.10?
Thanks

Not good

http://s26.postimg.org/jgpcnwcnd/cgminer310.jpg (http://postimage.org/)

Yeah it seems not good but testing time is to low. It should run at least an hour or so. 5 mins are just not enough to avoid variance ;)

LOL, it,s getting worse so i switch back to 3.8.5   :D

Don't forget I always create drivers for cgminer that report hashrate based on real shares generated, not some arbitrary amount of hashes done (that doesn't translate into meaningful hashrate). I can happily make you a build that shows you making 2.5TH if you like. Make sure you check your pool to see how many effective shares you're submitting with your other drivers... People don't like the truth.

Forgive me CK, doesn't mean to offend you.

But i got 2 different error message.
http://s26.postimg.org/n1ptkv6kp/cgfail1.jpg (http://postimage.org/)
http://s26.postimg.org/fa93mb2fd/cgfail2.jpg (http://postimage.org/)

Just for the reference Con. My intentions are not to offend you like Zich said:

With cgminer technobit i do have

According to kano suggestion

https://bitcointalk.org/index.php?topic=170332.msg3358609#msg3358609

Two nanos:

Elapsed: 45959
Dev1 Diff Accepted: 29,184.00
Dev2 Diff Acepted: 25,088.00

My worker is mining at 512 so variance is quite big but but both are above 2.5 Gh/s. Bitfury chip is capable of it for sure
So i am living math to you.
In general at osc 54 it should be a way more than 2Gh/s

Best
However i think that 3.10 when run for a little bit longer will have same results

Any way i will test it myself when i am able to and will let you know

Best

Loshia - just out of curiosity - can you give it a shot with the binary that I used during my initial development and testing? It is available at:
http://www.nanofury.com/cgminer-NanoFury-bin-2013-10-02.zip
It supports only one miner (and sometimes doesn't shut it off when you exit) but it would be interesting to see how that one compares. In my initial observations the same miners with bfgminer did a bit better... but it is also possible that just my initial coding was messy.

Window build  :o
Which one to use?

48-56?
can you post a screen shot of it?

They're real crashes. Could be somehow related to being on non-pc (which is the only thing I tested on).

Both of mine perform quite differently:
Alas I'm travelling from tomorrow so I doubt I'll be able to do any debugging any time soon.

I also had a few crashes - core dump. Maybe because of weird system. Try to figure out on weekend.

System load seemed much lower with the version of main.

This is due to nonce checking. For sure Con version is doing it faster. I am sure that in new technoit release this will be tweaked also

the 48-56 are the hard-coded speed bits. That's the same build I used during initial development - e.g. like in the screenshot here: https://bitcointalk.org/index.php?topic=291456.msg3280783#msg3280783

Success! Perfect! RaspberryPI Model B. 2 YellowJacket NanoFury's (powered hub required)! 2.2-2.5GHs each! No more 50% frequency errors!

Why doesn't master branch of cgminer have this patch included?

Congrats! :-)

By the way I think Con did incorporate exactly those changes into the mainstream cgminer (I think as of version 3.11 everything should be stable, he also added a speed override option)
Here is a link to the latest release - 3.12:

Hrmm. I didn't see any references to this patch in the main branch, which is why I created my fork (https://github.com/utdrmac/cgminer). If I clone ckolivas's main branch, and compile, I have to use hidapi and I get hardware errors or I get no detection of NFY at all.

Using my fork, with the patch from technobit, there's no need for hidapi (known to be cumbersome with RPis), YellowJackets are detected, run more efficiently and produce 0 hardware errors.

Take a look here: https://github.com/ckolivas/cgminer/network
Go back to Jan/2 and follow the branch that Con split off at that time, and which branch got merged back into the trunk on Jan/8. That's when majority of the work was done.

Also, I suspect Con based his work mostly on the changes that Marto did (which were based on LukeJr's code who reworked to a significant extent my nf_spidevc.* (https://github.com/nanofury/NanoFury_Init/blob/master/nf1_init/nf_spidevc.cpp) library). The reason being is - Luke used the hidapi library (which was something inherited from my code) while Marto's version switched to libusb and that's what Con used. I don't know whether Con or Marto was first with libusb though - it might be that Marto has just shown the way to Con and he has redone the entire library himself.

I haven't done a diff to see what are the differences between Marto's version and Con's, but I suspect I'll find many common blocks of code :)
(which however might be just a coincidence as both of them used LukeJr's code as a base)

And just for the record - I would be a bit suspicious about the zero hardware errors. No bitfury chip (to the extent of my knowledge) has ever worked ideally flawlessly in that regard.

i might have 1 "flawless" at least at 54 bits. been running for over 1.5 weeks with no hw errors. this would mean its been through 2.091 × 10^15 hashes in that time period at 2.2 Gh/s which i would think is enough for a hit on each of the 756 cores.


and now i have to wonder why Bitfury chose 756 instead of a full 1024. silicon die sizing constraints?

Concerning diffs between Martos and Cons Code:
I have several NF, working without any problems with both versions. But one NF doesn't work with cgminer 3.12.0 (and before). It workes with Martos Patch and bfgminer as well.
It doesn't bother me but it shows there are differences between the patch and main.

My NF1 crashes on cgminer 3.10, 3.11, usually when there is a stratum disconnect. Rock solid on bfgminer.

This bug is fixed in 3.12.0. It wasn't in Martos Patch either.

bfgminer rock solid? No frequency drops there anymore?

Looking at stuff like that just makes me proud with my Nanos :)

Ah - a very good explanation and details for a question that seems to crop up quite often - the "how many of those can I plug to my USB hubs" one:

So where are these for sale at reasonable prices now?

I have a few for sale but they run under the standard speed.  I have sold all my others and these are the leftovers from my batch of 110.
At the default 50 bits they range between 1.5 Gh/s to 1.75 Gh/s.
Overclocked at 54 bits they all run over 2.0 Gh/s and the faster ones run about 2.5 Gh/s.
I have about 8 of them and they run between $55-$70 depending on the speed.  Delivery is additional.

PM me if interested.

Excuse me? I'll give you the benefit of the doubt and assume you're mixing up my code with something else. There's no code in the master cgminer code that uses hidapi, nor was there any code that used it. It only uses libusb directly which is included in the cgminer source, for that is the driver model for all usb devices in cgminer.

For sure Technobit reworked Luke vs3 hidapi implementation to libusb and this is noted inside their code.
Peace.

$55-$70 is a bit steep, especially if they are under standard speed.  I was hoping $25 to $30, expected $40 or so. 

Is there nowhere I can get $10/gigahash?  That's the only price that makes any modicum of sense right now.

No.  The lowest available price I have seen is $95.  But most are well over $100.

That's craziness.  No way to ROI @ over $30 right now.  I don't see it being even remotely possible.

well, you have to also keep in mind that those were made with a bitfury chip which alone cost $25. And even though everyone's bottom cost will be different - in my group buy my estimates for a 500-1000-devices series were for about another $25-30 in all other pieces including PCB, components, assembly, stencils, shipping, taxes, etc.

I get it; assembly of these had a certain cost.  But at this point in time, it doesn't make sense for me to buy them even at cost. 

I agree, but you need to realize these are not made to make money. They are entry level miners that introduce people to mining at the lowest cost available.  Their usefulness is more for education or the hobbyist who just wants be a part of the network.
If you want to make money mining you need to spend a lot more money than $100. Like $10,000 at least to make the effort worth it.
But if you want to dabble a little, learn about the space, have some fun and mine some millibits in the process, for less than $500 in equipment?.....this is for you.

Here is hoping those who have produced and sold these do chip back into your development costs. Really a shame if people do not support great designs and engineers like this that make projects that really work. Maybe if the BitFury chips come down to $5 a pop then more can be fabbed and people can help cover more of the development costs. Support your local Slide rule give genrously!

http://1.bp.blogspot.com/_hpxaqo7FS9c/S4wzM2rv8tI/AAAAAAAABi8/OCbosx7Hdv0/s400/math%2B-%2Bslide%2Brule%2B2.jpg

Sure, makes sense.

A little teaser with some cool stuff in the making :)

https://i.imgur.com/4HkA3Pb.jpg (http://imgur.com/4HkA3Pb)
(from observations and depending on your bravery - each doing 4-5GH easily) )

+10 ;D

TAKE MY MONEY!!!!!!!!!!!!!!!!

is there a git to look at?

Not yet (as it still needs some cleanup) but it will get there.

Did you think about an optional second 5V access? Because USB is limited it could be a possibility to attach an additional psu.

I did think about it, but at this time it will most likely not be necessary - most people would either use a powered USB hub (and those now easily deliver 1A and more per port) or use USB3 ports.

In one of my tests I used a powered USB2 hub where the power brick provides 3A (plus 0.5A from the PC) and that was sufficient to attach and run 3 devices, and actually push them all the way to their limits. So by just not using all ports you leave the extra power for the ports that are in use thus achieving the same effect as if you use 2 ports simultaneously.

Plus, the extra USB connector would get in the way the simplicity idea :)

?
what u mean with that?

this is a trusted guy, making good boards.

Trusted guy DESIGNING good boards.  We make them.

I know. Vs3 is an excellent person. I purchased some from his first run of the nf1.

The comment was not made as an insult but rather encouragement and excitement.

I'm definitely going to get set up for a DIY build and my challenge to myself is to see how low I can keep the costs.  Now that I've seen that the NF2 is in the pipeline I'll wait for that to be available.

vs3 - I'm assuming it's a 2 layer board and that the back is just ground?  Thinking about the viability of etching my own boards :)

Can anyone point me in the right direction to source a small number of bitfury chips (probably about 5 - 10 of them) ?

If I'm successful anyone in the UK interested in a DIY build kit of parts?

Thanks
MonkeyThink

There are companies in the uk that will make boards, small quantities are expensive, but the price drops rapidly for bigger orders.
also lots of guides on utube for etching boards.

Indeed there are.  I fancy the challenge of etching my own for a small number of prototypes. 

If there's enough interest then I'd get a bigger batch made up.  I see there is one UK based company who actually has the work done in China or Singapore.  From memory they worked out to about £3 per NF board if you were ordering a few hundred I think.

For me the only doubt I would have would be supplying the right files for a commercial production run out of the KiCad design.

Regards
MonkeyThink

I looked at the NF rev7 and eventually came to the same decision as you.(It becomes a big investment risk to go commercial)

The prototype would have been interesting though.

If you do go ahead please keep posting.

DYI/making your own miners is always the most fun :)
Here is an example:

As for:
I got a bit lazy with v2 - there is one track on the back :)
I already got an idea on how to take care of that (in version 4 .. if we make it that far at all.. as that's going to be pretty much the only improvement and I'm not sure it's worth the hassle)

Some more of the NanoFury NF2 teaser :)  :


for those wondering - there are 3xNF1 and 4xNF2 miners. I'll leave it up to you to figure it out which is which :)

so the NFY driver works without mods?

The NF2 have 2 chips or 1 and this only lead to maximum performance?

This will wipe out any remaining value in a BFL jally  ::)

BFGMiner git (and therefore the next version) has added support for NF2.

Check out the design / prototype pics a page back.. that plus the informal name 'NanoFury Duo' should be a hint :)

How much and when?  ::)

BitFury chips are due in today and then we will start manufacturing a couple hundred.
We hope to have them out by mid March, but it's a new design.  We don't expect any issues, but there usually are in bitcoin.

Is there any way to update the firmware of these devices such as blocks Bi-fury, to reach some more Hasheo.

thanks

They don't have a firmware limiting the hashes. All you need is just the oscillator option in bfgminer/cgminer (a.k.a. osc6). Or if you want to do some serious overclocking you may also consider doing a pencil mod to increase the voltage too. Have in mind that in that case you will need to deal with a lot more heat, and obviously provide provide more power.

Thanks for responding, I already knew oscillator but the Mod with the pencil do not know, can you give me more information on obtaining more power even endangering miners.

Sure - here you go:


One clarification - on the early NF1s the voltage was 0.833V, while on the later ones it is 0.86V (R2 is 180 ohms) - it's a bit more stable there, and you also get a bit of a "factory overclock" :)

Any news on NF2?

I think https://bitcointalk.org/index.php?topic=250249.msg5480556#msg5480556 is interessting because the new chips need less power. So it should be easier to get two on a usb stick. Because they are pin compatible it should be possible to just put them in?

If anyone needs samples:
http://www.bitfurystrikesback.com/product/bitfury-55nm-rev2-samples/

Yup - I've ordered some samples - they're on the way. As soon as I get them I'll be testing them with the NF2 design.

VS3 i like your design and cant wait for NF2
I`m ordered 32 chips so have some to playing with it:D
How cgminer or bgminer detect  NF1/NF2 ?
are you planning  build more powerfull miner with more chips?

I'm working on a fork that's something like that.

Making my baby public

https://github.com/Taugeran/NFx

She's nowhere near as elegant as the NF1 or NF2 but that's what revisions are for. When I go home for lunch today I'll push a new commit that has some updates

I do not want high end pcb only working to study and in near future create my own modular miner (not for sale)
Thanx i will study that when you submit updates:)
Schematic is on kicad?

Hello

bfgminer will detect both.
cgminer as far as I know doesn't detect the NF2 (it detects only first chip so it will essentially behave as a NF1).

I am also working on a USB miner with more chips but that's a bit questionable design as you'll somehow need to provide 5V at 2-3A to the device (which a standard USB port can't do). So, we'll see where that ends.

so generally cgminer right now  dont support nf2 ?
Power is not a problem got  from LED boards some PSU 5V 120Amps:D
i`m trying how connect chips to get in chain with power, meantime wainting for samples chips :)

plenty of 5v & 12v here
https://www.dropbox.com/s/dpjtcouau6ah0kj/2013-11-02%2012.41.13.jpg

There is a different bitfury USB stick that cgminer supports with 2 chips. It would be trivial to add support for the 2nd chip to the existing NF1 driver as well, but as always would need testing on real hardware.

yes. schematic and board are in kicad. update pushed to github

Thanks i will study that tomorrow :)

where can these be bought if available?

Just to clarify, precrime3 is probably referring to the hex•fury.

( Original NanoFury-based products like the yellowjacket, hitchhiker and icefury can be purchased new from several sellers. )

Yes thanks for clarification xD

Ok my study schematics is complete but got quick question.
Is 3A DC/DC Swithing regulator will be good for one chip?
i got plenty 3Amps regulators and LDO :)

Yes 3A is what is used in the NF1. The NFx is basically 10 NF1 relaid and interconnected

Soft Release of Nano Fury 2 USB Miners – 5 Gh/s – Dual Bitfury Chip Miners

Please follow this link for UPDATED details:  https://drive.google.com/file/d/0B_tqfkR9Qrx0VVo5UGZFNnR1LVU/edit?usp=sharing (https://drive.google.com/file/d/0B_tqfkR9Qrx0VVo5UGZFNnR1LVU/edit?usp=sharing)

EDIT - PRICE REDUCED 20% TO 0.18BTC FOR THIS PRE-RELEASE

Thanks for your help vs3!

More pictures: http://imgur.com/a/WKycW#0 (http://imgur.com/a/WKycW#0)

https://i.imgur.com/P5GTePm.jpg

When is hard release  ;D willing to pick up some if price is right.

Friday, but you can order now if follow the link.  Price will not change.  We are just not broadcasting everywhere yet.

Awesome! :-)

From all pics on that album I think those are my favorite:
https://i.imgur.com/nljOPS7l.jpg (http://imgur.com/nljOPS7) and https://i.imgur.com/NNQJpg1l.jpg (http://imgur.com/NNQJpg1)

And just to add a minor detail - one of the factors that was considered during design was "The Overclocker" :)
So especially for those people I picked a really good heatsink - you can probably tell by the pictures already, but it's even more impressive when you see it in action!
In addition, it's a standard 40mm x 40mm one, so the extreme overclocker can even add a GPU fan (yup, some GPU cards use 40x40mm fans, powered by 5V - read USB will work! ;) )

As for overall dimensions - the board is 40mm wide and 50mm tall and as for thickness it's approx 16mm (the tallest component on the front is the 4mm inductor and on the back the 10mm heatsink).


And as always - HAPPY HASHING! :-)

Oh - and thanks for the awesome pics!

Can you tellme what DC Conventer you use to supply those chips?
I`m build my own minner with modular chips but now stuck with power supply :)

.2 btc per.... Bit pricy. Let me gather some finds xD

Just based on non-OC'd, USD/Ghash-wise, it's fairly competitive ($33), especially if you consider it overclocked ($25); Beaten only by the recent price drops on the Bi•Fury (also 2-chip, $27/$26), Yellowjacket (1-chip, $22) and AntMiner U1/U2 (1-chip, non-bitfury, ~$18 for either) products.  They all have other pros/cons, of course.
These figures do depend on things like bulk discounts (if you buy 500 AntMiners, the USD/Ghash more than halves - but who buys 500 AntMiners except for resale?  So I only included ones that can be individually purchased)

I'm really looking for them 15 gigahashes, but it seems that they would be .6 btc each.

directly PM user Intron or user cscape (cant remember if there is a hyphen so may be c-scape) they designed them and may be willing to sell one privately

I pm'ed intros, he said watch the forums, which I have been doing xD. No word yet.

Are the schematics of NF2 free like the ones of NF1?

Yep. I have a bit of work left to finish them (mostly un-customizing them) and will be uploading them on github.
I am mostly focused on testing another design right now though, so I'll probably get back to that in the next few weeks.

            NANO FURY II is released.

Follow link below for sales thread:

https://bitcointalk.org/index.php?topic=526925.0 (https://bitcointalk.org/index.php?topic=526925.0)

I stumbled onto a pretty interesting video review of the NF1 (IceFury version) :)
(including covering all the important how-to questions)

https://www.youtube.com/watch?v=7_HVT5nAZMk

Several people have asked me for a NF2 DYI kit.
So I ordered some extra boards - they're 8 boards per panel and should be arriving at some point later this week or early next.

I have a few extra panels left. I also have an extra steel laser cut stencil.

Pricing will be as follows:
----------------------------
8-board panel: $80
Stencil : $70
Components pack (all components for 8 boards excluding bitfury chip) : $100
----------------------------
Ground shipping within the US will be free. PM me for international.

Note: heatsink and thermal tape are not included.

TO PLACE AN ORDER:
PM me with what you want and I'll send you a quote with the final amount and payment instructions.
Please put DYI KIT in the message subject line.
For the quote I'll need a shipping address, and if you prefer UPS/FEDEX also a phone number.
Payment options (in the order of preference) will include: BTC (coinbase rate), check, paypal.

I anticipate having everything in my hands and ready for shipping at some time next week. Based on the interest I'll be ordering the missing components on Thursday (with expected delivery on Monday).

Below are images of the panel and actual boards with stencils. Boards on the panel have been specifically laid out for an easy automated SMT machine assembly.
https://i.imgur.com/9OXwcXdm.png (http://imgur.com/9OXwcXd) https://i.imgur.com/mzkm9wA.jpg (http://imgur.com/mzkm9wA)

edit: added clarification regarding heatsink and thermal tape

When will NF2 files be released ?
And for the NF1 are the files 100% ready for production? And can u confirm that the version on the files are supported by cgminer ?
Thanks,
JT

Got the boards and stencils!

And all I can say is - WOW! They're actually a much higher quality than what I expected!

https://i.imgur.com/0MnMpjIm.jpg (http://imgur.com/0MnMpjI)

Care to share from where? Hard to read the package.

To be absolutely honest - I have no clue where they got made! :)
I asked one of the companies that did assembly for a previous run to get me some extra boards and a stencil and that's what I got. What I suspect is that they're ordering those things from the far east somewhere (as it took a few weeks). I'm not complaining though - the quality was worth the price I paid and the wait.

I'll give them a call tomorrow to see if I can find out more details though.

Thanks man. I sent you a PM. Keep up the great work!

From China…

Well, there is a group buy for Rev2 BitFury chips.  Supposedly protocol and pinout compatible (datasheet says it is anyway), so... should be able to get those onto existing NanoFury(-derived) boards as well :)

Bitfury chips had a data-what? What sheet? :D

I'm buying a drink to whoever can point me to the "official" such thing.


Humor aside though, preliminary tests indicate that Steve's statement is correct.
Just put the new chip and everything else works as expected and better.

:D

At least for Europeans there is good opportunity to get hands on an engineering sample of max 32 Bitfury rev. 2 chips. Punin still sells engineering samples in his shop. Good for two 8-board panels. ;)

https://i.imgur.com/rmVYRa7l.jpg

yup - which is why I made the panels consist of 8 boards ;-)

(And coincidentally most people requesting DYI kits have asked for 2 panels ... I dunno why :D)



Speaking of the kits - I have 5 panels left and one stencil ...
(I wish now that I had ordered more panels ... if I only knew that they'll sell out so quickly!)

The boards are looking great VS3! Your work in designing the most simplistic USB miner is astonishing. Also it is good to see consumers liked the icefury :)

Any info on the NF6 yet?

Heh heh. He's working on it I promise. :P

Here is a NanoFury Sixer teaser :-)

https://i.imgur.com/35uCTijl.png (http://imgur.com/35uCTij)

Now the most annoying and nail-biting part begins ... waiting for the real boards to arrive ...

Looks nice vs3.

By the way: I wanted to ask if I have to flash the controler of the nf with a firmware....probably...but how is it done, if it is necessary at all.

Thank you and best regards.

Max

String-liked NF6?

If today had been any other day but 1st April, I would have asked how to input the other 40watts  ::)

Nice! Let's see if it can do 15+ GH;)

Has anyone managed to get more than 10 nf1 running on a beaglebone black.
It doesn't seem to pick anymore up when I run bfgminer and gets a MCP error

yup as in string-like NF6. :)helped him with some realworld data measurements to verify his ideas

Nope, no firmware necessary (since there is no additional microcontroller) - which is the beauty of the NanoFury design ;)

However, you do need to run a small program to initialize some settings on the board (namely the product string) - which is required for some miners (bfgminer) but not for others (cgminer). But all of this is already described and detailed on the project git page.

yup - they're chained (a.k.a. in a "string").

@ vs3: thank you very much....can't wait to build my nfs ;-)

Greez

DYI kits on the way! :-)

https://i.imgur.com/yu7R8Jxm.jpg (http://imgur.com/yu7R8Jx) https://i.imgur.com/MiI6F5Gm.jpg (http://imgur.com/MiI6F5G)

I have only 4 panels and one stencil left!
(I wish I knew that DYIs would be so popular and ordered some more panels...)

I would like to reserve 1 (8 board) piece + components, no stencile as i can do by hand.

That's what this one's for ;)
https://i.imgur.com/5i2jQa3t.png (https://i.imgur.com/5i2jQa3.png)

It needs a power lead, a wi-fi twig and some white led's.
Then it could be suspended from the ceiling and used for space heating and lighting as well.  8)  ::)

H-card to USB converter.

Letting it ride with 8 chips 50 osc bits(to stay in usb specs as i read on bfg readme)
http://s25.postimg.org/ijmbsppnf/bfg_8.jpg (http://postimg.org/image/ijmbsppnf/)

If someone could solve the 50% frequency drops in bfg it would run great and also overclock.
Or if there was support fort multiple bf chips in cgminer...

sweeeeet :-)

btw the 50-bit stuff is meant for NF1 - from observations at around 50bits it uses 0.5Amps (hence "within USB2 specs").
In your case this doesn't really matter so - as the guys from Eagles sing - "take it to the limit"! (a.k.a. girls favorite song about credit cards :D)

Good bye free time ...

https://i.imgur.com/sxlAFFQl.jpg (http://imgur.com/sxlAFFQ)

Hello fun! :)

Yay!/me prepares for 'visual history of NF6' ;)
( Though hopefully v0.1 is bang-on :) )

8 chips seems to be the limit, after that(currently at 12 chips) it gives alot of 50% drops and reinitialising, every time it reinitialises it brings a HW error with it, most of the times.
As i said, if it truly reinitialises the chip i think it breaks the chain and has to start all over again... i dont really know.
If i tweak the osc bits for all the chips to top out at 12ghs it works, no hw errors, no drops.

The H-card works good(all good chips) with an m-board, so the card is not the problem.

The mcp2210 should be able to take more but i dont know how to tweak the software, maybe someone could help.

I don't have one so cannot write the support for it. I only have 2 single chip icefuries.

I could send you a fully populated H-card+adapter, like you see in the picture all with good chips.
Default speed or a little overclocked from 25ghs to 30ghs(in chainminer).

pm me the address ... wish you were in Europe... but its probably US :)

Con - PM me a mailing address please - I have a few NF2s from my engineering units and I might be able to repair a few more. I'll send you also one of the engineering NF6 boards.

Speaking of which - mostly good news:

 bfgminer version 3.10.0 - Started: [2014-04-04 05:51:05] - [  0 days 00:09:14]
 Manage devices Pool management Settings Display options
 Connected to stratum.mining.eligius.st diff 16 with stratum as user
 Block: ...cc9d29fc #294158  Diff:5.01G (35.84Ph/s)  Started: [06:08:38]
 ST:7  F:0  NB:3  AS:0  BW:[253/ 19 B/s]  E:12.87  I:30.81uBTC/hr  BS:2.34k
 1/5           |  7.47/ 7.51/ 8.42Gh/s | A:59 R:1+0(1.7%) HW:43/3.8%
----------------------------------------------------------------------------------------
 NFY 0a:       |  1.54/ 1.55/ 1.83Gh/s | A:11 R:0+0(none) HW: 0/none
 NFY 0b:       |  1.54/ 1.54/ 1.66Gh/s | A:16 R:0+0(none) HW: 6/2.7%
 NFY 0c:       |  1.42/ 1.42/ 1.43Gh/s | A: 8 R:1+0( 11%) HW:14/6.3%
 NFY 0d:       |  1.48/ 1.48/ 1.84Gh/s | A:13 R:0+0(none) HW: 9/3.6%
 NFY 0e:       |  1.50/ 1.51/ 1.62Gh/s | A:11 R:0+0(none) HW:14/6.3%
----------------------------------------------------------------------------------------
 [2014-04-04 06:04:42] Accepted 0869332c NFY 0d Diff 30/16
 [2014-04-04 06:04:49] Accepted 0befdfb6 NFY 0d Diff 21/16


I'll have to find out tomorrow what have I messed up with the 6th chip (or it might be just a bad one). I also had to correct several component values, but that was the easy part :)

With those stats - power usage 1.05A (on a very crappy hub - voltage dropped to 4.1V!), hottest heatsink temp 68C, passive cooling/no fan.

http://s25.postimg.org/8322w0i73/H_card_to_USB.jpg (http://postimg.org/image/wwbmwo17f/full/)
img upload (http://postimage.org/)

https://www.dropbox.com/s/zdta1eu4ucgq7ai/H-card%20to%20USB.jpg

To be clear, the schematic is not mine, its an NF1 stripped down and active level shifting installed.
The original can be found here:
https://github.com/nanofury/NanoFury/tree/master/NF1/kicad/NF1-MCP-V06-2

The image on dropbox is 1:1 scale so you can print it and make the board.
2 layers, bottom one is groud, where you see a dot and the gnd sign, thats a hole that goes to the bottom layer(gnd) and has to be electrically connected to it.

From what i see i cand get past the 12ghs barrier, maybe because the frequency of the mcp 2210 is 12mhz or the usb bus....no clue, im no software engineer.

in your driver-nanofury.c

around line 255 or so what is

port->speed = 200000;

if its still at 200K try upping it to 500k or so, recompile bfgminer, and report back



EDIT: just tack on the end this will accomplish the same thing



this set the baud rate for your frankenstein nanofury, and effectively the data rate. if 8 manage to run on 200kHz i wanna know what 500kHz can support, b/c i had a few ideas (may or may not involve strings [beyond 6]. itll literally look like an empty board, with the part count.)

Hi Folks,

At first look on your schematics i think that B1 from level shifter need to go direct to SCK on MCP2210, GPIO5 from MCP2210 is used to toggle SCK and send reset to chips, so the connection is wrong ( please check that ).

Cheers,

EDIT: just tack
Code:
--set NFY:baud=500000
on the end this will accomplish the same thing

I use the 3.99 windows version of bfgminer and it just exits.
Anyone with another version? latest 3.10 i tried but no luck, just sees the 1st chip.


"Hi Folks,

At first look on your schematics i think that B1 from level shifter need to go direct to SCK on MCP2210, GPIO5 from MCP2210 is used to toggle SCK and send reset to chips, so the connection is wrong ( please check that ).

Cheers,"

On the original schematic that 1k R seems to be part of the passive level shifting.
Ill try connecting it directly. Thanks.

UPDATE: Direct connection to SCK - no go, it fails to detect nanofury.

You must leave 1k resistor, GP5 goes trough 1K resistor to SCK, B1 goes direct to SCK

Just saw, the schematic(printed pcb) is actually missing the last part... exporting issue i think
https://www.dropbox.com/s/uvril5p01dzd8k5/H-card%20to%20USB.jpg

Correct schematic uploaded, you do need a jump wire on the INSCK line

UPDATE: Connecting SCK directly gives bitfury_init_oldbuf error

http://s25.postimg.org/8322w0i73/H_card_to_USB.jpg (http://postimg.org/image/wwbmwo17f/full/)
img upload (http://postimage.org/)

MORE PICURES OF WHAT SHOULD BE CALLED FRANKENSTEIN NANOFURY :))
https://www.dropbox.com/s/mm3nv70c8gurtll/2014-04-04%2020.06.42.jpg
https://www.dropbox.com/s/l0i4n79hs8ey86q/2014-04-04%2020.06.50.jpg
https://www.dropbox.com/s/l0i4n79hs8ey86q/2014-04-04%2020.06.50.jpg
https://www.dropbox.com/s/ocs6e5t9am9m8fg/2014-04-04%2020.08.12.jpg

Well then it is connected correctly  ::)

anyway i am using this http://www.mikroe.com/click/usb-spi/ ( i am trying for now  ;D ) how you run it on windows ? any special drivers/setup needed ?

I use this BFGminer development version
luke.dashjr.org/tmp/code/webisect/webisect.php?dobuild=bfgminer

no special drivers like zadig, if you have zadig just click on update driver->manualy select from list->select USB input device

bfgminer -o stratum+tcp://............ -u ........... -p x --no-submit-stale -S nfy:all --set nfy:osc6_bits=45

this is for 12 chips, as chip count goes up, osc bits must come down to keep the 50% drop error down.

maybe ill try linux on a stick to be able to modify the baud rate.

Linux with 500000(tried 300k and 400k) baud its doing worse than windows at the same clock bits linux-44 to get stable, windows 47- to get stable

I still belive mcp 2210 can hack it, the problem is i dont even know how spi works let alone trying to figure it out.
Mcp is rated o 60mbits/s, there are devices out there that transfer 30ghs over usb so i dont think its a bottleneck...or maybe im wrong...im not the expert.

I just like circuits :))

i have finally menage to comunicate with my board over mcp , i have restarted pc  ;D

i want to have usb option on my 14 chip modules, at the moment i have only 4 chips running on module and that module is connected to spi click board, its working but i dont have accepted shares at all and i get error with bfgminer abot not recognized --set command ...

i dont know is it smarter working with mcp2210 or build new hardware based on pic32 maybe, is there anybody with some free time to write code for this ?  ;D

maybe to build controller capable to drive 16 boards with 16 chips or so ... but on one USB  :)

anybody ? some more ideas about this ?  ::)

you can skip the --set command and play with the bits after.
I belive the mcp2210 is good and 16 chiop resonable if it would be able to run them.
Given that bf chips are so picky, different from each other, iffy... and sometimes Lazarus style resurected :))
Its easier to mantain a 16 chip board and unplug it letting the others hash.

Im thinking more on easy mantaining and no spi errors and oxidation problems and so on.

Thats why i want to make it work, maybe there is another more capable spi to usb converter chip....

maybe this one
http://www.silabs.com/products/interface/usbtouart/pages/usb-to-spi-bridge.aspx

also to rehash what vs3 has said in the past there is a limit to how many check chip txrx can happen per second. this is based on what seems like an unforseen downside to using the mcp2210. since it is recognized by the OS as an HID (Human Interface Device), it is limited to approx. 64kbytes/s bandwidth. this is an OS level limitation built into the OS HID driver since most HID devices are low data rate e.g. keyboard, mouse, etc.

Each check chip txrx is I believe 500-800 bytes. So let's round up to 1kB. This would allow only 64 checks per second which need to be spread evenly over each chip.

This is compounded by the fact that the mcp2210 imposes an artificial 40-70 microsecond delay between each SPI Tx

http://www.silabs.com/products/interface/usbtouart/pages/usb-to-spi-bridge.aspx

Would this make a difference?

what i want to achieve has actually been done with bitburner fury... but thats out of my league of DIY

Maybe zadig and cgminer would make a difference...

That chip would absolutely serve your purposes. But a lowl driver and interface code for bfgminer would need writing etc, it would function faster than the Mcp2210 since it uses bulk usb versus OS HID. Only downside is dev time.


The bitburner fury uses a microcontroller if I'm not mistaken.  Which is not only hw dev time but the pita that is embedded development

There is one extreme option, cut the traces on the h-card dividing the chips in banks of 4, use 4 mcp2210 and a 4 port usb hub for each card.
As i said.. extreme
And as i said earlier maybe zadig and cgminer would make a difference...

Or into 2 groups of 8. 6 should be entirely stable. 8 is cranking it to eleven so to speak but more than 8 is most likely going to remain unstable with the current iteration of the mcp2210

You know on second thought... That's not as bad an idea as you think. If you are comfortable redesigning the hcard a little you could turn it into a 4-up of NFY NF4 with an inbuilt 4port USB hub and a pcie 6 pin connector for power. Replace the edge connector with a micro USB port and tada...

If zadig is installed is it still seen as a HID device?

For cgminer? Negatory.

After zadiging it, run cgminer. Should auto detect.  Also I'm not sure if cgminer has support for multiple chip NFY devices...yet

Not yet but i offered o send the FRANKENSTEIN to ckolivas for research.
Post no. 262.

Thanks for that. I don't have the power to run much more in my house and I'd hate to see hardware go to waste. Maybe the 2-6 units alone are enough to get me going.


Thanks vs3, I'll take you up on your offer.

I figured out what was wrong with the 6th chip :-)

bfgminer version 3.99.0 - Started: [2014-04-05 02:07:05] - [  0 days 00:09:16]
Manage devices Pool management Settings Display options
Connected to stratum.mining.eligius.st diff 16 with stratum as user
Block: ...484af8b3 #294330  Diff:5.01G (35.84Ph/s)  Started: [02:14:47]
ST:8  F:0  NB:3  AS:0  BW:[203/ 27 B/s]  E:20.23  I:40.96uBTC/hr  BS:1.72k
1/6           |  9.72/ 9.55/10.65Gh/s | A:75 R:2+0(2.6%) HW:43/3.0%
-----------------------------------------------------------------------------------------
NFY 0a:       |  1.64/ 1.66/ 1.77Gh/s | A:17 R:1+0(5.6%) HW: 6/2.4%
NFY 0b:       |  1.64/ 1.65/ 1.91Gh/s | A:12 R:0+0(none) HW:14/4.7%
NFY 0c:       |  1.53/ 1.54/ 1.59Gh/s | A:10 R:1+0(9.1%) HW:15/6.4%
NFY 0d:       |  1.57/ 1.61/ 1.66Gh/s | A:11 R:0+0(none) HW: 4/1.9%
NFY 0e:       |  1.64/ 1.66/ 1.83Gh/s | A:12 R:0+0(none) HW: 5/2.1%
NFY 0f:       |  1.68/ 1.70/ 1.83Gh/s | A:15 R:0+0(none) HW: 0/none
-----------------------------------------------------------------------------------------
[2014-04-05 02:12:52] Accepted 0c1ee27c NFY 0c Diff 21/16
[2014-04-05 02:12:52] Accepted 099ab634 NFY 0d Diff 26/16
[2014-04-05 02:12:52] Accepted 0c06c938 NFY 0d Diff 21/16
[2014-04-05 02:12:52] Accepted 0d599353 NFY 0e Diff 19/16
[2014-04-05 02:12:54] Accepted 0ce9c358 NFY 0e Diff 19/16
[2014-04-05 02:12:57] Accepted 0826a4eb NFY 0b Diff 31/16
[2014-04-05 02:13:05] Accepted 047c994e NFY 0a Diff 57/16
[2014-04-05 02:13:06] Stratum from pool 0 requested work update
[2014-04-05 02:13:10] Accepted 0dae182d NFY 0e Diff 18/16

I'll have to try with a different hub tomorrow as this one is quite crappy (at 1.15A dropping down to 3.9V).

I have ever tried 5-chipped NFx. (NF5?!) last week with bfgminer and gotten 12+G/h. But I forgot to take a picture to show what I said...sorry.

looking fwd for a diy kit for nf6. rework station & some v2 bfchips too is on its way.

I do not use any hub, but use a kit I made. It's simple and works.
https://dl.dropboxusercontent.com/u/23682587/DSC_0223.JPG
https://dl.dropboxusercontent.com/u/23682587/DSC_0222.JPG

https://dl.dropboxusercontent.com/u/23682587/cgminer-bitfury-3chips.png

3-chipped NanoFury (NF3!?)

I will post 5-chipped picture later...

Okay, I think it's time to get some feedback from the public -

the 6-chip design will most likely be limited to around 12GH. The reason being is - with 6 chips each chip gets around 0.8V in which case the bitfury chips do around 2-2.2GH max. For more gigahashes we need higher voltage. (the NF1/NF2 are running with 0.86V).

And the 0.8V per chip is the ideal case - most USB hubs will probably crap out and instead of 5V will provide somewhat less (thus even lower voltage per chip and lower gigahashes).

The alternative is to try a 5-chip chain. In this case we'll have around 1V per chip and they will do significantly more than 2GH/chip (likely in the vicinity of 3GH+).

Pros and cons:
6-chip:
pros: relatively low power usage of around 1-1.2A, probably 1.5A max when pushed, resulting in relatively low heat production
cons: wasting an extra chip ($5-6)

5-chip:
pros: will almost certainly do more gigahashes with one less chip per miner
cons: will use a lot more power (likely over 3A) and will be very very very hot


so - the question is - should we even bother looking at a 5-chip version, or stick with the 6-chip even though it has some disadvantages?

Considering neither configuration is going to ROI unless a lot of crazy things happen, NF6 design is well underway, doesn't require active cooling (which the 5-chip one sounds like it probably would) which tends to be something people seek to avoid if they're just doing it for fun, and the 5-chip*~3Ghash/s would only be 3Ghash/s faster,  I'd say don't bother.
Would it not be easier (not sure if feasible with the string design) to optionally allow an external power source to use a higher voltage for those wanting it running faster / daring enough to overclock?

Are there any other pros/cons you can think of?

Though if a 5-chip chain is possible, I guess those DodecaFury panels could be sold separately after all ;)

lol @ dodecafury.


i concur with steve. the NFY i have presently dont need active cooling but perf is better with it.  same with the OneStringMiners. the DIY0 with no heatsink no A/C does about 25-27Ghash. with some A/C and vga heatsinks, jumped to 30-33

Im all in for max heat max power max ghs :)

As for my plans for the H-card-usb converter....
An H-card can take 12 chips@0,9...0,95V and you get about 30ghs/card this way.
My plan is to split them in 6+6 on two usb converters/card.

Thats is if cgminer doesnt change the playground.

Okay, so we'll stay the course with the NF6 :-)

I personally prefer the option that uses less power - 1-1.2A is just above USB3-specs for power and most "mining-tailored hubs" would have no issues with 1.5A, but going for 3A+ will be a completely different challenge by itself (plus most cheaply-made connectors may not take that very well).
So in terms like that the 6-chip is much more end-user friendly (although not necessarily very geeky :)).

As for over-voltaging - that is a viable geeky option - most PC electronics will tolerate up to 5.5V. So if they do let's say 5.4V this will result in 0.9V/chip where they can definitely perform a lot faster (almost certainly beyond 2.5-2.7GH/chip).
Alternatively, they can do a relatively easy board mod and remove the fuse and supply any external voltage they want (although I would caution against going with much over 5.5V as there are some inter-chip dependencies and with higher voltage things may very quickly go beyond what chips can tolerate)

Just out of curiosity - would that (the geeky 'over'-volting, or at least stable-volting, not going above 5.5V) be as simple as disconnecting the USB Vcc (cut the trace/whatever) || dropping in a schottky, hooking up the external power supply and share the USB GND, or are there some other considerations?

yup - exactly as simple as that

Yeah, then I definitely recommend just sticking to the NF6 design.  If it is still in the design phase, you could make that a little easier by including two pads to solder some external power leads into/onto (bonus points if you can make it fit some common connector part), but I'd suspect there should be some pads for parts that could be soldered as well, just a bit more tricky for people new to soldering :)

I hope the NF* variants have a unique identifier on USB to make them easy to distinguish on USB.
The nanofury currently nicely identifies itself as:
Managing variant devices is much easier when some forethought goes into it by the hardware devs thanks ;)

I can send you my code if you like. I modified cgminer 4.2.3 for NFx already.
I get the number of chips by finding "NFx" where x is for chips, i.e. NF1 for 1 chip, NF2 for 2,
NF3 for 3,...NF5 for 5, NF6 for 6...and so on...
The freq of all chips are the same. I have no idea to add an optional to set different freq for each
chip. As the picture I showed that is tested on "NF3".
https://dl.dropboxusercontent.com/u/23682587/DSC_0226.JPG
https://dl.dropboxusercontent.com/u/23682587/DSC_0227.JPG

Later, I will show NF5...

I would be intrested in the code if youd like to share it, id like to test it on the h-card-usb thing.
Just point me where i have o change the numbers.

Thanks

Download this tool to modify the product string of MCP2210 :  http://ww1.microchip.com/downloads/en/DeviceDoc/MCP2210Utility_v1.2.1.zip
Read, modify and configure... quite easy. If you use zadig to access MCP2210, you have to remove/uninstall it and change to windows default HID driver such that microchip can access MCP2210 correctly. I will put my source to github here tonight, 'cause that is at home, not in office.

First - CONGRATS! :-) Nicely done!
And that's exactly why those 3 extra test points are there and how they were meant to be used :)

Oh - and I think that's the first case when I see someone adding a 40mm fan :) (which was another one of the other design considerations)

Just a word of caution - I see that you're powering the 3rd chip from the same power regulator - be aware that the regulator is rated for 6Amps so if you push the chips a bit too much you may exhaust it. (nothing scary is going to happen - just some serious voltage ripples which will mess with bitfury's internal clock generator and you'll see very low hashing speed and lots of errors)


As for the NFx numbering - that is correct, the "x" was meant to indicate the number of chips. This is all part of the "product string", which is a customizable element.
Luke's bfgminer looks for the text "NanoFury" and that's how it decides to do any further tests with that USB device. There is also an additional test that checks the device (a hardware check). The NFx part is not mandatory for bfgminer, but I've been adding it anyways for consistency.

I've been using the https://github.com/nanofury/NanoFury_Init program for setting the "product string".

The product string can be up to 25 characters long, and I've been adding manufacturer-specific identifiers after the NFx part (e.g. "NanoFury NF2 v0.5 by MLM").

I don't know if any of those strings have been used by cgminer and how it detects the devices (although I could probably take a look at the code and figure it out).

There is one addition that I've asked Luke to add but it is not there yet - I want to add a "default speed" indicator, which will be in the form of "/N" after the NFx part - e.g.
"NanoFury NF2/t" where the letter will represent the default speed, calculated by the formula 30+modifier, where upper-case letters are negative (@=0, A=-1, B=-2, etc) and lowercase are positive (a=+1, b=+2, c=+3, etc), and in this case t=20 so that's default speed 30+20=50 bits.
And the speed designator can be one for all chips, or one for each chip - e.g. if I have 2 chips and if the first should default to 48 bits and the second to 50 the example would be "NanoFury NF2/rt" ("r" is 18+30=48 bits for the first chip, "t" is 20+30=50 bits for the second).
I realize that this has a limitation and will work for individual settings of up to 12 chips, but for anything with that many chips it may just be sufficient to have one default speed for all chips.

Also, there will be one more hardware change specific to the NF6 design - I need an inverse PWRON signal (where ON=0 and OFF=1 or input state) and I need the signal inverted so that I don't have to put an extra transistor (and the signal that is currently in the code is not inverse = ON=1 and OFF=0).
Currently there is no power control and all chips start working and consuming power the moment you plug them in the USB :)

And that will probably be the only notable difference in v0.2 of the board (plus moving some stuff around).


@simoncc - I know I'm repeating what's already said but could you please share your multi-chip version of the code?
I'll be messing with Luke's code tomorrow and want to try cgminer as well :)

Hi vs3,

1. About freq issue, your ideal is good. I will modify my code to solve this issue.
2. After modifying and finishing test simply, I put my code into public.
3. Regulator on NF2 seems supporting 3 chips and working well. I had made them working for +10hrs last week, and found the heat-sink was very hot. That's why I put a 5V 4-cm FAN on the heat-sink. After that, it's quite cool now.

simoncc

Download this tool to modify the product string of MCP2210 :  http://ww1.microchip.com/downloads/en/DeviceDoc/MCP2210Utility_v1.2.1.zip
Read, modify and configure... quite easy. If you use zadig to access MCP2210, you have to remove/uninstall it and change to windows default HID driver such that microchip can access MCP2210 correctly. I will put my source to github here tonight, 'cause that is at home, not in office.
[/quote]

I thought cgminer was suppose to work with the zadig drivers... excuse my ignorange im not very good at software.
As Taungeran said earlier in the thread, mcp2210 is limited by the hid drivers as they are designed for relatively low output of information, limiting the NFx to about 8 chips... as i found out.

Is this the final results?
Dual Chip?

https://www.youtube.com/watch?v=BxyzR_2IYG8

Thats BIFURY, same design but communication is done through another chip and through firmware.
NFx design comunicates with the chip aproximately directly...just an usb-spi converter.

NanoFury and RedFury..icefury...etc are not the same.

some more observations with a slightly better hub:

bfgminer version 3.10.0 - Started: [2014-04-07 02:56:32] - [  0 days 00:16:58]
Manage devices Pool management Settings Display options
Connected to stratum.mining.eligius.st diff 32 with stratum as user
Block: ...6f069900 #294632  Diff:6.12G (43.81Ph/s)  Started: [03:01:58]
ST:23  F:0  NB:4  AS:0  BW:[ 58/ 52 B/s]  E:162.41  I:  130uBTC/hr  BS:4.28k
7/21         | 41.73/41.54/43.68Gh/s | A:338 R:3+0(.71%) HW:208/2.0%
-------------------------------------------------------------------------------------------
NFY 0:       |  2.51/ 2.52/ 2.43Gh/s | A: 18 R:1+0(3.3%) HW:  3/.50%
NFY 1:       |  4.50/ 4.50/ 4.48Gh/s | A: 36 R:0+0(none) HW: 12/1.1%
NFY 2:       |  4.26/ 4.26/ 4.40Gh/s | A: 22 R:0+0(none) HW: 10/.95%
NFY 3:       |  4.24/ 4.25/ 4.18Gh/s | A: 43 R:0+0(none) HW:  2/.20%
NFY 4:       |  4.50/ 4.50/ 4.40Gh/s | A: 28 R:0+0(none) HW:  9/.86%
NFY 5:       | 11.12/11.24/12.13Gh/s | A:105 R:2+0(1.7%) HW:101/3.3%
NFY 6:       | 10.26/10.27/11.72Gh/s | A: 88 R:0+0(none) HW: 71/2.5%
-------------------------------------------------------------------------------------------
[2014-04-07 03:14:45] Accepted 02780fff NFY 6e Diff 103/32
[2014-04-07 03:14:48] Accepted 070813ac NFY 5f Diff 36/32
[2014-04-07 03:14:50] Accepted 0605a0b3 NFY 6a Diff 42/32
[2014-04-07 03:14:58] Accepted 019b50b7 NFY 0  Diff 159/32

In another test I tried two different heatsinks - one from the NF2 series (very high performing - probably the best I've seen so far) and another generic one. The generic one had about 8-10C higher temperature, and when pushed very hard (and no fan) the board got to 105-108C and I started seeing a lot of hardware errors.

I've relocated it vertically - again no fan, and that seems to be fine - temp is 76C as of this moment while it does:
 NFY 0:       | 10.37/10.22/10.89Gh/s | A:87 R:0+0(none) HW:75/5.0%

which looks like this:
https://i.imgur.com/MLj5gg4m.jpg (http://imgur.com/MLj5gg4)

I put my code at this : https://github.com/capricorner/cgminer (https://github.com/capricorner/cgminer)
About the number of chips and frequency in MCP2210 product string, that was added already. For example, "NanoFury NF2/E v0.5 by MLM"
stands for 2 chips by freq 54. the token in the product string is "NFx/y" or "NFx-y", x is for the number of chips ( 1~99 ), and y is for the number
of frequency ( A ~ F ==> A for 50, B for 51... E for 54, and F for 55 ).
I believe there must be some issues in the code and I will try to fix them ( I found several tonight...)

I hot 6 of that new bitfury miners. Can overclock them up to 6gh/s whit fan and 5gh/s no fan.
Can u guys get more hashing speed out of them? Is this the same chip as the nanofury?

Video of my bitfury miner:
https://www.youtube.com/watch?v=stbhrGpfq1w

The chip is the same, but the board is not (there is a dual chip NanoFury, the NanoFury II (https://bitcointalk.org/index.php?topic=526925) as well, you can compare the two)
For the device in your video, try this thread instead:
[SUPPORT THREAD] BFx2 Bitfury USB stick miner (https://bitcointalk.org/index.php?topic=524521)

Last two panels and a stencil left! $200 shipped!
(I don't have enough from all components though - so no components, even though I'll probably throw in a few of the more important ones)

Has anyone got Cgminer to work with more than one chip?
I cant figure out how to make the code simoncc provided.
If someone could post some instructions... i would be very happy :)

You can download cgminer from official Con's git, then replace all files related to mcp2210 and bitfury from simoncc git then compile with --enable-bitfury.

I think that should work ...  ::)

One day someone there asked me one question "How to chain NF2s together and access them by one USB interface?"

I had a look at my NF2 board, and found out how to do that. Let me show you the picture:


WARNING HERE:  I HAVE TO SAY ONE THING THAT I DO NOT REALIZE NF2 CHAIN HERE BECAUSE I HAVE ONLY ONE
BF2. SO YOU CAN TAKE THIS FOR REFERENCE ONLY. NO GUARANTEE AGAINST DESTROYING DEVICES...https://dl.dropboxusercontent.com/u/23682587/nf2-chain.jpg

1. unsolder the parts in orange rectanges on 2nd NF2. ( 1st NF2 does not need any modified, only 4 flying wires later...)
2. there are 4 red circle on 1st NF2, and there are 4 blue circle on 2nd NF2. Connect red 1 to blue 1, R2  to B2, R3 to B3,
    and R4  to B4  by flying wires. Note: make sure wires 1 2 3 are the same length.
3. if you want to chain the 3rd NF2, unsolder the parts on 3rd NF2, and connect Red circles on 2nd NF2 to Blue circles on 3rd NF2,
   and so on...
4. if you plug these chained NF2 to the same USB hub, GND plan should be the same. No GND wire should jump each other.
    Yellow circle in the picture is GND point I use.

AGAIN HERE:
WARNING HERE:  I HAVE TO SAY ONE THING THAT I DO NOT REALIZE NF2 CHAIN HERE BECAUSE I HAVE ONLY ONE
BF2. SO YOU CAN TAKE THIS FOR REFERENCE ONLY. NO GUARANTEE AGAINST DESTROYING DEVICES...


You do not populate the parts in oragne rectangles on the rest 7 NF2 boards if you
have DIY kit from vs3... and you can save a little bit of money by this.

First: simoncc - nicely done! :-) And thanks for sharing!



Onto the NF6 story - v0.2 of the design is now confirmed with both gen1 and gen2 chips!
https://i.imgur.com/30huWA8m.jpg (http://imgur.com/30huWA8)

Here is how things look like with Gen1 chips:

This is with a crappy power supply (4.6V) and 54 bits, somewhat inferior heatsinks, a tiny bit of airflow, temp 46C and 65C (on the second one heatsink faces downward and that's chips temperature). I haven't checked exact current usage but from indirect measurements it is around 1.3A (which is a bit higher due to the lower voltage, or around 6W total)

Or to summarize it - with Gen1 chips 9GH would be around the norm, and over 10GH with a bit better hub (that provides actually 5V). On a slightly over-voltaged hub (5.25V) it should be able to hit even 12GH and above.


And the screenshot below is with a mix of chips: one NF6 with Gen1 chips and one with Gen2 chips:

Same hub, power usage a bit higher (and voltage down to 4.3V), speed 54 bits for both, a bit better heatsinks (the Gen2 with the best I have around), temperatures: 50C on the gen2 and 73 on the gen1 (which is also facing down so not the most efficient position anyways).

Judging by the low error rate I can probably squeeze out another 0.5-1GH out of each.

And I'm actually rather surprised that I see a stable 11.5+GH with such a low voltage on the Gen2 chips. On a better hub it will certainly exceed 12GH and with a slightly overvoltaged one (5.25V) it will very likely exceed even 15-18GH.

The Gen1 chips in this screenshot are from the older series with datecode 1337.

There are also differences between the various production dates in the Gen1 chips - the first 2 boards have chips with date code 1404 (which were from the last series - I got a few of them forwarded from bitcoinvalet's last group buy) and I am a bit surprised that they perform slightly worse than the older chips that I had around with datecode 1337 (which were leftovers from the NF1 series).

Bitfury Gen2 chips were from the first samples available for purchase from punin (datecode KCTMS):
https://i.imgur.com/DBgV4yfs.jpg?1 (http://imgur.com/DBgV4yf).

Nice, well done:-)

Thank you, Sir! :-)

And to be honest from the few people that deserve a lot of credit for NF6' existence there are two that stand out:
you - for proving that it is actually doable, and
KNK - for confirming the minute detail that makes it all possible! (which also saved me a dozen chips:)

So - Thank You! :)

Should our paths meet - drinks are on me!

Praise should go to Valery (bitfury), he
came up with this brilliant 'string' idea:-)

And while we set up the NF6 production line and clarify all details ... I got bored and tweaked the board a bit (removing a few extra resistors and some other minor optimizations) .. and figured - it would be interesting to see how did it evolve :-)

So - here it is - version 1 vs version 2:
https://i.imgur.com/rkToD6Am.jpg (http://imgur.com/rkToD6A)

then, version 2 vs 3:
https://i.imgur.com/edVF1Yxm.jpg (http://imgur.com/edVF1Yx)

or all changes between 1 and 3:
https://i.imgur.com/wuPq9Epm.png (http://imgur.com/wuPq9Ep)

I take it that you would do a diy kit for the NF6?

I register my interest in a possibly upcoming NF6 DIY kit.
 :)

I have some questions about the NF1 (0.7). (yellowjacket)

According to the parts list, CF1.8-1 is a 100nF capacitor.

https://i.imgur.com/7dx862b.png


But it looks nothing like the other 100nF capacitors on the schematics.

https://i.imgur.com/bgr3zzv.png

Also, what are those small gray circles on the left.
And what is the purpose of the contacts right of the bitfury chip, no components are going to be placed there.

thanks.

The footprint is supposed to be 0603 and at various points I was experimenting also with 0805 - it is possible that I may have forgotten to switch the package back to 0603. Either way - the two are very close and 0603 will work with either one.

Also, the component label isn't perfectly placed and that might be adding to the confusion.


gray circles are vias. I think I left them visible on some of the assembly drawing pictures so that you can identify a bit more easily which component and track goes where.



Those are test pins (or pads). They're there for testing purposes. And they're also very handy if you want to add more chips to the chain as they expose the OUTMOSI/OUTMISO/OUTSCK pins (as well as OUTCLK if you want to check chip's internal oscillator frequency with some external tools).

Parts list might be mistaken - according to the schematic (https://github.com/nanofury/NanoFury/blob/master/NF1/kicad/NF1-MCP-V07/NF1-MCP2210-v07-SCH.pdf), it should be a 22uF capacitor (same as e.g C21 close to it).

They're vias - small plated through (meaning there's copper around the inside) holes that connect the copper on the top layer to copper on the bottom layer.  You usually place a bunch of these around - you can see more near C1, by the NanoFury text, etc.

Those are P1 through P6 P7 in the schematic - they're just test points, e.g. for checking performance and behavior with an oscilloscope or logical analyzer

Edit: What he said! :P

Thanks guys,


So suppose I had a 100nF cap. Could I place that in CF1.8 instead of a 22uF one?(or the other way around)
That is a big difference, won't you totally blow that one up.


I have one more question, in the red square is a very big patch of conduction material(with the holes), should you fully solder that or not,
since the chip already has connector pins?

https://i.imgur.com/MEL6VUl.jpg

I'll leave that one for vs3, given the conflict between schematic and parts list :)

Those holes are for thermal transport (from the chip, through the board, to the other side, and to the heat spreader/heat sink.  You can fill them with thermally conductive material, that should actually improve thermal transport a little bit.  Filled and plated over boards are relatively expensive, so you wouldn't see that happening very often (might see it in e.g. BGA designs where you've got via-in-pad nastiness).

Both CF8 and CF1.8-1 are in parallel, so it doesn't really matter which goes where. They now have the same pads, so you can put either one at either location.
From a practical point you may want to put the 100nF (CF1.8-1) closer to the bitfury chip as the capacitor is a bit smaller and easier to place there. And as for CF8 - since the bitfury chip draws very little current you can even experiment by not placing it at all. I've experimented with various values - 1/4.7/10/22uF and things are fine with any one of them.
Just make sure to put at least one of the two capacitors - the 1.8V regulator may start misbehaving if both are missing.

Aside from that - I'm not sure where you got the BOM from - on my schematic C8 is 10uF. (And CF1.8-1 is incorrectly 0805 on both the schematic and PCB - it should've been 0603).


As for the big pad - TheRealSteve already mostly covered it:


Most of the vias would get at least partially filled with solder during the process - which causes also some solder to seep to the back side resulting in a bit uneven surface there - keep it in mind when selecting the thermal pad!

Below is an image of how things would look when you overlay the solder mask and solder paste:
https://i.imgur.com/nj7556Wm.png (http://imgur.com/nj7556W)

Out of curiosity - where did you get that board from?

From my recollection those boards were black ...

http://shop.bitgtr.com/yellowjacket-pcb-diy-educational-edition/

He started selling some bare boards :)

Yes that is where I got the boards and the BOM from.

Ok, the soldering took some time.
The main problem was that the solder paste didn't wanna let go of the tip of the syringe.
Therefore it was hard to precisely solder everything.

Here are some pictures after heating.
http://imgur.com/a/5jSxL (http://imgur.com/a/5jSxL)

The usb plug is not soldered yet. but that is not such a problem.
The problem is the bitfury chip, some bridges have formed.
What do you think is best, use a soldering iron to remove the bridges?

Solder wick, or re-do (chipquick is some nice stuff for that).

http://www.theledart.com/blog/archives/1301

Thanks, will add that to my bookmarks.
Nanofury bookmarks now totalling over 100 haha. :)( all those tutorial videos lol)

You should definitely bookmark the source of the video - http://curiousinventor.com/ - there's plenty more (de)soldering tutorials in there, and they're a Bitcoin enthusiast as well.

+1 :-)

Solder wick is your best friend!
And an old fashion toaster oven might be your second best one :)


Also take a look at this video:
https://www.youtube.com/watch?v=c_Qt5CtUlqY
(the interesting part starts after 3:30)

Also, I've been using SMT stencils for majority of my work and that's how I've been making my prototypes.
Here is how I did mine: http://imgur.com/a/5hXU4

Solder wick is your best friend!
--
Cohesion is your better friend;)

I finished one nanofury.
But both my pc and rspi are not seeing the device. Also the led is not on while plugged in.
Is this a driver issue, or probably a malfunctioning component?(or something else).

Start it by one step at a time.
1. Plug it into the PC (I'm presuming running Windows). It should beep and say a new HID device has been found. You should also be able to see the HID device in Device Manager. (if that doesn't happen then most likely your MCP2210 isn't working - check voltages, shorts, oscillations on the crystal pin, etc)

2. Run the nf1_init program. (source: https://github.com/nanofury/NanoFury_Init ) There are two things you're looking to do with it:
2a - it should recognize the device and print the serial number and Product String. The Product String is important.
2b - run the program again with fixID parameter and it will change the ProductString to "NanoFury NF1 xxxx" - the NanoFury one is mandatory for bfgminer to recognize the device. NF1 is optional and is used by other mining software, so make sure to put both (if you're customizing the string). A few pages ago there were links to other programs that can be used to change the string.

Also, at that step the nf1_init program while testing the device will make the LED blink.

If you make it that far - then your device should be good to go.

3. Run BFGMINER and experiment with the osc6 speed. If you haven't put any heatsink then don't push it way beyond 40-45. If you have an infrared thermometer you can check temperature - try to keep it below 75C. If it gets beyond 85C it will start misbehaving and you'll start seeing tons of errors.

If BFGMINER recognizes the device and can communicate with the MCP2210 but prints lots of init errors - that's an indication that there is something wrong with the communication between it and the bitfury chip. Check the resistors between the two, check for shorts, check voltages, etc

You can use the nf1_init program with the test parameter to get the device to turn on the various voltages and measure what's going on.


Good luck! :)

Well it is not in devices in windows,
But I Noticed a difference in your work and mine.
I did not solder these pads since there is not going to be placed a component anyway.
I don't see a point in soldering those, the are already conducting.

Notice the red circles:
https://i.imgur.com/xRZwUJZ.png

that is correct - there are two sets of pads for the MCP2210 depending on which package you get. You should only use one of them and leave the other untouched.

That's in the v0.7 change - basically, the MCP2210 USB bridge was difficult to source in the package used in v0.6 and earlier, so he added a footprint for an alternative package type.  If you look closely at a board/design file, you'll see that the pads just connect straight to the alternate one.

Edit: What he said - again.
I swear this used to tell you somebody else posted while you were editing - what happened to that?

Allright, I thought that was the case.
It is compatible with 2 types of MCP2210 sizes.

The MCP2210 is probably the issue, I will try and fix it tomorrow I think.
As you may have noticed I am a total noob at this and it is mostly trial and error for me haha.
Placing solder paste on the board is a total disaster without a stencil.
The paste is not sticking to the board, so I looked it up on the internet and my solder paste is probably dried out.
I also read that you can dilute it with some flux, maybe give that a try.
( flux are 2 different things apparently, the watery stuff and the fat yellow goo)

where are you getting the stencil from? and they should be asking for a gerber file of the top cream layer

Hi vs3,

speaking of your toster oven: Do you just use an old oven or did you modify the oven too?
I'm building one for my self and didn't see on your pictures any thermocouples or something like this, how do you control the temperature?

Thank you and best regards,
max

It happened to be on sale (for $20) so it was a brand new one :)
No mods whatsoever.

It has a temperature dial - which is way off but it worked for me.

- preheat: dial at 150C (which is closer to 180C), leave it there for 5min
- melt: dial 200-220C (you'll have to experiment and find out at what temp the solder starts melting - add 20C on top of that)
- measure 1min from the moment it starts melting, then unplug oven from outlet, open door and let it cool

Hmm, getting kind of annoying.

I Didn't find anything that was wrong so I soldered a whole new nanofury.
It looks identical to the old one but I am seeing some improvements in soldering. No bridges as far as I can see ;D

But unfortunattely the miner is not detected by windows/raspberry.
I am hearing no sound when plugging it in, and led is not on.( I don't know if the led should be on).

Led is not normally on. Until software is running

I've added support for NF2 and NF6 into cgminer git master branch. I'm not getting much love out of the NF6 just yet but they may well need so much power that even a regular USB3 port is not enough. The NF2s are behaving very nicely. This is at --nfu-bits 54:

NFU??

Is that a question, statement, exasperation, disgust, or what?

Heh :) Fun with abbreviations!

Con - I suspect he wasn't sure what the "--nfu-bits 54" part meant - which is the osc6 bits, and the "nfu..." prefix is what is used in the source code - e.g. as in this line:
mcp->value.pin[NFU_PIN_PWR_EN0] = MCP2210_GPIO_PIN_LOW;

you would see also this in the source code:

/* Identify number of chips, and use it in device name if it can fit
* into 3 chars, otherwise use generic NFU name. */
val = sscanf(bitfury->usbdev->prod_string, "NanoFury NF%u ", &info->chips);

I meant the identity was changed. that's great. not for NF1 only, but for MCP2210 based devices... great thing

This board hashes at 4.4-4.6 GHs with 5W consumption?

I think it is unlikely, but suppose both my MCP2210's are faulty, how could I measure this?

In what file? I get 0 results when try to find in cgminer-4.3.3 repository.

https://github.com/ckolivas/cgminer/search?q=generic+NFU+name&ref=cmdform (https://github.com/ckolivas/cgminer/search?q=generic+NFU+name&ref=cmdform)

Looks like those are different versions. I downloded 4.3.3 bz2 package from here: http://ck.kolivas.org/apps/cgminer/ and in bitfury driver.c at line 438 I have this:

EDIT: no occurence of "generic NFU name" in any file... Searched by notepad++ in all files...

It's just updated hours ago. I guess it's not ready for next rev. package. You should go to github to get the latest source.

Didn't notice that sentence before. THX

In the last few days I build 2 NF1's.
Unfortunattely they are not recognized by windows/minepeon.
They don't beep when plugged in, and don't show up in devicemanager.

So I double checked, triple checked the board and there are no bridges, misplaced polar components etc...
It is identical to the schematic I used and pictures I find on the internet.

I think it is unlikely that both MCP2210's are not working as they should.

Also when building I wore a anti-static wrist band just in case, so I am assuming nothing is fryed by ESD in the process.

The process of reflowing may not be ideal but it was alright I think, I used a baking device that looks something like this:
http://s.s-bol.com/imgbase0/imagebase/large/FC/9/3/0/8/9000000011158039.jpg

The max temperature is 215 C, which apparently is enough to melt the tin. (yes the joints looks shiny)

I have no idea why the miners are not working, so comments/criticism are very welcome.

Don't you have to program them or something?

https://github.com/nanofury/NanoFury_Init

This is a windows based c++ program that uses ms visual studio to compile. This should program the nf1 with the proper naming convention needed to get cgminer and bfgminer to recognize them.

Did the parts come in a kit or did you have to buy them separately through a distributor

By what you said so far it seems the MCP2210 is not talking to the PC. Can you take some close-up pictures from your boards?


Start by checking one thing at a time.

First - connect 5V from external source - just solder 2 wires onto the USB connector pins. (anything 3.8V-5.5V will work, but don't push it beyond 6V!)
- Check the input and output voltage at the 3.3V regulator. Between the USB port and regulator there is a ferrite bead (should be equivalent of zero ohms for this test).
If both 5V and 3.3V are good - move to the next step.
- Check for oscillations on the 2 pins where the crystal is. You will need an oscilloscope for that. (note - one of the pins is more sensitive than the other, and by touching with the probe you will probably kill the oscillator - that's fine, as soon as you lift the probe it should start oscillations again and they should be good 12MHz ones on the other pin)

For the MCP2210 chip to be recognized by the PC all you need is the 3.3V and that crystal. If you have another board you can assemble just that part of the board - the ferrite bead, USb connector, 3.3V regulator, at least one of the capacitors between 5V and GND and at least one of the ones between 3.3V and GND and the crystal along with the MCP2210.

Check those and see where that leads you.

Is the crystal orientation sensitive?

No. But MCP2210 chip is :)

Pictures of the board:
http://imgur.com/a/zYquJ#2 (http://imgur.com/a/zYquJ#2)
http://imgur.com/FT7KJFB (http://imgur.com/FT7KJFB)(solder 3 capacitors under microchip?)

I will check into those voltages tomorrow.
As power supply, a usb port supports up to 5V, that should be enough.
Unless you want the full 5V potential between ground and the '5V' connection.

I don't have an oscilloscope unfortunately to check the crystal.

This soldering is terrible ! It's clearly visible that there's no solder paste on some pins of MCP and there's brigdes on bitfury chip. No wonder it don't want to work.

Bridges only matter on one edge of the chip