Bitcoin Forum

We are glad to announce ASICMiner's 4th generation chip. From the physical testing data
of our verification sample, BE300S, the silicon results matches the simulation results faithfully,
and even outperforms our simulation in some cases. BE300S achieved the lowest energy
consumption per gigahash in existing market, getting the joule/gh ratio down below 0.2.

Furthermore, our string-based power solution eliminates almost all power losses and almost
all electric components on the board. Therefore on-wall power consumption would be very close
to (on-chip power/PSU efficiency), and the overall system cost is squeezed to its minimum.

Name: BE300S (Sample)

Technology: TSMC 28nm HPC

Package: FCLGA 5mm x 5mm

On chip efficiency (first board):
    3.0GH/s | 0.1872W/G
    4.8GH/s | 0.2275W/G
    5.2GH/s | 0.2469W/G

On board efficiency (average of 4 boards):
    2.8GH/s | 0.1961W/G
    3.2GH/s | 0.2026W/G
    3.6GH/s | 0.2095W/G
    4.0GH/s | 0.2145W/G
    4.8GH/s | 0.2204W/G
    5.2GH/s | 0.2257W/G
    5.6GH/s | 0.2314W/G
    6.4GH/s | 0.2363W/G
    6.8GH/s | 0.2439W/G
    7.2GH/s | 0.2495W/G

24 chip board efficiency:
    118.56GH/s | 0.26W/G
    155.04GH/s | 0.27W/G
    176.64GH/s | 0.31W/G
    273.60GH/s | 0.38W/G

With more accurate control on the variance between chips we could in principle gain less power consumption in advance.

We are still testing more voltage-frequency combinations, as well as more chips.
But the test results are stable and solid so far.

After the thorough testing of single chip boards, we are going to test boards with chained chips.

Single chip testing board picture:
http://i61.tinypic.com/28k6m51.jpg

Testing going on:
http://i61.tinypic.com/npp9o1.jpg

24 chips testing board picture:
http://i58.tinypic.com/28rgnxl.jpg

http://i60.tinypic.com/14v5h7k.jpg

More data and pictures are to be updated.

Awesome! They're finally in!  ;D

Looks very promising. Looking forward to seeing what hardware will be offered from these chips. Will follow this one closely.  8)

that's a step up. well done FC.

if i'm adding up right - that chained board (filled with chips) should be ~135GH/s @ ~32w?
Network is going to go mental. I can't wait to see what the competition brings to the fore.

so asic miner is promising a .3 watt miner at the wall socket.

if you could build one that works and does not burn up caps .  I will buy it.  but after my tube and then my prisma troubles I will wait for others to buy them first.

would be nice to have a .3 watt miner. an evga 1300 g2 could power a 3000gh unit easy.


Right now I can run 2444gh on an evga 1600g2  ..  I can do this by under volt of the sp20.  2 units will go to about 625 watts plus 625 watts = 1350 watts at the kwatt meter plug

So this is about the best a home miner can do today on 1x 120 volt  15 amp circuit.  need a 1500 watt psu to do it.


If you can do .3 watts a gh it may be possible to do a 4th machine  for 1200 watts .  that would be a nice improvement.. it would allow the evga 1300 to be used as that psu can do 1200 watts non-stop.

lets see if you do it.

There are still several ifs for the chained board to achieve this result. Generally we are
positive towards it.

Anyway we will report more data whenever we have new.

In that case, i wish you well with your tests and look forwards to seeing the results.

There are two reasons contributing to the current RMA problem of BE200-based chain design. The first is firmware problem, which we solved for Prisma 2.0 and are solving for Prisma 1.0 and 1.1. The second is that BE200's load constantly changes when switching task, introducing some big ripples in the whole chain and making the board unstable. It is also largely solved for Prisma 2.0, and is going to be totally eliminated by BE300 since its load never changes.

Also for chained design the output voltage of PSU, instead of software controlled DC/DC, determines your frequency and efficiency. For example, using a 12V-output PSU results in a higher frequency but higher W/G compared to a 11.5V-output PSU.

Only one question. Samples?
You didn't provide any previously. That ended with two designs only, one foulty as hell. Maybe this time?

beggars everywhere  ::) i don't want a sample, i just want to see if they make specs.

We will see if we can allocate some from the sample batch. The problem with MPW is that we only got
100 chips in hand and we need at least a large part of them for chip-wise variance testing.

i am in

Wow. Looking forward to testing these out first hand.

Grats FC ! Nice specs !

Any date estimates for miners with these chips?

i'd speculate, spring 2015.

So what are the chances of the be200 users getting compensation and good boards before the be300 comes out?

I'm not saying that you should provide samples now (but it would be great)

Who said that I want it free? I wanted to design a miner with a BE200 and pay premium price for a small batch of sample chips (lets say 10pcs). But no response from AM. How it ended to AM we already know. Only 3 designs and troubles with selling those chips...

fair enough, apologies for my presumptions.

FC earlier stated that mass production of the BE300 chips will be finished in Feb 2015. Given that everything seems to be well on track (actually a week ahead of schedule!) we might see miners in April 2015, or even as early as March 2015!

FC, very good news! It's great to see some pictures of the board and the testing, as well. It makes me feel being "more part of the whole thing" :) Keep us posted on further results, please!

nice chips, get frying them mr cat.

Great chips, 0.3 W/GH. Looking forward final result  :)

I believe i cat fry :D

Nice job. Waitng for prototype of final miner. Hoping you will show it faster then Bitamin and Spoondolies.

Interested in designing a miner board with these
Any info for samples?

Finally! This is great  . . .

Looking forward to seeing what comes of these.

I hope that AM sticks to just selling the chips and not making miners again. Amazing how bad the Prisma experience was...

Do you have enough test units to send one so I can try to have BFGMiner support at time of production shipping?
How about specs?

I was hoping they were gonna stick with pin-compatibility with BE200 chips like was mentioned when they open-sourced stuff earlier this year.

Additionally, a small suggestion for strung miners - fuse the power rails per bank. If something fails short and parts start popping off, it'll blow the fuse instead of catching the board on fire and then you're down a bit of capacity but still operating.

^^^ single chip board the man

Or at least drop a publicly available datasheet with pinout and data protocol....plz

a thousand times this

Yes!  Please provide samples to Luke-Jr and ckolivas so BFGMiner and CGMiner are supported at launch.

And so it begins! It will be interesting to see the formfactor devices take as GH / chip [at these speeds] is much lower than BE200 which was ~7GH. Power per chip is also down so the hashing and power density of BE300 will be quite significantly lower. This is of course that AM doesn't throw out W/GH for $/GH.

We achieved 7.2GH/s per chip already, increasing W/G by about 10% compared to 4.8GH/s per chip.

There's a large room for $/GH optimization for mass production, and even if we do not do it, the $/G on 7.2GH/s setting is still decent.

So once you get everyone's money will you and your crew go silent for weeks\months on end again?

Yeah sell chips and make this opensource THATS the only way hobbiest mining will survive. Groups of people can build their own miners =)
Make the chips cheap too of course.

Hmm... LGA vs QFN..  does that mean the heat goes out the top of the chip now?
 

Negative ghost rider. Lga is similar to qfn in that they both eject heat out the bottom

This isn't the "same" LGA as like the Intel cpu sockets

Fantastic news Friedcat!   It looks like a great chip, on time and on spec.  

Now, as a shareholder I have one request:  please hire the best sales team you can find to promote the BE300, and capitalize on this engineering success!  

ASICMiner has the best chip, it should have the best *everything*.  Including the best bottom line.  ;)

Congratulations and keep up the good work!

I'm surprised folks are so excited about this considering all the problems with under-performing BE200 products. I'm certainly glad I never purchased a Prisma and had to deal with all of those headaches! Also, Asicminer overpromised and underdelivered big time with the BE200.

Given their track record, I don't trust Asicminer enough to order a BE300S product right now. That could change in the future, but only after others confirm that the finished miners are actually a quality product.

i am exceedingly cautious these days. my rule of principle is that I never buy batch1 of anything.
and NEVER pre-order.
i'll wait to see what happens.

I bought a tube right when they came out and it has worked perfectly. Never got a prisma due to all the delays.

It should be interesting to see what happens in the next couple of months.

From pictures you can see that heatsink is under a PCB. So propably it's old way to dissipate heat thru bottom...

Very good news Friedcat  8)

Keep up the good work.

Will these chips be packaged in a miner for standard users or the development is already pointed on a datacenter product like SP did? (maybe it's too early to ask about it)

Indeed something to look forward too ^^

Keep up the good work Friedcat and thanks for the update.

@Friedcat: what is the qfn next to the pin header that's labeled PMS01?

that would be a simple 8/16 bit micro-controller, iirc similar to what they used in the usb block erupter, gets the uart data off the chip then passes it over to ttl..

i could be wrong

now i do have a question about these chip, they are not a straight drop in for the BE200's yes?

re-read, "FCLGA 5mm x 5mm" thats smaller then the 8mmx8mm QFN package

Probably March for mass shipping

It was a strategic mistake to switch to chips only, that's why AM started producing be200 based devices (tubes/prismas).  You should expect miners and chips to be available with be300

+1. nobody knows better how to design a PCB than the guys who design the chips. Relying on 3rd party groups (technobit for example) to invest in prototype boards for multiple chip types simultaneously is less likely to get an ideal resulting design - ie the series-based design for voltage control, exactly what bitfury managed to do

You would think that but then you run into a company like AM that produced the prisma. Multiple people including myself have had one spontaneously burst into flame. Then AM has promised compensation they have not delivered. For myself they promised a new prisma for compensation on the burned up boards that I paid to ship back to Hong Kong from the US. Instead they shipped replacement boards leaving me to repair the damaged heatsink and pay for shipping from the distributor they shipped the new boards to. Then when they started acknowledging the other boards that have caught on fire or just don't work they promised to pay for return shipping to Hong Kong and to compensate for downtime. Now if you look in the thread where they posted that, there are no responses from AM. emails and PM's go unanswered and aside from a few boards they sent which didn't cover all the damaged ones there is no compensation either.

AM's track record:

prisma ran hot a lot of them burned  and popped caps.---------- a little bit slow to deliver

tube ran hot a few of them popped caps.       controller sucked only one pool no auto rollover.-----also a little bit slow to deliver

cube ran hot a few of them smoked some psu's

blades ran hot but these worked okay if you cooled them

 the usb sticks ran hot but were decent performers. I would say they were an important pice of gear for many miners to join the game.

Canary
Crazy
Silent Sonic Boom  sold me hundreds of them.  all good sellers.



Am's support and replacement refunds etc. --------------meh not terrible.

I will wait until two or three brave souls  run them  and let us know how they work.

It's our own string-based power management chip for solving the Prisma-related problems.
The single-chip board is not a string, but we used many of its functionalities anyway to simplify
the PCB design.

One of the problem with Prisma is that it involves too many separate components that might
go wrong. We concentrated them into one chip and only need to test one chip.

The revised Prismas (2.0) are based on this chip. From our large-scale stress testing, we
are confident that the instability and burning problems of original Prisma (1.0 and 1.1) are eliminated.

Look at AM's track record compared to:

BFL
KnC
Black Arrow
Cointerra
Hashfast

(and about the same compared to)
Bitmain
BitFury

Okay bitmain has less fire issues and burnouts then AM.

If I ranked gear /companies  bitmain + spondoolies are 1 ,2 or 2, 1   and AM is in third place.

This is why I would still buy from them which I said in my post.

Well I purchased both a Tube and Prisma and both were shipped with issues.. Why should I trust you when you say it is all fixed now? My only experiences with buying AM gear have both ended in me taking a loss. You guys have 0 support to help us too. I must say it has been very frustrating buying your gear to say the least.

And here we go again. Posting about how things will be better but still ignoring the existing issues and promises. Why are you still not responding to emails and PM's? Are you just hoping all the prisma users will forget about the compensation or the lack of communication and just fork over more money to buy new gear and hope that it will work better than what we already paid for?

What prisma? what fire? huh? ive payed 100btc for many of these chips so i can swim in them :D

jk, you are right though, i have seen this happen too, companies ignoring their users/consumers to move onto other things, that's why they all go register up as a limited liability company (LLC).. i know little about the LLC legal form, i've known nothing like this in Australia.. (or i could be ignorant of this?)

Lemme just clarify, i have no idea where ASICMiner sit in the business model, what laws they are governed by, and what they care to do as a business.
Just stating others are also doing the "ignore customer issues, make new stuff" game

10Gh USB miner? :D Maybe?

Maybe this could be a return of the super cheap USB miners.

https://bitcointalk.org/index.php?topic=889147.0

Friedcat is an honorable person.

I think returning that btc from the transaction error a while back cemented that.   

The component overhead of USB miners is still a problem.

1 MCU, 1 DCDC, 1 USB-to-UART chip, for each 2 chips, is so much cost to take.

If we make a 8 to 9 chip USB miner (about 40G), DCDC cost can be saved. But USB2.0 does
not have enough watt capability to power it.

Indeed that was a ridiculously large mining fee for people that recall that event.
It's good on Friedcat to do this as well.

USB 3.0 is being used more prominently nowadays, could it be useful?

Yes. But in low power mode we think it's possible to make the system without heatsinks
in either side.

The heatsinks in bottom just mainly serves a lifting purpose and meanwhile maybe dissipate
some heat.

We tried hard and we can't. The overhead of QFN-based package is overwhelming. And this time our die
size is much less.

Good work friedcat.
What time frame do you expect to release ready miners?
And what form factor it would be? Tubes?

Is producing a small, low cost USB 3.0 miner actually being considered? Sounds interesting but would need at least ten of them to be productive.

USB  Bitcoinlottery devices aka solominers would be awesome :D

As an aside: both the hex•fury (https://bitcointalk.org/index.php?topic=523063) and the NanoFury 6 (https://bitcointalk.org/index.php?topic=665927) already do 10Gh/s+.
I'd certainly welcome another addition, though :)

One that did 40GH to 50GH at a reasonable cost and power level would be nice!  ::)

One could drop the MCU and use a USB-to-GPIO (or similar, depending on chip interface) to offload work onto the host....

I think something small like Rock Miner did was pretty cool (the 100GH unit). Maybe make a pod like Bitmain...something that does not require a computer PSU to run.

As long as someone remembers to program a manufacturer and model identifier!

To not require an external PSU means they have to put in an internal PSU.
Which means additional certifications are needed, and can sometimes delay customs/shipment...

(that being said, I agree it's nice to have)

And unique serial number

I would be very interested in purchasing a sample as well, if you do end up selling some. Chip docs would be appreciated in the mean time though.

Just did a quick calc and a rig in the style like the bfsb 48card bitfury rigs could pack 3.5TH into ~1Kw

Excited

I think what he's saying is, not necessarily an internal PSU but also not a separate external high-current PSU. Probably still an external supply, but something more like a brick. A 200GH VRM-less system would pull about 70% load on a 12V 5A brick and could probably run pretty quiet if built right.

We'll likely see retail products sold closer to .4.

This is great. I will be following this closely. Congrats on the preliminary results and I would also be interested in getting my hands on a couple samples.

Good Luck!

Any word on low volume sales?  It'd be nice to see a decent chip available in less than 10,000-piece quantities...

Very exciting, I'll be following this closely as well.

MrTeal, Any chance you would call the resulting product the "Cayenne", or would it be more of a "Serrano"?  Either way, sign me up - I am on board if you end up using these chips to create the successor to the Chili and Habanero.

E.g. Like bsfb when they were selling Gen2 bitfury chips?

The Scotch Bonnet  ;D

Yep, good data (especially on the function of the PMIC) would be pretty nice, at least to get a rough idea of feasibility.

Oh yeah, it's mentioned that the hashrate is proportional to the core voltage. I'm guessing that's got an internal voltage-controlled PLL scaler? It'd be good to know the formula for that, or put some Vcore data on your first-post W/GH chart that could be extrapolated. It'd be a good reference to get people thinking.

I would really like to see the core voltage at each level. 
I'd like datasheets on both chips too, but even knowing core voltage vs. GH would be super useful.

--
novak

5.2GH/s | 0.2469W/G@0.55v sound not so good, and the cost is high...
bitmain's 3rd Gen chip have achieved this goal @0.6v and  the miners in mass production, i think FC is late at this round...

Sorry, this just smack's of FUD. Can you post a link to the Bitmain gen3 chip, it doing <0.25J/GH @ 0.6V, or someplace where FC stated what the price of these chips is going to be?

just some information from friends， I believe its truth. bitmain will sell S5 miner with gen3 chip in the next few days，wait to see...

The PMS01 IC provides all the functionalities needed except the mosfet for chaining the chip into a string.

For the one-chip board, we still choose to use it instead of a bunch of buffers and LDOs for board simplicity.

The problem is that there may not be widely available high efficiency power bricks.

Competing with 70% efficiency bricks agains professional miners with 92% efficiency PSUs is hard.

If you want to see the numbers of home miners continue to grow then you must admit it will be from newcomers to BTC. An introductory sub-$100 miner is still going to be the best proponent of that. SO, would you like to watch Bitmain and others develop loyal customer base from providing an introductory device? Or do you gamble on breaking that relationship/loyalty between them and get the business from their pockets when they mature into highrollers.

look into, i think Delta's, laptop/desktop external PSUs, they hit the 87-90% but they are a higher voltage DC.
There is also nothing wrong with using a PCIe power adaptor off a desktop internal PSU

There is if the aim is a starter device. To us the idea of paperclip jumping an ATX PSU is second nature, to newbies it's daunting.

I've been around personal computers since they were invented (Imsai and Altair). The idea of jumpering the PSU with a paperclip was daunting to me. The first time I did it I half expected the PSU to blow up. LOL.  :o

Coming up on 24 hours since I sent a email to AM and just like the other countless times I have sent them emails there has been no reply.

For all of you praising AM, this is the third time compensation has been promised. The deal looks great and they have managed to make many of you think they are a great company standing behind their products and taking care of their customers but there is no follow through. Just a bunch of empty promises.

Perhaps domestic drivers for LEDs would be a good choice - if they produce 12V and don't try to drive a constant current.

Do you exactly expect them to file and respond to 1000 emails in a day? Its just PB being bombarded with emails, give him a chance to work through them. I received a response late last night.

So, all the level shifting IO lines and clocking and such? Handy. I'll have to take a closer look at a Prisma board to see exactly what the FET would be doing (I have ideas). I am looking forward to the release of documentation.

Adding the core voltages to the GH/W table would still be quite nice. That would help people get a jump on figuring out designs for these chips.

So, how chained chips board performs?

Yes. The LDO output for IO and PLL are included as well.

The FETs help balance the voltage. For BE300 we may not include them, but we need experiment to see if it's true.

No data yet. But soon.

Will do the study. Very interesting.

12V is also not a must. We only need to adjust the chain length.

All mining devices try to draw constant current anyway, so if they try to drive constant current it's also OK.

That's what I figured, the FET would act as a dummy load shunting extra current when the chips were underloaded in order to keep the string powered properly. Still like to see a hashrate/voltage chart or formula.

Ignore those things, they have exposed terminals and are NOT user friendly. I don't know why they're being discussed for an introductory consumer device.

perhaps looking into 48V power supplies wouldnt be a bad idea? Effectively driving 4x12V chains in series. I imagine the lower currents in the PSU and lower change in voltage may give good efficiency and perhaps lower cost.

for smaller miners though, 12V pcie connection is best. hobby mining isnt really meant for anything under 300W anyways, and there are some good, cheap 450W power supplies available in most countries

There's no point chasing marginal savings if we're going to pay out of our asses for power infrastructure that doesn't exist. That is, in regards to scale, maturity, availability and cost.

Exposed terminals - shock horror - and why the shouting ?

You should lie low after having recommended duff 12V supplies.  We're talking reliability here.

1) Yes, exposed terminals on a beginner's device is not a good idea - that should be obvious.
2) What do you mean shouting?
3) What 'duff' 12V supplies? You mean the ones on Amazon, that sell 100s a month and have a 4+ star rating?

I agree on exposed terminals - bad idea for beginners. I really like the Rock Miner 100GH unit, but it needing a computer PSU sort of makes it border line beginner friendly in my opinion. If AM could make something that uses a power brick and had a built in controller similar in size to the RM device...that would be the perfect unit for newbies! Or make it USB based to keep prices down, and have a simple GUI for cgminer/bfgminer included on a disc.

Edit: Hate to say this, but something like that software BFL used

There would be no need to create a GUI if the driver support was provided to CKolivas and merged with mainbranch CGMiner instead of creating another fork that gets abandoned 3 months later too never be updated again like what Rockxie has done over at Rockminer.

Well for new people to mining having to use cgminer could be a little confusing. I am not saying fork anything just have a front end for it that is simple.

1)You are the first person to have suggested that this is a beginners device in this thread - One other person used the term introductory.
2)Shouting - the use of capital letters.
3)Yes the ones on Amazon that you get commission on that people said failed after a short time.

You have made a number of loud negative remarks in this thread showing a lack of respect for other people's opinions.
This time you use bold type to emphasise your opinions.
The discussion in this thread has been about obtaining samples, making boards, FETs, power supply efficiency.
You are out of order Dogie

Let's keep this to talk about engineering and assuming a level of competence.

Yes, I agree. I miss the days of even GUIMiner.

I haven't tried them, but there are a number of GUIs available for BFGMiner (yes, including BFL's EasyMiner - which should work for any supported device).

I like the idea of building a simple brick-powered introductory miner with USB connection. Something a n00b can buy, inexpensive but reasonably efficient and not requiring advanced computer knowledge or miscellaneous hardware stashes to use and maintain. Just saying, Dogie's not the only one thinking it. I've always said there wasn't enough (or good-enough) hardware in the mid-range consumer sector between USB stick miners and farm-grade hardware.

the problem therein lies that mid range miners, like USB stick miners are pretty much novelty items now. Sales volume won't command the demand on build investments, and making power bricks is just more manufacturing expense.
I do believe however, that there is no requirement to build humongous large-power miners with such efficient chips. FC would do better to keep product under 4TH/s... even that is a good home mining rig at these efficiencies and which can also be deployed in a farm setting.
I guess we'll have to wait and see.

I disagree.

These chips are ~2-3x more efficient than the next competitor's chips. With mining margins very thin, if you can get 2-3x the hashrate with the same electricity cost, why wouldn't you?  Of course, the price needs to be competitive as well.

i think we are getting ahead of ourselves a little. on-paper specs show these chips to be ~2-3x more efficient than the next competitor chips, but while competitors do not have the requirement to release their own specs, then by the time these are baked and ready, that ~2-3x will certainly not be the case.
then we have every h/w manufacturer on the planet vying to build 500TH/s rigs which in turn puts difficulty way up to a level where no mining is profitable.
we are approaching times where manufacturers will need to remain cautious. Super-efficient chips are all very well, but they still have to sell them to a market who are looking for profit.

I agree.

The spring will be an interesting time for sure.  I hope FC and AM build a decent size self mine with their cheap power in China.

Right now we are looking at the 2nd difficulty that will +/- 1% or so.  So that would suggest that mining has reached an equilibrium with price for the current generation of chips.  I fear that if price doesn't increase, not many people would be willing to purchase new chips.

i concur.

But therein could lie another issue, with spring comes the warmer climate in a lot of the world. it's going to be a very interesting time when it comes to deployment.
I, for one think that it's time all manufacturers should be working in tandem, working together to assure their own profits... the pulling ahead by any great margin by any one could cause problems for the entire network. To be fair, difficulty can be commanded as I believe we are currently seeing with what you comment on regarding difficulty remaining pretty stagnant is something which should be worked on by all. I want every miner to make their profits, and to ensure this, I feel that some form of collaboration between manufacturers is going to be required. what is your take on this?

I have some miners in my house acting as space heaters to help with the heating bill.  Once spring comes and the miners are no longer profitable (better to buy the coins), I probably won't buy many large ones because of the dual cost of electricity and cooling.

anufacturers working together to fall under some kind of Game theory model.  This would be interesting to discuss though.   I doubt it will happen.  Human greed is always present.  If I can buy more of these chips to 'print money' for myself, why wouldn't I?  Short term gains are, for the most part, more important than long terms gains for many people. 

A Bitcoin mining cartel? I like the way your thinking on that.   And I think we're overdue for one.   They could fix the supply as well as the price...   ;) I think we're off topic at this point though.  

yes, agreed, way OT. apologies, my brain often goes off on a tangent and i find it difficult to stop thinking... i'm hoping that one day it solves a block.
 :)

No, no they're not. Its just the test range displayed is at the very low end of the chip to demonstrate best W/Gh. The actual release W/GH will be significantly higher, as will the GH/chip. But even at .25 at chip, there are already chips that can achieve that at low clock speeds, and many more to come.

According to previous updates, the production chips would have have greater efficiency:


So, why are you assuming that they will be less efficient?

And looks like real performance of samples is better than predicted:

You're missing the point. There are 4 variables, not 2. W/GH, GH/chip and $/GH [and so $/chip]. In order to maximise $/chip, GH/chip will be raised much higher for a retail product than you see in those tests. That in turn, increases W/GH.

I'm not saying the chips can't achieve those lower W/Gh, I'm just saying don't be disappointed when the products based on those chips aren't .25W/GH. [This is of course ignoring that chip W/GH also isn't = to product W/GH.]

while the power numbers are certainly good, there are two questions i'd have.

the main goal when doing bitcoin mining is balancing the capital cost and the operational cost.

the capital cost can to a large extent by expressed as a performance/density figure, GH/MM, and the operational cost expressed as a Power figure W/GH/S (aka J/GH).  the remaining system costs (fans, power supplies, pcbs etc.. are largely constant regardless of the silicon used, so the main factor is efficiency and cost of the silicon itself)

clearly, many designers are working on asics that run at low voltages and thus achieve a headline low power figure.

The key is to achieve LOW POWER, while at the same time, ensuring that the silicon cost is low.

So it looks like the nominal performance of the be300 chip is somewhere between 3 and 6 GH/s (or 2.8 to 7.2 gh/s) per package.
And we've been told the package size is a 5x5mm square.   Since we havnt been told the die size... lets try and guess.  clearly its nowhere near 5x5 mm... so perhaps its 3x3 (=9 mm) or 4x4 (=16 mm)?

it makes a difference how big the die area is, as the gigahashes per mm need to be significantly better than last year's asics if its to be competitive in 2015.   it needs to be above 3 GH/mm to be competitively priced, and ideally north of 5 gh/mm.  And if... completely guessing... the die is... lets say its a 2.5 x 2.5mm, (= 6.25mm) then its going to be lower than 1 GH/mm, and thus will be expensive compared to other competitors coming out in 2015.   Bear in mind that last year's chips were all above 1 gh/mm, also in the tsmc 28nm process.  this year's ones should be significantly more efficient in order to bring down the cost of mining.

Possibly. You still need to know the ratio of silicon to other system parts costs. If a $100 miner is made up of $1 worth of chips, and $99 worth of other parts, doubling the cost of the chips only makes the miner cost $101. Unfortunately I don't know how much the raw chips cost to know if it significant or not relative to total system cost. The silicon could be a significant portion of the total system cost by now. It didn't used to be.

its the goal of the system designer to minimise the system and maximise the contribution of the silicon to the system.. and asicminer have been good at this.  their recent systems are quite efficient and their new one, with string support, should be even more so  - making most of the cost of the system will be the cost of the silicon.  that silicon needs to be extremely cost efficient on a gh/mm basis (which translates directly to gh/$), especially as so many mining asic suppliers are using the same tsmc foundry and similar process, thus their wafer costs are near identical, so the only differentiator is going to be the silicon efficiency of both power and cost.  ultimately, its gigahashes per wafer that directly affect the cost.

@aerobatic: how much does slicing, packaging and testing contribute to the production cost of Bitcoin ASICs?

And separately, how about the effect of yields on cost?

I would assume that AM has experience now in this field, after 2 generations of chips and battling with problems?!

bitcoin mining chips, with their large number of replicated cells (hash engines) are very good for yield - probably the best possible case that a foundry could wish for, as often they can still find useful and functional chips even if not all the hash engines are operational... whereas if you're making Apple A6 chips, even a single defect is fatal and the chip will be rejected.  you need every single circuit to work and any single defect on the die spoils the entire die, wheres for bitcoin mining, they can cope with some, and perhaps even a lot of defects and still have a useful and valuable asic.  for that reason, its not uncommon for bitcoin asics to have exceptional yield compared to regular asics, thus the yield factor doesnt become as important as the cost of each wafer and the efficiency of the overall design.

that said, as you know, wafers are round, and asics dies are rectangular.. thus the larger the die, the less number of dies fit onto a wafer.. because of wastage around the edges of the wafer where the circle intersects the many rectangles.. and if any part of th rectangular die touches the edge, the die is of course unusable.  asicminer uses very small dies, which should mean they will get less wastage at the edges of the wager and should have more usable dies per wafer.  on the other hand, the edges of each die where they're sliced puts a border around each die, which itself adds some wastage (die streets) and a larger area of the wafer is given over to edges versus usable die area on bigger dies.

i think asicminer uses small dies now because they have always done... whereas there would be less packaging cost, and lower board costs, jf they used larger dies with more hash engines per die.  there's some ineffiencies in using small chips that are not fully optimum.  in bitcoin asics... there are also various efficiencies of using small dies.. in that they might be able to be air cooled without individual heatsinks.. which saves cost and manufacturing time...   however, having each tiny die have its own package could become more costly than having larger dies with fewer packages.   personally, i think that larger dies are the way to go.  i think having tiny dies, thus high numbers of packages and larger boards etc.. adds more points of failure, and more board power losses.. so I'm in favour of larger dies in general.  not necessarily huge ones any longer (as they require exotic cooling, which we've all seen isn't as reliable).

Large dies are very hard to power and cool. Just compare the companies and die sizes and you'll see a very strong corelation.

Late Companies
BFL: Large die
HF: Large die
CT: Large die
KNC: Large die

On time companies
Avalon: Small die
Bitfury: Small die
Asicminer: Small die
Bitmain: Small die
Spondoolies: Medium die

It's pretty clear that the companies that tried big dies has the most delays. I do agree the Asicminer dies could grow quite a bit and still work well.

I don't know about a sea of tiny chips, but I'd much rather see a multi-chip high-hashrate design than a single- or few-chip high-hashrate design. Increased surface area decreases cooling requirements, distributed power decreases PCB requirements, and increased modularity increases reliability and repairability. From an owner/hardware-maintainer standpoint, the most problematic machines I've run were the ones with hot single-die setups.

i had always thought it the other way around.. the larger the surface area, the more cooling required... you are not spreading that heat from the centre to the edges of a large chip, you are seeing equal heat across the chip - so a bigger chip = increased cooling requirements.

You're both saying the same thing, distributing hashing power over multiple smaller chips increases the surface area / GH = 'easier' to cool. That being said, its not difficult to cool large chips at all, it just means more cost and weight. [First step up is copper base, then copper heatpipes].

i didn't say more difficult to cool, but more cooling required, which would be obvious?

raskul, I didn't mean per-chip surface area so much as per-miner surface area. It's a lot easier to supply and maintain a handful of 4x4 heatsinks to cool multi-chip blades than it is to supply and maintain a waterblock and radiator for a 400W 2cmx2cm die. That's what I'm saying. When I said "increased surface area decreases cooling requirements" what I meant was increasing the board area over which you have to spread cooling - take the difference between an S1 versus a Sierra board. Similar power draw, but one's spreading it out over a square foot of PCB instead of a square inch of silicon. The AM Blade would run passively cooled with its ~5x8 inch heatsink at 85W, where a Minion requires decent heatpiped CPU coolers to draw the same power from a single postage-stamp area. Not that my opinion is particularly valuable, but I'll pretty much always endorse multi-ASIC designs with simpler heatsinking requirements over things that require complex or cumbersome cooling setups.

Having single ASICs that run about 5GH at 1.2W per chip, one could stack them about as densely as the BE100s on the old AM blades and make a decent machine. It'd run cool, but not terribly space-efficient - about the same as the BE200 chips, I think. Maybe a little better.

ah, gotcha, in that case, i concur.  :)

AMleak

FTFY

Not only were some later with a larger die they were scamming people.

when you put it in black and white i agree with you, those big chips were way too big and relied on exotic cooling, which turned out to be costly, complex, and unreliable.  in the small chips versus big chips experiment i think its now been proven that extremely big chips are too difficult to cool reliably so yes, i agree with you.  but there's still a very wide spectrum of possibilities.  asicminer has gone for a very tiny die with only 6 gh per chip.  if it runs at their claimed, lets say 0.3 watts per gh, thats less than 2 watts per chip.   thats probably much too small to be efficient.  thats quite a bit less than bitfury's old chip (2.5 watts).  i wouldn't be surprised if bitfury's next chip will be more like 20+ gh per chip.  its all about watts per package that can be cooled in a low cost way.

i think the medium sized chips are probably the most efficient way to go.. the chips are still air cooled (with heatsinks and high cfm fans).. but they pack a lot more gh per package which reduces the number of boards that you need for a given gh's, keeps cooling simple, but also keeps the system size manageable.  less board losses, etc.  spondoolies was onto something with their package size and cooling.

Missed that - anyone an idea why he took it down?

It's been like 15 minutes in first post. Then replaced.

More interesting is that PMS01 chip is AM design? If so, it looks like friedcat is talented ASIC engineer...

PMSL  :D

On pictures it's clearly visible PMS01...

PMS. indeedy.

I think what he meant was, BFL and HashFast (as indicated) were revealed not only as "late" but also as "confirmed scams". I'm no shill for AM, heck novak and I spent a week and some out-of-pocket coin assembling Tubes for hosting and trying to find fixes for their epically-broken stratum implementation, and then rebuilding shelves to fireproof for Prismas (we haven't had any blow out yet). Not impressed. But I do like these chips, it might be a win.

Please recall, however, that this is a thread for discussing BE300. I think what you want is the other thread - the one about compensation and RMAs for existing hardware. Also please note that AM has pretty much always been bad at communication, and that there are likely a few thousand people trying to sell back their Prismas so a bit of patience will probably be necessary.

I was on your side and you lost me with this post.  
BTW BFL fully refunded my money after 16 months, but to say BFL is good or okay and AM is bad or terrible will lose most of us.

Just say AM fucked you and you want a refund and I will be back on your side.

Now to this chip I will most likely buy a piece of gear made from it. I will use CrazyGuy or CanaryInTheMine.  I will wait for others to buy it I won't be a Guiana-pig or lab-rat.

But I like new gear.

Prisma's have been a pain in the ass. But mine have been working steadily for two weeks now so I'm over it.
Can't wait for these chips to hit the market. Time is money FC. Make it happen!

that's very harsh mate, if they are providing customer service of that level for one product, they really should make it across the board and honour it to your own product.
it's like a bike shop offer a repair on a flat tyre - but only if it's the back wheel, front wheels don't get fixed.

If this was a Spondoolies thread, these completely off-topic posts would have been deleted by the mods by now. I guess FC needs to donate a miner.  :o

if this was Spondoolies thread, that unit would have been RMA'd by now.

Off Topic
We all know about BFL, i am also BFL victim.
But what with technobit? My order with technobit is fine until now.

Sorry for not reading up on the whole thread, but when will the first miners with this chips hit the market?

estimated ~April 2015.

Probably due to this.
https://bitcointalk.org/index.php?topic=682105.0
Boards were supposed to be shipped two weeks after the chips arrived (so early-mid August or so), and people appear to still be waiting. Not sure if Marto's made arrangements with them, but the people in that thread don't appear particularly happy.

We've even volunteered (yes, free of charge) to do some of the work fixing their boards so US customers (including the 15-odd KW worth that are supposed to be running in our hosting facility) could get them quicker. Marto gave us some info and then stopped talking to us and Minersource, over a month ago.


We're considering trying to build a mining board with BE300 chips, but the biggest fear is we'll accidentally do something like technobit with the minion boards. I don't think we're capable of screwing up that badly though, especially not where communication is concerned.

Can someone enlighten me what the problem is with the AM miners?
Ok, they failed bad, but it seems that FC's buy-back offer is more than adequate.
What's your issue?

Basically, if you buy a batch 1 rig, then you should just finger cross that you are lucky.
I'm the bad luck one and also the good luck one.
I bought 8 Prisma and some of them not work very well. (Bad luck)
Now I have send three of them for exchange the new 2.0 Prisma, and if those three all good as expected, I will exchange other 5, otherwise I will refund those 5. (They will not refund the rigs if you choose exchange instead of get refund). ---so I am also good luck that I have the chance to replace the new batch ones or get refund.

The 6-chip chain has passed the test:

http://i58.tinypic.com/28rgnxl.jpg

The next step is to test the board with fully inhabited 24 chips.

Funny, I just wanted to ask how test of chained chips board are comming ;)
Results?

Great news!! 

We are waiting for a compilation of data set.

But 7.2GH/s per chip is already achieved, so at least the chain is stable.

Indeed, good news. When can we expect chip documentation, voltage requirements and such?

and financials plz.

I've raised this question before: This PMS01 chip is proprietary design of AM?

Have the optimizations for the final BE300 design already taken place and has the order been submitted to TSMC? When will it take place?
How big is the risk of the final BE300 design not working or underperforming (like BE200) after receiving the final chips?

I would suspect BE300's risk is terminal, which isn't always a bad thing. Some times you have to go all in, and with careful planning and prep, clear the table.

Missed that, thx.


Those chips are MPW samples. Friedcat posted that mass production chips will be better. Propably this samples are just to "proof of concept" and mass production chips will just have more "cores" inside as adding those is almost like CTR-C CTRL-V

Naming a chip PMS is just asking for intermittent trouble.

LOL agreed on that one!

Will the BE300 boards fit into the Prisma fan/heatsink body?

I would like the cat to make an outstanding miner.  maybe the size of the smaller tube do 2th using 600 watts at the wall .

Noyce!

whats the power consumption?

Yes!  That would be great since there are probably tons of prisma heatsinks and fans laying around with the buy-back.  Since you would only need to ship the new boards, you can save money with shipping.

AM has said before something about wanting modular style miners.  There aren't a ton of improvements that can be made over the Tube/Prisma cooling design, I could see them offering simply boards, fans, and controllers as bulk buy/piecemeal similar to the way they have done before, but without having to buy the heatsinks.

From earlier statements it appears that the power density of BE300 miners will be lower, and so won't need nearly the same levels of cooling. Sure if a few of the screw holes match up you could put them onto Prisma heatsinks but at this point it may just be overkill.

Or ship the heatsinks, which would be an improvement as well.

Considering the overall failure of the safety design of Prisma's, overkill may not be out of order in the cooling department.  Besides, the move to 1KW+ miners isn't unwelcome, at least not in my case.  Less miners, less cables, less points of potential failure...


Agreed, if only they were further in development. They could have made an upgrade/trade-in plan for Prisma owners instead of having to pay all shipping costs back, and presumably re-use the chassis and re-ship to customers. Such is life, by the time these are released I'll be ready to upgrade my Tubes anyways.

If AM doesn't do it, someone else could. With low power density and over-heatsinking comes very quiet miners that probably won't burst into flames.

What do you think about the smaller packaged die size and less over all components on the boards due to PMS chip,  combined with the lower heat density..  potential for x48 to turn into x96 maybe?

7.2gh x 96 = 2.8TH icy cold Prisma's at ~850w?

First post:
7.2GH/s | 0.2495W/G -> ~1.8W whole chip

The tubes are currently being used to dissipate -conservatively- 720 Watts (180 Watts per board) of BE200 chips on x24 boards based on their testing of on-board power consumption at the time of development (0.9 J/G).  Extrapolate 180 Watts @ the 1.8 Watts per chip & 7.2 GH/s voltage/frequency combo tested so far, and you could dissipate the heat of 100 BE300s chips per board in a tube-size chassis. An X96 board revised for the smaller die size and made to fit the Tube heatsink package would work splendidly.  Theoretically we could end up with 690 GH/s per board, or a 2.7 TH/s miner and less power draw than a current Tube.  This doesn't even take into account that cooling is easier when spread out over more points of heat source on a heatsink.

I think with the current problems with getting things stable, that the most sensible solution would be a smaller board. If it dies but you've got 15 others in a miner, the consumer is going to mind a hell of a lot less than if you had 4 and 1 was bad. Think Avalon1, mini modules within modules.

That's sort of the design we're toying with, configurable strings that can be tied together. Our idea is, each module could be run as an independent consumer miner or they could be lumped together into a bigger unit.

The package turns from QFN to Flip-Chip-LGA, which means that
most heat dissipates to the upper side of the chip instead of PCB,
so... not very appropriate to stick to the piple-like heat dissipation
design.

And the chip size is 5mm x 5mm ?

FC, you can probably dispel a little confusion here. The MPW samples you have and are testing are not in the same package as the final chips will be in. The production chips will look significantly different than the test chips, and the board package will be different.
Is that correct?

Package propably will be the same. I'm quessing that die inside will be little different (more "cores").

These two are inherently incompatible statements :)

I was reffering to this:

The ones that are mounted onto the boards in the pictures shown are the MPW samples, and they aren't FCLGA. If the production chips are to be FCLGA then it would make sense that the package is different between the MPW samples and final chips.

No pictures of chip itself so we don't know. There's no reason that it couldn't be a FCLGA... It's not a big processor with over thousand pins so it dosn't need to look like one....

There are pictures of the chip mounted to the board, as well as the board mounting pattern. While you could do a flip chip like that an encapsulate it, it would seem weird to do so with those massive power and ground pads at the bottom if you're planning on pulling most of the heat out the top. You can get overmolded FC packages, but the thermals out the top aren't usually great.

I will concede it could be FCLGA though.

I'm also curious how friedcat want to achive better performance in mass production chips. As for now we only know that with samples he achieved better specs than simulations. 6GH/s | 0.343w/g  vs.  7.2GH/s | 0.2495W/G

Sometimes things are different than they seem, especially in China.

I have seen these small LGA packages in sensors (from Analog Devices) for instance.

BE200 -- QFN 8x8 or 9x9
BE300S -- FCLGA 5x5
BE300 Final Version -- FCLGA nxn (n not determined yet)

QFN in a low voltage setting introduces too much overhead, so although we wanted compatibility hard, we chose a different package.

i thought Flip Chip Land Grid Array (FCLGA), would require pins, or solder balls on the PCB, to connect to the land grid at the bottom of the package. Something like them socket'ed intel cpus, you know, they flipped the chip for the bond pads to face down, for the bond wires to be all inside the substrate.

these BE300S's still look like Quad Flat No-leads (QFN)

now Flip Chip Ball Grid Array (FCBGA), like you'd see on DDR2/3/4 ram, is what i thought the setup would be like.
..if you want exposed dies, you could follow the path of what intel does with one of their gigabit Ethernet controller chips, have the dies backside flush with the top of the substrate. or you could copy knc or cointerra and have bare dies just sitting on the substrate and held by glue.

on the note of compatibility, there was no compatibility between the BE100 and the BE200.  we all like the "drop in and run" idea, but i've never seen too much backwards compatibility with packages like this. so don't kick yourself over it, make the package the way the chips need it, even if you have to go 40 pin DIP, people will work around it

... am i rambling on?

Yeah, I miss alot 40 pin DIP packages...
Those were good times of C51 uC...

Heck yeah. PCB costs would suck but I would totally build a miner with DIP. BGA, not so much.

The backward compatibility idea was pretty great, especially when they open-sourced BE200 design, because a lot of folks were working on stuff hoping they'd be able to reuse the same design. Would have been nice, and I'm glad they at least tried to make it happen.

So uh, so when do we get to see the Vcore/GH curve?

AMHash takes 47$/Ths/month (0.8W/Ghs) = 0.08 USD/kwh - is it a cheap power?

You have to figure maintenance, internet, staff, datacenter are all included in that $47 / Th/month.  Cheap power would be ~$0.03 / kwh.

There is cheaper power in CN, however there are plenty of other reasons that manufacturers want farms in other countries even if they're more expensive.

It might be misleading to post it for the BE300S if he thinks that they're going to be significantly different than BE300 - which its suggested it will be.

What do you expect the Tjc to be for the top and bottom on the production chips?

Right, but at least it'd be a good starting point. We have really good data on expected W/GH at various GH setpoints, but without knowing either the I or the V, the W is basically useless except for cooling concerns and overarching "efficiency" marketability. Sure we can plan heatsinks, but not string lengths or VRM outputs or anything. Even in accurate data is still data.

the only reason FC is obliged to release any specs at all is due to the fact that the company is in perpetual ISO.
we all know the initial specs will be a wide variation of the finished product, it's always the case with AM.

They have to find really cheap power, if you ask me! i want to see 0.01$/KWh! That way they could build a really big farm an mine the hell out of the blockchain. AM isn't a small player, they should go as far as possible and re-claim those 40% in 2015! Also they could continue to offer a cutting-edge cloud mining!

I LOVE this talk.
Guns a'blazin.

*cough*decentralize*cough*

Even if the network wouldn't grow much from today on, claiming 40% of the network possibly means building 100-150MW worth of new Bitcoin mines in 2015. That's about half of all of Google's capacity (265MW) by power combined. And twice the size of Facebook (78MW). And it remains a moving target during and after.

I had a post here with some numbers I put together recently:

Energy Consumption of the Bitcoin Network
http://bitcointalk.org/index.php?topic=520977.msg9825138#msg9825138

Never going to happen, they simply don't have the cash to do that on this generation. Even Bitfury's level of cash would only just allow them to do that (if they started from 0%).

 BitcoinBrothers out of Germany claimed in November this year to have an ASIC based on a 16nm 3D finFET and have plans to unleash 300-400 PH onto the network in Q1 2015.   They also claim each of their 6 PH machines will consume 0.9 MW.  So using your initial assumption, it appears they are poised take a little more than 50% of the network (assuming they add 300 PH) using only 45 MW.
 I haven't dug very deeply into this claim but they seem to be echoing the claims of KNCminer as far as the 16nm 3D finFET based ASIC.  I haven't seen power consumption estimates from KNC however.
   

Not sure where (electricity rate) they plan to drop their bomb, but the numbers you have posted put their power consumption at 0.15W/G.

Official claims we have (also from my internal report):

- Spondoolies Tech 0.05W/G mid 2015
- KnCMiner 16nm 0.07W/G early 2015
- Cointerra 16nm 5X (?) Q1 2015 (5X refers to their TerraMiner IV I guess)
- Bitfury sub-0.1J/G mid-year 2015

Not sure what to put on the Chinese ASIC manufacturers when it comes to official claims (Bitmain, Avalon, InnoSilicon, who else is significant or still in business?).

Results? Competition says 0.5J/GH. AM says......... ?

BlackArrow has made claims using a finFET asic chip as well at the .07W/gh - claiming 20TH from about the same draw as the 1TH of the x3.  BS or not - we will have to see.

BlackArrow still has money? How?

When i will see intel and amd have a mature desktop cpu at 14nm, then! and only then i will believe those numbers!

The smaller we go the harder it is!

Just a shot in the dark:  under .1w/g in 2016!

I think it will happen in 2015.

It will likely be possible, but being possible and being the shipping clocks for a product is another thing. Maybe we will see both, maybe just one.

My money is on Bitfury and SP-Tech

But if BitFury does it, nobody gets to share the wealth?

I suppose it begs the question: can these any of these companies get chips manufactured quickly enough to account for their lower power? 

most likely Q3 2015.

sp tech has not met a target hash rate or power draw yet

Not my thread, I'm not sure I should reply here. Anyway:
We exceeded our 1st gen goals, we've missed (and compensated) our 2nd gen goals.
Regarding our 3rd and 4th gen ...  :)

so what were the expectations of the one generation that was released without customers funding it? it's pretty easy for you to say anything as it was not pre announced or pre sold.

You missed the mark on the "Slow Corner" May SP10's as well.

Yep...anyway this thread is about ASICMiner not SP  ::)

That's a reasonable position to take, as manufacturing does become more difficult as you shrink the architecture, however, there is quite a bit of difference between a SHA256 ASIC and a CPU.  The CPU is much more complex, as most of a SHA ASIC is simply repeated hash cores.

I think many manufacturers missed their efficiency targets at the last generation though (ASICMiner, Black Arrow, KnC, Spondoolies, and probably more), so that's not an unreasonable guess.  However, reading this thread from ASICMiner indicates that they now have enough experience to design an ASIC at 28nm and hit or surpass expected efficiency.  Whether or not companies can now drop down to 16-20nm or so and get the efficiency they think they can is the million dollar (actually probably a lot of millions) question.  It looks like we have announcements from multiple companies attempting it by mid 2015, so it is entirely possible.

--
novak

+1. Bitfury was doing extremely good numbers (~1w/GH) on 55nm in the first version, and improved that by about 15% in the 55nm//Rev2. undervolting allows something around 0.85w/gh on those chips.

a 28nm chip with exact same design is theoretically capable of almost 4x the efficiency. that would achieve around 0.2w/gh on an underclocked chip. 20nm would achieve around 0.1w/GH with slight underclocking I expect. implementing improvements like chained power distribution or perhaps larger die sizes could also improve on the numbers.

(in contrast, the Bitmain 55nm chip from the S1 is not much different. 2w/GH at full steam, but with a bottom value of about 0.9w/GH if undervolted to about 40% of the full speed. Thier 28nm chips (S3 and S5) are capable of 0.5-0.8w/GH and 0.25-0.6w/GH power ranges also - clearly showing the improvements found by thier 28nm//Rev2 chip in the S5)

AM is shy to present us their 24 chip board.

I concur. SP is the most transparent for both upcoming tech, support and refund policies. They continue to keep the industry notified and always respond usually within 24 hours to anyone's comment.

Whilst other competition is coming along nicely, they aren't transparent and simply don't communicate keeping us up to date. We get an email once a week if we're lucky, usually once every 14 days or so. This level of communication isn't satisfactory in today's highly competitive market.

If AM could either hire additional staff, or devote more time themselves, that could answer our questions, openly and in a timely manner, then they are likely to receive a warmer following and brighter future. SP is the leading example in all these areas. Bitmain towards the top, now falling like a stone. AM whatever you do, don't follow bitmain's example, rise above, communicate with us and show us what you can offer.

 

+1

I'm also interested in some samples, will check usability BE300 for USB dongles :)

About USB miners -

Bi•fury use 1A for ~5GH/s - USB 3.0 or powered hub needed
Hex•fury use 2A for +10GH/s - strong powered hub needed

USB 2.0 will deliver 0.5A on 5V = 2.5 W
USB 3.0 will deliver 0.9-1A on 5V = 4.5-5.0 W (especially in PC, not all notebook/laptop USB 3.0 will deliver 1A)

One BE300S chip deliver ~7.2 GH/s = ~1.8W

According to USB power we can use:
1.39 Chip on USB 2.0 = ~7.2 GH/s with 1 chip, ~10 GH/s with 2 chips.
2.78 Chip on USB 3.0 = ~10 GH/s with 2 chips, ~15.6 GH/s with 3 chips.


That look nice, and according to string design we can save some on BOM,
BUT according to new information about FCLGA - PCB and heatsink cost can pass out "reasonable" cost of USB miner on BE300 (not count AM itself of course) :(

PCB and heatsink aren't much more difficult if you put a bit of work into it. Heatsink on top of chips is not really any harder to accomplish than heatsink on top of board. Just stick your ASICs on the backside, control hardware on the top side; grease your chips, sink the backside and screw to it through the board.

For a string design to work AM have PMS01 chip. Without that chip string design is impossible(?) or more complicated if you want to replace that chip with avaivable IC parts. Either way string design on USB stick is like firing to fly with big gun...

Yep, there are some options, that why I ask about samples, all other we will see


Will check, and talk about this Dex :)

How would you run a string design? If you want to avoid DC/DC altogether you'd need 8-9 chips to get the 5V down to 0.55V or so. I'm not sure how you'd get away from DC/DC on a USB miner.

Depending how many you want to string, it might not be too bad. But stringing works best if you can divide the rail voltage, which for 5V means more chips than he's talking, so you'd need to build a VRM anyway. At that point it's much easier to just kick the voltage down to single-chip level and run several chips in parallel, standard topology.

Pretty sure the NF6 is strung chips, but it's also (as noted above) a very high current requirement. Not caring about standard USB port current limits changes the ballgame.



Also, he beat me to the punch but MrTeal is pretty much always right.

As a complete and utter aside - note that these specifications are usually per port.  Old portable 2.5" HDDs and optical drives would often include a split cable with one end of the Y-split being power+data, and the other just power, allowing a second USB port to deliver the extra oomph needed for the optical drive / HDD spin-up.
( Looks like most current external HDDs no longer use extra power, and some optical drives seem to come with a USB-USB cable and a USB-barrel cable )
I'd certainly view a USB miner that used shenanigans like that as still being a USB miner.  Might have to draw the line at utilizing 4 ports, though ;)

Nanofury 6's using the rev.2 bitfury chip already achieved 15gh/s.
It might be a fun project but I don't see a market.  I produced 100 of these and while we sold out, they did not go fast. That was 6 months ago.
I think you need to achieve 75-100 ghs to have anything marketable.

I've got several NF6 in the museum, not sure if they're Rev2 chips or not (USB miners is more novak's project). Swimmer, you built miners? Nice. We should talk sometime.

I'm was not the tech guy sidehack. Vs3 handled the designs. I just managed the build and did all the sales. USB miners were big business for 8-12 months. Love to figure out a way to bring them back.
But with difficulty now they would have to be able to do 8 times what they did.
The guys that bought them usually got 5-10. Folks that could not get away with spending much or having a small jet engine in their home or office. Still a big market there if we can figure out a simple, cool, small form factor that will do 75-100 ghs.  Still have the website setup.

Novak and I are toying with the idea of a configurable USB-connected single-board miner for BE300 chips, something that could run off a DC brick (or better PSU, it'd have a barrel and screw terminals) with full under/overclocking options from the command line. Basically the Jalapeno market sector, except with tomorrow's power efficiency. Drop me a line if you think there's merit.

Just be careful about assuming 3.0 can do high numbers. Even if a single port can, it doesn't mean the array of ports / other electronics or even the PSU can take many many maxed out 3.0 ports. If you tell a consumer that it will work fine on USB 3.0, they'll assume that if one works then they can put 6 in.

Even the 'big' PSUs like the CX750M only have 25A on the 5V rail, of which a proportion will be taken up with other components.

I'm not a EC designer so I can't answer now, If it is possible and how, especially because I'm not work with any string able chips beside BF.

I was working with Intron and C-scape, they make first string design on BF chips - OSM.
Next they propose me to produce extended version of Bi•Fury - Hex•Fury and that was first ever dongle USB Miner with string design, on 6 BF chips.
(I'm not sure but I'm think that design of NF6 is familiar).

But I want to produce product especially for "miner wannabe" market(like Bi•Fury), that why I stick to USB dongle design, taking max from USB spec draw.
This kind market customers don't want use special DIY hubs for power it and etc.

Since BF I don't see any chips for simple, low cost USB device that can give this kind of customers some scratch's of BTC form there "miner fun game".

But producers going down with power draw (Yes I aware about high current) so I'm still searching, asking etc. ;)


Finalizing this "My story" speech I have proposition for all You EC designers:
If You can deliver some kind of design on any producer chip that suit market needs I describe, I can take all production operations(finance-production-sell) on me :)
[Interested in co-op? drop me a PM]



Yes I'm aware of all this this dogie, since Bi•Fury I had 1k+ working USB ports under control for testing devices.
Many HUBs models, OS and hardware setups(mobo, PSU etc.) that can get stable max form Bi•Fury.

That why I know that this kind of information You mention above, need to be placed in product describe for customer awareness.


[FC sorry for little off topic here, and still waiting for more information about test ;) ]

so? no news good news?
FC last logged in today.

It would be nice to hear some word from FC, atleast we would know that he didn't died at "drawing board" ;)
We know that sample chips works better than expected, string design also works, so propably next news will be announcment of new product. Prisma 3.0?

The expected date of first batch is April 10 to April 30. Potential delay will be notified three months ahead.
Hashrate per device and power consumption:
  Normal mode - 1.0TH/s guaranteed. 1.6TH/s maximum. 0.25-0.27W/G.
  Turbo mode - 1.2TH/s guaranteed. 2.0TH/s maximum. 0.30-0.33W/G.

this is a update

 This is very good news but is it an official update?  Cand ai inceput sa lucrezi pentru Asicminer?

Many things don't add up, so I think this is speculative update. Having chips in February and miners in April? When everything is prepared upront? Chinese fabs can produce hundreds miners per day, next day after FC say GO there will be thousand miners ready on shelfs.

im not so sure about that.
from previous experiences, it seems their assembly chain is not that effective.
remember when they used to just send hundreds of screws to assemble the miners they could not handle by themselves?

"Burn in" ...

Lets hope they stick to self mining/cloud/bulk chip offerings. They obviously can not handle customer service when selling to individual miners. Let companies like SP-Tech and Bitmain sell to the general masses.

Ohh, right. Forgot that magic word... But history of AM doesn't show evidence of such a practice... As hdbuck said they even hadn't time to screw screws..

Where is the 24 chip board?

on the tank track road towards tibet  :-\

24 chip (6 chip chain x 4) test result posted.
The mosfets and big capacitors are safely removed due to the constant workload of BE300.

http://i60.tinypic.com/14v5h7k.jpg

Results:

273.60GH/s | 0.38W/G
176.64GH/s | 0.31W/G
155.04GH/s | 0.27W/G
118.56GH/s | 0.26W/G

With more accurate control on the variance between chips we could in principle gain less power consumption in advance.

It's right ^there^.  Duh!

Any more questions ChoadStress?   ;)

It's seem that the color on some area of the board become darker. Maybe due to high temperature.

Shadows? It's such a crappy picture, it's pointless guessing at that...

There is only a fan above cooling that board, and they tested it at 273 GH/s - 11 GH/s per chip.
Single chip board test report was up to 7.2 GH/s per chip.
They must have stressed this 24 chip board up to the limits.

Im gonna guess its solder flux.

ROADSTRESS..... where are you?

I am right here don't worry. Only took FC one full month to publish the results. If you 2 are happy that AM has a 0.3W/GH chip for at least 8 months from now then I'm glad for you  :D Time to buy some AM shares!  8)

Looks authentic - photo taken 2015:01:12 15:28:21 in Binhai Ave, Shenzhen according to the EXIF information.

I wonder how well the mass-production chips will handle dissipating 4W each, being a different (smaller?) package and all.

great work friedcat! my prismas are working very whell and whainting for your new release!

This chip looks very promising for AM. It's good at both ends. Efficiency and performance if needed...

pity that for decent miner you need 100 chips...
costs matter, those miners cant be cheap

look at bitmain they have so few to get 1.15TH

Right, only 60 chips that are larger.

Cost of that chip for AM is likely ~1$. PCB in large quantities is something like 5$. Rest of parts (few caps and resistors per chip is 0.1$ most), so 100 chips board will be around 120$ (assembly included). Bitmain sells for 0.27$/GH? If AM do the same it's over 100% profit.

I'd like to see about a 30-chip board that can be software volt/clocked between 100 and 300GH with a quiet fan and can run off a DC brick. That'd make a really good entry-level miner, Jalapeno or U3 market sector.

Maybe I should just go design it. But that's about three projects away. But I think it's a good idea.

Yeah, this chips scales quiet good. Bitmain's chips don't do well in that matter. Going to double hashrate means decreasing efficiency almost twice.
https://i.imgur.com/sipgJ07.jpg

With AM chips double hashrate = only 25% worse efficiency.
 3.6GH/s | 0.2095W/G
 7.2GH/s | 0.2495W/G

Yeah, it is fairly flat. The upperbound on clocking is going to depend on a lot on how easy it is to keep the chips cool. If they're like AM's chips of the past though, they won't mind high temperatures too much, but in a 5x5mm package (if I'm remembering right, it's several pages back) that's a real question. I know BE100 chips, in I believe a 6x6mm package, it took some rigging to make them run reliably at 4W dissipation. That was a PCB-cooled chip with not-the-best PCB-heatsink contact though (at least on Blades), so maybe a top-cooled chip will have a better overall thermal resistance between case and sink?

5x5mm is a 30% reduction in surface area which is a huge hit. Getting that heatsink right is going to be mighty important as to the formfactor of the device. Nice thin blades like in the Cube at this point are really required to get any sort of 'next generation' density out of these things.



That's not really a fair comparison when the BE300 is running / being tested in a range far far lower than the BM1384 is running at. If the graph continued back towards 0 it would probably look very, very similar to that of the AM chip.

That's my main concern, but hopefully the chips will be priced low enough that density isn't prohibitively expensive. Novak and I have discussed working up a design for a 30-chip board which could be roughly the same size as a Cube board, and several could be stacked together in modular miner configurations for devices of different hashrates. Nominally 150-200GH per board is probably a decent setpoint. If we could build them to be compatible with S1 and Tube heatsinks could make for some fun upgrade kits as well.

This x1,000,000.  I'm tired of excessive depreciation of redundant hardware, when the only portion of it that is redundant is the PCB's/Chips/Controllers. If upgrading made financial sense, more miners would do it (I.E. NOT the Bitmain S1 upgrade flop).  The savings on purchase and shipping of frames/heatsinks and fans would also give the manufacturer an edge in production cost of working miners compared to competitors, and give them more room to mark-up the important parts while still saving us money at the door.  Do this!

I think if you're expecting $20 over the cost of the chips you're going to be a little disappointed. You aren't going to get PCBs, other components, PCB assembly, heatsinks/fan, controller, unit assembly and testing for $20, even in quantity in China. At current prices they'll still be able to make money on BE300, but 100% markup in this market is wildly optimistic.

Hey Icebreaker, please leave. Nobody wants you here if you're not going to be polite.

He's mostly mentioning because the single-chip data was posted last year and the prototype string board was already being tested at that time. Basically it's a commentary on how inept ASICMiner's PR/communications can be, more than any questioning of their technical competence or engineering efficiency.

If no one talks to him, he goes away. Just like grade school.

You must mean Spondoolies. The SP20 uses 4 chips while a Bitmain S5 uses 60 chips, which is less than AM but still, that's a lot of chips.

I'd much rather see a "many small chips" design than one relying on the operation of a single large chip - but this has already been discussed extensively in previous pages. Cooling is a lot easier, as well as increased redundancy (which is somewhat negated by a string design, especially with no cap/FET buffering). Providing a single incredibly-high-current low-voltage power source is also very problematic and prone to resistive-loss inefficiency up the butt.

Only if integers matter on their own. The (cost per chip per GH) or [(cost per chip per GH) + (board components per chip per GH)] is the most important thing.

Edit: Seems I imagined you said something totally different, but my comment still stands on its own :P

Kudos on what looks to be a promising miner! Thank you for being transparent in the process.

Okay 155 gh  x .27 = 42 watts   once you figure psu in  47 watts .

lets say 5x =

    775gh at 235 watts on the low clock.
  1365gh  at  576 watts on the high clock .
 this is a 120 chip design.

right now I can buy the sp20 for under 400 usd.

It will do 970gh at 450 watts

the s-5 will do 1300gh on overclock at 676 watts  cost shipped right now is 342 usd if you have a coupon.

For a 120 chip AM to be worth buying it needs to come out soon and be cheap.
  At least the board did decently (not .2 watts) more like .3 or .33 watts. 
The sp20 can do .46 via underclock.

Since I have owned every AM miner I will buy it just to fuck with it.  Hurry up with one. :)

Price of 120 chip design should be in principle at 230-250 usd w/o the PSU.

If the mid-term future Bitcoin price allows significant quantity, the end-game price can be below 200 usd.

Does that mean you'll sell the chips for a sub 1200$ for 1000 price?
If the answer is yes, I may be interesed to buy a few samples at sample price when they will be available

I understand squeezing the competition.
We just witnessed the squeeze for the last 10 months.

When your next competitor is charging $342 for a product that is less efficient,  why would you price your product at $250?
I don't think there are going to be enough rounds of these chips to keep weeding out competitors before you can enjoy a little harvesting.

Maybe I misunderstood your comment.
Maybe you were just saying you could lower your pricing down to $250 and still make money.

But, please.  Cheap hashing power isn't going to do anyone any good.
Pricing war.
Difficulty into the stratosphere.
Investors make shit returns.

Really nice Friedcat

Wondering do you have a estimated date for the release ?  ;D

  In a nut shell cheap mining gear killed off LTC.  Gridseed blades were 3000 usd and ltc was 34 usd  .  gridseed blades dropped to  150 usd and ltc is 2 dollars.
If AM puts out a .33 watt 1.5th miner for  250 usd Shipped it will only serve to lower the price of BTC.
BTC diff-Price ratio is fucked up as it is.  I can not see the btc network working much longer without demand for the coins lifting price up and over the 500 dollar mark.

That price would be appealing today, hate to ask the question with the impossible answer, but when will they be available? :)

There is this post, not official, and I don't know who the poster is (tutorialvideo)
https://bitcointalk.org/index.php?topic=888260.msg10123817#msg10123817

Good to hear this update. The power draw is still higher than single-chip solutions, but looking good, still. Any news on the expected mass delivery of first BE300-final chips?

That is what I'm worrying about, with endless price war, the one that can stand many years of loss is the one with bank loan support, so eventually banks will take over all the mining chips business

Unless the major chip makers make an alliance like OPEC, they will be take over by banks one by one, but unfortunately bitcoin is free to mine for anyone, not like OPEC countries which have geographic access limitation

In fact, knc is most likely to be already running on a large sum of bank loans, so that they can operate in a loss for at least a couple of years

There's no reason to think cheap mining gear would cause the price of BTC to go down. LTC might have gone down in price with cheap hardware, but that's more than likely a coincidence.

I for one would like to be able to get good gear at that priced shipped.  I really did not like the tube I bought and was glad that caused me to skip the prism. 
This price point with stated hashing rate and power consumption might bring me back to try another AM product. 

Banks don't lend money to unprofitable businesses.  It's not their job to support years of losses through loans and they certainly wouldn't do that for Bitcoin mining businesses - something which banks have yet to warm up to!

All of that aside, a loan is just that - it's meant to be repaid.  What mining hardware manufacturer could afford to take a couple of years of losses, with any hope of repaying them (plus interest) from future profits?

VC funding is similar in this regard; it's not there to cover losses.

April if I remember correctly!

Hopefully April, as depressed btc price can be a bit of an issue since factory needs to be paid in fiat

I think it is a really bad idea to compare the BTC price dynamics with LTC or any other ALT. The factors determining BTC price is much more than just difficulty or the cost of miners, which is not the case for LTC.

April or May

if price of btc will be <200$ this time
miners will be not sellable in large quantities
maybe for hobbist

Miners didn't kill the price of LTC the market did.

True.  Little to none adoption of Ltc is hurting its value.  But it is also designed to stay within btc ratio which it is.

Thank you for sharing progress FC, communication from the team is always appreciated, even from the small investors, observers or the mining community.

It's not designed that way, it's just the way the whales manipulate the market.

I am curious as to why everyone is listening to a unknown source ( tutorialvideo ) about the chip expectation dates.
Friedcat said February.  I think it better to listen to him, until confirmation of the other date.

Are you another angry AM shareholder?


If you think that AM can manage to get a whole batch in a couple of days after receiving the chips you are lying to yourself.

Using a loan to suppress the price in order to squeeze out the competitor is a common practice in large enterprises. They could operate years with a negative earning to drive out competitors with weak cash flow (they either take loan or have earnings in other area to cover the loss in this area, if they think it worth the effort). It is not for profit in tomorrow or next month, but for a profit in 5-10 years

From banks point of view, it is even more possible, since bitcoin is said to be their competitor. Why shouldn't they take over competitor's infrastructure using money out of thin air, and have total control over it so that it will not make trouble for them in case something went wrong?

Because this is Bitcoin!  Still an experiment.  It's not drilling for oil.

the problem is johnj is right   killing off btc is in the interest of banks and flooding the market with cheaper and cheaper and cheaper gear is not costly.

Still all and all I can spend 400 a month on power as long as I feel like it.  So if this comes out in april I will get one.

IMO it's the other way around. Expensive/overpriced hardware is killing the price of BTC.

When the hardware is sold with a massive margin, there's no way for small/medium scale miners to compete with manufacturers.

So all the people who would likely hold their BTC are forced out of the game by manufacturers who will likely dump their btc to reinvest in hardware/next gen development.

I'm not sure that argument holds water either. Margins on mining equipment have been steadily going down, yet the price is also continuing to drop.

Regardless of what the price of mining gear is doing, the market forces that determine the price are much bigger than whether half of the 3600BTC produced per day gets held or it all gets sold. A year ago the 12.64M BTC mined were worth US$10.55B. Today, the 13.74B BTC mined are worth less than US$3B. That's more than just the extra 1.1M BTC pushing down the price.

Of course there are tons of factors that affect the exchange rate and hardware markup is probably a tiny one. I just don't see how overpriced hardware can have a positive effect on the exchange rate.

Friedcat isn't suggesting to undercut the market, he's projecting likely market prices at the time of release.

I think we will see further price dropa in coins and gear.  by this summer all .5 watt gear will be worth zip .  and the beat goes on. 

In the meantime.   Mr Fried Cat

  A) how abount a miner with a good controller  at least 3 pool choices with a simple failover option
  B) also not crazy loud
  C) not a fire hazard like some of the other gear you put out.
  D) I have owned everything you sold and plan on owning this.

That only works in a closed industry with you having the only source of cash. There are at least 5 mining companies in the industry with heavy heavy VC investment and most of those can raise more if required. You can't empty a sink if the tap is still running.

Also not a botched stratum implementation. That would be pretty helpful.

You've been around considerably longer than me Phillip, but I believe we are approaching the point of diminishing returns.  The improvement in miner efficiency is already slowing down, and is bound to continue to do so. Mining chip technology is rapidly approaching current processor technology.  We are bound by physical constraints, and it doesn't seem realistic for a niche market (in comparison to major processor manufacturers) to overtake companies like Intel in technology, even if ASIC chips are considerably more simple.

That is what i'm trying to tell them ::) it's like trying to reach the speed of light, the faster the harder!

Plus, check out btc price!

No, another happy shareholder.
I guess I missed that bit of info.
On the bright side, I found 1 of your posts that was actually usefull, congrats.

Is the datasheet available for the BE300S ?

Maybe banks already commanded more than 50% of hash power, just we don't know. Most of the actors in bitcoin mining space are short sighted and they only want to make some fiat profit, so banks will take it over easily

One benefit of the bitcoin is that it can solve banks headache of endless monetary expansion: Unlike house/national debt, bitcoin can absorb any scale of fiat money inflow without a total crash, thus the bubble of bitcoin will always become bigger after each downturn, and banks need such endless bubble to make themselves rich

However, if bitcoin is going to replace the fiat money as major payment medium, then fiat money will have no use, that will not be welcomed by banks

i would tend to argue this bolded statement: its no headache.. its their privilege.
+ you are wrong thinking mining companies are short sighted. on the contrary, they know they are the next payment processing industry (cf. bitfury).

thats a good point.  does anyone know if asicminer has 'taped out' and produced masks for the be300?  ie, is it in production?

the samples that have been shown so far are from a shuttle, and thus they wont have taped out, paid a multimillion NRE and produced masks to make those samples.

Last info was that first batch of BE300 will be done in February. They must be in production now.
https://bitcointalk.org/index.php?topic=833704.msg9678661#msg9678661

thanks... its not quite the same thing.  do we know who amhash is?  do they represent friedcat or are they a customer?   all we know is that someone (who wasn't friedcat) said in november they would have chips back in feb.  but until we hear it from the horses mouth or from someone definitely representing him, then i don't think we know for certain that they've taped out and will definitely have the chips back in feb.  or whenever.   all we know is they continue to test the samples, but the samples aren't from a production batch and are independent from that.

 

If you read relevant threads, you'd see answers to your questions.

thats cryptic and not exactly helpful.  have you seen a tapeout announcement or friedcat recently stating availability of production chips?

Tldr AMHash = ASICMiner

anyway, my point is that unless you hear it from friedcat that they've taped out, why would you assume they've taped out yet?  no where does it say it.   and even in the post from amhash in nov that says silicon due in feb.. that was presumably before tapeout had occurred so theres a lot of uncertainty in that statement and its datapoint is old and stale.  it costs a lot of cash to tapeout, and with bitcoin at the price it is now, the economics to justify taping out are stacked against you.  sure you could do it, but would you spend millions pushing that button in the current climate?

They got those chips about a week early it seems so they seem to be on track and we've heard nothing that would suggest any delays.

those are engineering samples.  they come from a set of wafers called a Shuttle - or MPW - Multi Project Wafer.  these are wafers that contains lots of customers samples on the same wafer.

They're not production wafers, and the huge NRE will not have had to be paid for those samples.

For the real production, a multi million dollar NRE will have to be paid, plus an order for several 'lots' of wafer production (also multi million dollar).   its a big commitment at a time when the price of bitcoin is very low and the economics of taping out a new 28nm chip are a painful decision.  It'd be surprising if friedcat has pressed the button on this expensive tape-out and production, especially when he knows that bitfury has a better chip (and presumably others) also out in the same timeframe.

Like I said, full production was scheduled for Feb-March and samples were early. There's been no indications of any delays so why would you assume that the schedule has changed?

Do you have any evidence to show that BitFury has a better chip that will be out in the same time frame? BitFury (and others) said that they'd have 0.2 J/Gh ASICs (based on simulations) in a similar time frame. Have you seen any evidence of that? We know that AM have had 0.2 J/Gh chips since Dec 10.

i have no evidence, only hearsay.

we do know, as you say, that AM has samples back that are 0.2 j/gh... but as mentioned, we don't know when they taped out (nor if they have), and production will be approx 2 months after they tapeout and pay the NRE.  they could still in theory deliver march, just, if they tapeout this month.  but unless they've already taped out in december, the chance of them having production silicon in feb isn't too good.  and now its march you say, someone else was quoting 'amhash' as saying it was feb.  both dates are possible if they've taped out.  so do we know if they've taped out?

The AMHash quote is from a later date than FriedCat's so it could be based on more recent information. Like I said, there's no reason to assume that the schedule has changed from a chip-out date of Feb-March. They got the chips on Dec 10 and testing was supposed to take 3-10 days.

Asic Miner will be late not BFL late but late as in 2 to 5 weeks.

Zero evidence on my part but they have had a pattern of a little late more then once.
So I figure history will repeat.

Hoping this gear does not over heat and run crazy hot. I want to have one in my hands sooner rather then later.

So as most of us I guess.  ::)

How can you have sample chips without a tape out?

They have produced samples using a multi project wafer. They haven't had to pay  the high nre for the masks to be produced.

A shuttle run (mpw) costs only hundreds of thousands of dollars as the costs are shared with lots of other manufacturers. Each customer only gets a few samples of their chips.  Maybe a few hundred or less.

Once the samples are tested and everyone's happy with their chip. And no further changes are made then the nre is paid and full mask set is produced. That costs ballpark 2-3 million and the first order for a bunch of wafers probably also could cost 2-3 million so the initial order with nre and mask set could easily be a 5-6m cash payments (in advanc) for 28nm. Thus it's unknown at this time whether asicminer has taped out and paid the nre for the masks or for production wafers to be made. All this takes 2-3 months.

In bitcoinland many ASIC companies havnt bothered making a shuttle and have gone direct to production. Higher risk but it has usually worked. 

Tweaking/scaling up the mask happens after tape-out because without a (preliminary) physical design, sample chips can't be fabricated.  The RTL code can't go straight on the silicon!   :D

I believe they taped out the final design of the BE300 in mid/late december and it's already submitted, the masks have been fabricated and the chips are in production right now. Come on, FC knows his job, man :)

I think that tape out was done earlier. SHA256 chips are easy design (FC already designed few) and you can go straight to mass production without sampling. Sample chips were only an answer to how much these chips will be off specs and thus how much to order. FC learns on his mistakes and this time didn't ordered tons of chips that are 2 times of specs... Samples proved that chip is even better, so he could finally put a number into ordering form...

could we get it over with the "i think/hope/blah" rhetoric regarding AM?

Hellooooo FC what's up?!

Jutarul maybe? any tangible infos?

 :-[

this isn't a case of friedcat not knowing what he's doing.  Of course he knows what he's doing, which is WHY he might not have placed a huge multi million dollar order at a time when bitcoin is very low priced.  He may be waiting to see what happens to the price of bitcoin, and he may also be tweaking the design, after seeing the sample chips, to make it as good as it can be before he hits the button on producing millions of them.

im not in any way accusing FC of not knowing what he's doing.  quite the contrary, he KNOWS what he's doing, so he may not have taped out yet.  The economics for spending millions making a new bitcoin mining chip are a much tougher decision today than ever in the last 18 months.

If the man in the street can buy bitcoins, at retail, for less than it costs to mine them... then you've seriously got to question whether its worth producing mining chips and other equipment.

and its a different decision for cloud mining as the economics are different if the product isn't intended to be sold at retail and can be operated in a low cost geo location (iceland etc) and host them them at scale to bring the cost down.  but then, everyone is not keen on cloud mining right now as there are too many scams and not enough evidence of actual mining going on.

it's just a case of the usual schizophrenia...  between fc's announcements... his announcements are meds against certain forum member's schizophrenic panics and FUD.
the longer fc spends time working and not posting here (on every demand), the more numerous are the outburst of schizophrenia here... it never fails.
now, i'm not diagnosing or labeling anyone with schizophrenia, just using the word to describe crowd behavior here... so don't take it personal... just making a point...

sounds good ! hope i get my  moolah in time to invest in some !

its not binary.  there are plenty of other options.   for instance designing the best possible chip that could be less than half the operating cost of the present sample.  other asic designers may have demonstrated power numbers well under the 0.2 J/GH of FC's current sample, but not in the same timeframe... so FC might be thinking one of the many options is to go with what he's got as its good enough and matches some of the announced competition's spec, but equally he could be thinking its worth taking time-out while the price of btc is low and there's no hurry on the next generation to come out, and he could take that time to design a more efficient chip that might take a few more months, and that would be better suited to our present market conditions, etc.

its also incorrect to say that either the price goes up or it dies.  there's plenty of business opportunities that don't care about the price of bitcoin to succeed, and most people on these forums always say not to worry about the price of bitcoin.. for instance, one of the near-term potential businesses is the remittance market.  a specific example use case: lets say there's a nanny from the Philippines who works abroad and wants to send money home to her family.  right now its western union or some other expensive money sending options.  a specialised remittance company could utilise bitcoin transparently as the payment rails, and offer her the facility to send person to person money across country boundaries, by matching the transaction via an exchange at both ends, and allowing the nanny to send dollars in one end, and have Philippine local currency come out at the other end for their family... and yet, the nanny hasn't had to know anything about bitcoin, nor even be aware it was being used by the remittance service company... and said company has done all of that for her, and taken no currency risk nor had to hold the bitcoins for longer than milliseconds...  just long enough to do the exchange conversion at each end.  the result, much lower cost remittances without any knowledge of what bitcoin is for either sender nor receiver.

Did anyone have an estimate of the efficiency at the wall? Will be about 0.3W/G?

Tks

Depends on a lot of variables - miner topology (string vs VRM vs hybrid), volt/clock setpoint and PSU efficiency to name the most significant. I'd say anywhere between 0.2 and 0.4 are possibilities.

No updates no estimated time?

HT

This has been asked by others and me before, any info yet ?

I've been unable to raise them for quite some time now. They're ether super busy or hunkering down.

Most probably will came online in some time with the miner ready to be sold.

Juan

That's how they usually do it.

So we know there are issues going on with ASICMiner at present. What's the word on the existence of these chips?

Mystery in Bitcoinland…. the disappearance of FriedCat

http://www.btcreporter.com/2015/03/04/mystery-in-bitcoinland-the-disappearance-of-friedcat/

Can anyone confirm/deny?

Recent events make you wonder if there was ever any independent verification of the success of BE300S.

He seems to be paying for  his cloud mining adventure. I don't think we are going to see much if anything from his group in quite a while.

 

Source?

Sorry I mistyped it . I  corrected typo .


I am  still a bit hungover from my wife's birthday party.

 We're more concerned with where you got your information.

PS Happy birthday wifey!

There was this posted it lee's group buy. Someone also said that friedcat goes quiet to try and solve issues before he comes out with the whole story.

I can verify I heard this too from an inside source...the project is on hold until FC reappears