Bitcoin Forum

I'm branching this off from the old group buy thread, as that has a lot of posts that just aren't that applicable to the actual board design.

Project Description
This project is being developed by Mr Teal and Chip Geek, with the purpose to provide a high performance mining platform for the Butterfly Labs ASICs. By leveraging existing infrastructure in the PC space, the plan is to provide the chips with high levels of thermal and power headroom without spiraling cost out of control.
Once the design is proven, orders for the boards will begin. Chips can either be ordered through myself, or you can order chips and have them resent to us to have them populated onto miners.

Product Description
The final design is still under development, but it will broadly incorporate the following features.

• 8 chips - nominally 30GH/s, extra hashrate depends on grade
• Dynamic clock and voltage control
• USB control, 12V 6pin PCIe for power
• Hole spacing designed to allow many standard CPU coolers to cover the ASICs
• 3 or 4 phase VRM capable of providing well excess of 100A to the ASICs
• Each module will be designed to be run stand alone, or chained together through a single USB port
• Preliminary size ~ 100mm x 150mm, standard 3.5" HDD footprint

Pricing is still up in the air and will depend on many factors that will be clarified in testing, but a good estimate would be $300 for an 8 chip solution.

Test Board Description
A test board has been designed to test the BFL ASICs, and is slated to be done manufacture July 18th. It incorporates a 3 phase VRM, many extra test and debug features, and space for 4 or 1 ASIC.
http://i44.tinypic.com/b7j4nt.png
The primary aspects that testing will focus on will be:
1. Validating the design and the power supply, testing the current limits and temperature rises.
2. Ensure the microcontroller maintains safe control of voltage with no unexpected excursions.
3. Test the microcontroller, verify the code base and communication with peripherals.
4. Test ASIC communication, and hashing.
5. Characterize the BFL chip, examining the chaining of Done signals, whether the onboard temperature diode works, and external clocking.
6. Test board chaining using the Zlink (chaining) interface

Project Status
Done:
June 1st - BFL announces chip sales, basic design ideas begin
June 16th - Group buy initiated
June 22nd - Group buy order placed
July 3rd - Sample chips shipped
July 12th - Samples received, final PCB review
July 14th - Test PCB sent for manufacture
July 18th - Estimated PCB ship date
Upcoming Estimates:
Week of July 22th to 28th - Begin building and testing of PCBs
Week of July 29th to August 4th - Continue testing, incorporate results into the production version
Week of August 5th to 11th - Begin taking orders for PCB sales, send out a small quantity of production boards for manufacture, finalize assembly information
Week of August 12th to 18th - Receive and build production test boards. Test, and if necessary, modify and respin board
Week of August 19th and beyond - Prepare for production, further testing and rework if needed, extra lead time. Final PCBs produced. Components ordered and stocked.
Mid September - Chips arrive and head to assembly

Feel free to ask any questions, and I will keep everyone updated as milestones are retired and the project goes along.

hm... interesting.
good luck

Looks promising. Good luck.

MrTeal, this looks great! Keep up the excellent work.

If you have a list of users looking to have units assembled by you (self-supplied chips), please add me to it.

Keeping an eye on this still, hope I'm not too late by the time my liquidity situation eases.

Excellent update MrTeal.

Just one thing, that I hope I did not miss somehwere else... when is the time frame for the remainder of the 50% BTC for those who participated in your chip buy. I have yet to see or hear when I am supposed to send the remainder of my half for the chips. Or is that just incorporated into the final payment you need to assemble and ship a working unit?

Watching

Good but price is very high
I think better price will be at 250 dollars or 200 dollars

great work here!

I like the idea of the 4 chip board, that keeps the heat down to similar levels that CPU coolers were designed for.

Would it be feasible to remove BFL ASIC chips from BFL units and put them on your boards? Your board design may end up being more desirable than BFL's, especially for the clock/voltage control.

I'd call that feasible but not without risk and cost. Kinda like if you wanna swap hearts with your identical twin.

Without special reworking equipment, you would most likely destroy them.

Watching

@daemonfox, once everything is ready to take board payments I'll collect both the last half of the ASIC money as well as the board money (unless you just want the ASICs). No point in doing all that twice, and I don't want to take anyone's money until we've demonstrated a working product.
@jelin, I'm sorry you feel that way. We'll try and get the price down as much as possible, but even if it is $300 that is still competitive per GH/s with other board assembly services our there.
@erk, exactly how we implement 8 chips on a board is still up in the air. I would like to put all 8 under a large heatsink simply because it will keep the size down, but we'll have to see how well the chips respond. Getting two LGA1156 heatsinks on a 100mm x 150mm will be impossible.
@keefe, you could remove them, but you would need to reball the ASICs before resoldering them. I don't think it's worth the risk.

Any chance you will offer cooling installation as an option?

ps-I doubt it but wonder if anyone will offer to provide a hosting service for boards like these...

Probably. Just because of how mounting the heatsink works, some LGA1156/1155 heatsinks won't work. The ASICs only sit ~1mm off the board while a CPU sits considerably higher, so some heatsinks would sit too high depending on their retention mechanism. As well, some might have small interface areas or weird irregularities that could negatively impact cooling. To take some of that away, we'll probably offer a standard and tested cooler as an option along with the bare populated boards so people can just plug in and go.

I also plan to test the idea of a copper shim/heatspreader to get the height correct so that pretty much any cooler would work. Having another interface will degrade performance a little vs one, but it might be acceptable tradeoff to some people.

I'm sorry to hear that about the height of the chips. I would prefer a design that would allow any standard CPU heatsink to be used. Very interested to see the results of the shim/heatspreader idea... I expect it will be less than optimal but I'd recommend that you make this option available as it would make mounting a water block much easier for those who would like to go that route...

Have you considered a custom designed attachment kit? Perhaps something like a back-plate with a few simple bar springs to apply pressure? I'm thinking that a thermal pad interface is going to be necissary but it would be awesome if you could figure out a way to have a more direct connection using thermal paste...

If you do indeed find that a custom heatsink is necissary, please go with a copper base plate with heatpipes transferring the heat to a large aluminum radiator with fan. pretty much the standard premium heatsink design and in my experience it is extremely effective.

Thanks again for all your hard work, this is a very exciting project and I will be watching with great anticipation.

So would I, but it's just not possible I'm afraid, at least not without a shim or a custom bracket. That being said, there's a lot of great coolers that should just work. I grabbed these from Tom's since they had a 10 cooler LGA1156 roundout that shows great pics of the mounting mechanism.

Arctic Freezer 7 - Would need a custom bracket.
http://media.bestofmicro.com/A/Q/236258/original/arctic-cooling_freezer-7-pro-rev2_installed.jpg

CM Hyper 212+ - Should work, I actually bought one to use
http://media.bestofmicro.com/A/V/236263/original/cooler-master_hyper-212-plus_installed.jpg

Noctua NH-D14 - Should work, I have one and will test
http://media.bestofmicro.com/B/0/236268/original/noctua_nh-d14_installed.jpg

Scythe Mugen-2 - Looks like it would
http://media.bestofmicro.com/B/5/236273/original/scythe_mugen-2-rev-b_installed.jpg

I won't show all the pictures, but of the others the Thermaltake Frio, Tuniq Tower 120 Extreme, Xigmatek Thor's Hammer look like they'll work while the Thermalright, Sunbeamtech and Zalman coolers wouldn't. There won't be a custom heatsink; there's no need since many CPU coolers will still work and there's no way I could offer the performance of something like the Hyper 212 Evo for the $30 that Newegg wants for it.

Would a Hyper 212 eve work? It's a direct heatpipe design, which doesn't really work for a multiple chip design unless you've got some sort of heat spreader between the chips and the heatsink. Otherwise, some chips could be cooled better than others.
http://www.maximumpc.com/files/u139222/cmhyper212evo-02-big.jpg

The Noctua looks like it isn't a direct heatpipe design, so it might work better by way of not having to use a heat spreader first.
http://xeemod.com/wp-content/uploads/2012/10/noctua_nh_d14_3.jpg

I'm a little concerned about some direct contact heatpipe designs since there often is a gap between the heatpipes and the aluminum plate between then. That doesn't seem to be the case with the Evo as there isn't a gap between the heatpipes. The copper heatpipe area width is ~28mm, so it's likely that there might be a millimeter or two of the outside dies contacting the aluminum  on the base plate instead of the copper, but the difference between the two is going to be inconsequential compared to the thermal resistance you'll get from your interface between the silicon and the heatsink anyway.

Ya, I realized the Hypte 212 isn't a bad one, but some of those direct heatpipe ones that work for a CPU would be devastating on an 8 chip design! 

http://img.hexus.net/v2/cooling/cpu_coolers/Xigmatek_Achilles/basefinish-big.jpg

I wouldn't expect very good performance, that's for sure. Again, the plan is that it will be something that's compliant with the mounting hole locations for an LGA1156 heatsink, but things like clearance and having a large enough base plate will be hit or miss. I'll do my best to provide a reasonable selection of coolers that will work with minimal or no tweaking, but it won't be something where you can grab any CPU cooler and toss it on there. There has to be a balance between providing options to people who want to try high end cooling or water blocks, and making the board larger and more expensive.

mrteal, can you look at my PM to you?

I just realized that I never posted this here. Must have something to do with getting home after working on it at 4am yesterday.
http://i43.tinypic.com/2cdtpp5.jpg

PCBs are delivered and two are populated, other than the precious ASIC samples. I plan to build two more tomorrow (one with an ASIC mounted, one without) and ship two to Chip Geek for firmware testing while I characterize and test the power supply and tweak the power supply compensation loop. This project might take a little longer to get running once hardware arrives as it's a different MCU and there are several other features to add (external clocking, onboard temperature diode reading, dynamic frequency and voltage control, custom chaining interface, etc), but so far the chance of sticking to the timeline seems pretty good.

Very cool.  Looks a lot simpler than the jalapeno board (below).  Do I understand correctly that this is a test board that will run with 1 or 4 chips (not seen) but the final board will accomodate 1-8 chips?

ps-will be interesting to see what hashrate you get with each sample chip given the range advertised by bfl...


http://mineforeman.com/wp-content/uploads/2013/03/Jalapeno1.png

Great

But the board take final 4 or 8 chips?

Hi, could you give an est price for the populated board with chips thru you?  I'm not a part of a group buy.
Thx

Jelin, the final board will take 8 although you can mount however many you'd like on it.
iikun, we're still estimating $300 (plus chips) for a fully populated 8 chip board.

OK, I claim dibs on part of the first production run - let me know when I can buy 10 or so of them. :)

I presume $300(plus chips) might be a typo. since 8 chips already costs $500.

$300 for board + $400 for chips (using chip credits) + cooling ($30?) + psu ($50?) = $780ish?

Chriss99 at BFL forum is also designing a board for BFL chips but I think MrTeal is a little farther along.

*tingling sensation detected  ;)

I think BFL are going to have to halve their chip price to remain competitive after Oct when KNCminer and possibly Hashfast are shipping.

MrTeal - anything new happen in the last week with the boards??

This looks very interesting.

A lot of hard work has gone into it so far, can't wait for the final product.

Seems to be a better board compared to the BFL Jalapeno as dwdoc mentioned.

i would laugh so loud if BFL will deliver chips on time and go over avalon  :D

FWIW, outside of these particular intended purposes, the Noctura products rock, I've always used them, and they are super quiet. You probably know this though, but in case you didn't and ambient noise matters to you...

Oh, I know.  It would just be too spectacular for words. 

I'd hate to see it only because of the amount of great work put into the Klondikes and the whole Avalon-chip open source infrastructure, but, increasingly, team Avalon needs a good, solid, Fuck You, and I can't think of one with more iron-clad irony that that.

Sorry for the lack of updates, I was in the mountains and out of easy internet access over the weekend. Testing is still underway, and coding/testing for all the peripherals is in progress. Most of the low level ASIC code is done and looks good on a scope, so hopefully we should be fully hashing and tested this week, and get the 8 chip board out to fab so it can be built and validated for production.

I am guessing they will start with 1 month delay, then 2 week delays, with 2 week intervals :) Today, BFL doesnt seem to be able to produce chips fast enough for their own devices. I feel it may be a miracle if they can ship the chips in time. But, if they can ship, then chip projects would produce products quicker than BFL and that may anger their pre-order customers...

What makes you think that "BFL (hence Globalfoundries) doesn't seem to be able to produce chips fast enough for their own devices?"

I am under the impression that delays there are due to other part supply and assembly issues.

Well this seems like an interesting project so will observe
Bet you guys will be faster than BFL at their rate
I wish you the best of luck on your progress and will be watching as the milestones are reached

Sorry for this mini threadjack, but is there anyone still buying BFL chip credits? I'm trying to get rid of mine for 0.00625 BTC each. Send me a PM so it doesn't interrupt the thread.

a) this really isn't the place for this and

b) the going rate is much much less than that...

EDIT: Unless I misread that, and it was supposed to imply that you're selling multiple credits for ~$10, in which case that's about right if the number is >4

Watching - If your successful perhaps you could get BFL to transfer my orders over to you.  ;D

Awesome - will watch it too as we have 128 chips ordered.  :)

It's more of a delay than that - BFL is quoting 100 days from order date to delivery.  In BFL time units, that's around February 2014.

Another quick update. I've made some good progress on the DC/DC converter, I think most of the issues are now resolved. I'm now going to start tweaking the control loop and doing some load and transient tests at which point I can finish that part of the circuit for the final design. Meanwhile, work on the firmware continues.

So you stopped burning 7805s now? Awesome  ;)





(Just kidding, translation for the non-geek, 7805 is a linear power regulator that you'd never use i) Since 1990*, ii)for anything more than an amp iii) if you cared about conversion efficiency .... *okay, they're maybe still handy when you just want something quick and dirty.)

I went with the 7905 since the number is higher. I think the 7805 is last year's model.

LOL good plan :D

That 7905 is going to be a big negative....
.
You should consider using something more sturdy than those flimsy semiconductor things....
http://www.ebay.com/itm/ws/eBayISAPI.dll?ViewItem&item=271247474477

Thanks for the update, looking forward to hashing demos soon

Time to market is everything IMO.  Why do an 8 chip board instead of getting this 4 chip design production-worthy?  Seems to me that the 4 chip board that you already have would be production worthy sooner.

What's the rush? Chips won't be delivered until late September at the earliest.

I am a cheerleader for Lucko, MrTeal and chris99.  Would be nice to have more than one option for BFL chip boards if possible...

I must say that MrTeal is really nice... He was ready to send us chips to help when it looked like ours are stuck in US... But it was only tracking... Received my today and say thanks to him just to find out he had things ready to send me samples...

I'm think chriss99 in UK could make use of those sample chips.  He just posted board pics but he has a shortage of sample chips for testing.  If there are any sample chips in USA I would be willing to buy/ship them from Houston. Nice to see a few different board options coming along.

https://forums.butterflylabs.com/bfl-forum-miscellaneous/3831-board-bfl-chips-group-buy-chips-9.html#post50833

in case of board project on BFL-forum i would be less enthusiastic until they post some actual photos of the board. and some more trustful info maybe.

strange is the sequence of the latest posts and explanations about BFL-logos on the posted photos there... after getting PCBs from manufacturer?  ::)

so, no comment for now except of this one...

Yes. Chris99 says those are modified jalapeno boards for testing with their 8 chip board design pics pending next week. It would be nice if he, MrTeal and Lucko all had working boards we could choose from.  Also, Lucko is offering a hosting service which is a nice option...

Just checking in to see how things are going?  Do you still think you will have samples up and running by the end of the week?

They're coming along slowly. The work on the high current DC/DC supply is progressing much slower than expected, so it's holding things back in some ways. Other aspects of the design are being worked on, and some work has started on the next revision of the board.

Thanks for the update. Keep us posted!

Just a bump to push this one back up.

Mr Teal, How is everything progressing. Thanks!

Slowly, but going decently. We're talking to the chips, though we're still working through a few issues communicating with a vanilla build of cgminer using the BFL communication protocol so we can live hash on the network. Rewriting the mess of firmware BFL released is a bit slow going, but it will definitely be worth it once it's done.
I've moved on to now working on the design and layout of the 8 chip board so we can get it sent off to fab.

We're rooting for you!

Thanks for the updates, really looking forward to the boards. Now if we can just get BFL to hurry up.

I've been waiting over a year for my Singles. :(

And we're hashing. There some interesting things that are coming out in hashing, I'm really excited to push the chips and see how they react even if they are the dregs of BFL's barrel.

In other notes, other features are looking like they should be working great. It appears that the on die thermal diode is functional, so we should have per-chip on-die temperature measurements, which will be much more accurate than a thermal sensor just placed on the PCB.

The latest board revision is almost done and should go off to fab this week. The major changes are dropping the onboard external clock generator which we aren't using, as well as moving to 8 chips. It will also feature a major revision to the power supply to solve some problems and ensure it's exceedingly stable.

Great news!

Let's face it, complaining about delayed BFL chip delivery won't be as much fun without having some cool boards sitting around!  ;D

I am looking forward to the Teal vs Luko head to head comparison one of these days!

MrTeal vs. Lucko vs. Chris99 vs. any others?

Terrific job  8)

Must admit I am curious, will the rewritten firmware also work on BFL boards?

No. We're using a completely different MCU, and the hardware is very different. Some features (like the temperature diodes) require additional support circuitry that isn't on the BFL boards.

Fair enough, hopefully we'll see some of the innovation eventually merged back into the BFL firmware, if anyone has time to write the code!

I'm curious too. Did you got engine 0 working. I'm pretty sure it is a hardware problem but still interested if I'm wrong...

That is extremely awesome! A big well-done and thank you!

Evil, but quite right!

Unfortunately the chip I have mounted is terrible, 8 engines pass the test cases. It's pretty apparent that the samples shipped appear to be the rejects that didn't pass QA for the BFL products. Those that do run do so pretty well, running at 330MHz ~2.64GH/s. Testing continues...

That's pretty much in line with what Lucko discovered.  It seems the sample chips are likely those that didn't even make D grade and would have been discarded.

Not yet. We might do some firmware fiddling to see if we can get it to work, but enabling E0 just causes it to spew errors.

Same hire... And I fell like doing search for God. There is no way you can prove that he doesn't exist. You can only prove where he can't be... But get it to work would mean that you get one free chip for every 15 chips... Not such a small gain...

But since BFL given up on it I would guess it is hardware error...

I thought it was mentioned that engine 0 was not working on the first run of chips and it would be remedied in subsequent runs

I have seen Conman mention that if engine 0 is enabled all the other engines do not work. That was awhile back, and maybe things have changed but I am sure he knows

That is both interesting and plausible... I wonder how this fits in with the chip orders - has it been confirmed that they are being made in a separate run?

I think whether or not it is fixed depends on what layer the error is on. It is not unrealistic to expect that they would change the metallisation layer to enable the engine, as this gives a ~7% boost in speed, but there is no way they are going to have a full mask made up to fix the problem. Has anyone read anything about this? Is it known where the error lies?

Because some people wanted pictures. :)
Test setup
http://i40.tinypic.com/2n6t3lg.jpg
Running at 363MHz (11.36MHz * 64 for the PLL clock, with 2 divisor)
http://i44.tinypic.com/20qnkg3.jpg
Putting out 2.9GH/s on 8 cores
http://i42.tinypic.com/2884j06.png

Very awesome! Keep up the work

Nice work MrTeal, congrats.

This sounds very exciting!  I'll keep an eye on this thread too  :)

363MHz is incredible! Did ChipGeek manage to get similar clock speeds on his board with 14/15 engines running, or is this only possible because of the much lower TDP due to the low engine count?

Isn't that 2.4?With less working engines you can get higher frequencies but even more engines will fail tests... So hi frequency is not necessarily good. We got it as hi as 428 but got less out then at lower frequency... You need to find a balance how many engines will work at how many HW errors and frequency...

Feel free to ask any questions, and I will keep everyone updated as milestones are retired and the project goes along.

Was wondering where in the milestone list the pictures above are and nice setup :D

Project Status
Done:
June 1st - BFL announces chip sales, basic design ideas begin
June 16th - Group buy initiated
June 22nd - Group buy order placed
July 3rd - Sample chips shipped
July 12th - Samples received, final PCB review
July 14th - Test PCB sent for manufacture
July 18th - Estimated PCB ship date
Upcoming Estimates:
Week of July 22th to 28th - Begin building and testing of PCBs
Week of July 29th to August 4th - Continue testing, incorporate results into the production version
Week of August 5th to 11th - Begin taking orders for PCB sales, send out a small quantity of production boards for manufacture, finalize assembly information
Week of August 12th to 18th - Receive and build production test boards. Test, and if necessary, modify and respin board
Week of August 19th and beyond - Prepare for production, further testing and rework if needed, extra lead time. Final PCBs produced. Components ordered and stocked.
Mid September - Chips arrive and head to assembly

No, it's 2.9. The one showing 2.4GH/s bounces around all the time depending on when shares are discovered, but the average was 2.9GH/s.

Incredible job, getting 3 GH out of a reject chip is better than I expected. imagine what you will get out of better grade.

Unless you have a build of cgminer that has this turned it is the other way around first one is showing 5s average and second is average...

EDIT:
If that was the case I had 10GH jala(with our firmware)
http://www.prorang.si/jala.jpg

But the pool say I don't. It has up and downs but it average on 8,5 to 8,6...
http://www.prorang.si/jalapool.jpg
Still not bad since stock 1.2.5 got only 7.3 out... And modified got 7.8

So it should be 2.4 not 2.9
http://i42.tinypic.com/2884j06.png

Ah, you're right on the order, my mistake. If you'd really like I can take a video of it hashing on 8 cores at 2.9GH/s on Monday, this time making sure I zero the stats after fiddling with things. I'm not sure why you find it so hard to believe that it's producing 2.9GH/s though considering the cores aren't giving hardware errors and that is the expected hashrate for that clock speed and number of engines.

I have no problem believe that. I'm just saying I don't see that... I know you get 1MH per core for every 1MHz... So yes it should be 2904MH

But I must say your CGminer is realy strange. Data on top and at device should be about the same for only one device... Unless you pluged it in at later date and mined before that. But since you have same number of accepted shares this is not the case...

Awesome!
Congrats MrTeal...  ;)

Someone on the BFL forums asked for pictures, so I grabbed a couple in the middle of testing. I'm not sure exactly what I was doing at the time. I don't bother restarting cgminer, I just flash the device; it drops out and gets redetected once programming it complete. That's also why it shows as BAJ2 when it's the only one in the system instead of BAJ0. As for why the shares are the same, I'm really not sure. I won't be back at the office until Monday so I won't be able to test flashing it in the middle of hashing to see what's going on with the number of shares.

Don't get me wrong... I do believe you(it would be in line with what we learned) and I know how you got BAJ2 since I got up to 12 :) . But I don't get it how can you get different data in summery and device with only one device send all shares... If any device should be higher not other way around since data for the device is from plunging in the device and summery is from cgminer start... And it might get someone else confused as well...

And sorry for forgot to say good job!

Roughly 30 days till the BFL chips show up (If they keep the 100 day delivery). Do you think the PCB and design will be finalized and ready to go? Thanks, jborkl

Do you think the chips will arrive on time?  ;)

Though joking aside, I'm genuinely curious about what people's opinions are on this... There are arguments both for and against, so it would be interesting to see where people stand on this.

Great work, MrTeal!

BTW, how shitty were the samples I sent?   :D

LOL... Don't know yet. The 8 chip board is being fabbed as we speak, so I should know more one they're built.

Did mine arrive with you yet?

Hi MrTeal, How are my sample chips working with testing boards? Are they also defect ICs?

To be honest, I haven't been segregating them and marking which samples are my original ones vs which were sent to me. They're all mounted to boards in whatever order they're sitting in trays.

That's fine, I just wanted to make sure that BFL actually sent mine to you, otherwise I'll chase them up! Do you know if mine ever arrived?

I do. The 8 chip design is in fab right now, and while it's not the exact version that will be in production (there will be less debug headers and whatnot) the production version is being finalized right now and will be sent off pending confirmation that the changes made for this board perform as expected.

Some of the changes include switching to TI NexFET integrated driver MOSFets (with peak efficiency around 100A for the 8 chip system and max current of 160A), switching to a 6 layer PCB to provide lower inductance and resistance in getting all that current to the chips, and tweaking the layout and optimization of the chip layout.

Yep, they sure did. Thanks!

To clarify one thing, when the __PRODUCT_JALAPENO or __PRODUCT_LITTLE_SINGLE  is selected, the crystal used for the MCU should be 16MHz. However for SINGLE or MINIRIG models, then the crystal
used should be 12MHz, otherwise the Frequency of the ASICs is going to be incorrectly calculated (that is where you get 320MHz speed report, etc).


Regards,
Nasser

Thanks for the reply Nasser, although I think you're mistaken here. There is no Product_Jalapeno or Product_Little_Single, the firmware is entirely custom on an ARM chip. The 320MHz was measured at the clk_out pin as the divided by 64 version of the onboard PLL while using divide by two internal divisor.

Is there an email I can reach you at? I would like to ask a few technical questions about the chip in regards to setting it up to use an undivided version of the onboard PLL as the chip clock. I will PM you my contact details.

Awesome, glad they made it to you! Now I've just got to cross my fingers and hope my chips aren't too far behind...

Probably more than is happening wrt shipping singles...  :'(

Just wondering when the first shipment of chips is due to arrive? Obviously with the emphasis on the word "due", i.e. what is the date of the first chip order + 100 days?

And has there been any communication from BFL about how things are looking for this date? (Come on, a guy can dream, right?)

BFL started taking bulk chip orders 6/10/13. 100 calendar days after that is 9/18/13.

Well, the 8 chip boards shipped from the fab today. Give them a couple days to get here and we should have a few ready this week.

The first group buy I placed was on June 26th, so I'm not expecting chips prior to Oct 4th. The first chips are scheduled to be ready mid September though, so if that date is met we could be seeing custom BFL miners ready by the end of September.

Other little bits of news, the production board waiting to be sent off pending production test includes 53.2mm hole spacing, so you can mount (as long as the contact pad is large enough) aftermarket coolers that fit many ATI GPUs. Pics to come soon.

Very nice! And Josh has just announced that their latest batch of chips has arrived, which is "tons", which if it includes the chips for bulk-buys could mean that they're ready even soon than that!

I believe Joshua is having a chip shortage on the production line so don't count your chickens.....

Although there is $$$ in play (the balance of the 50% payment) so he will probably ship them to raw chip orders as that will generate income as making customers who have already paid wait a few more months won't cost him a dime.

Well it's win/lose, win/lose for me as I have eggs in both baskets, so either way is both good and bad for me. But I agree that something that benefits them whilst throwing their very first customers yet further under the bus doesn't sound to be hugely out of character...

[Speculation]
A play on numbers:

BFL has 75 wafers with a maximum yield of 1,100 chips per wafer at nearly perfect yield.

3000 Single SC units will consume: 16 x 3000 = 48,000 chips

3000 Little Single units will consume: 8 x 3000 = 24,000 chips

2000 Jally's units will consume: 2 x 2000 = 4,000 chips

Total tally so far: 76,000 chips required.

Minirigs are not counted. (Minirigs should require 8 x 16 per box = 128 chips...384 per order)

===================

Question, where does BFL get the rest of their chips if wafer yields are 90% or less? Where do the mini rigs get their chips from?

You, of all people, I am surprised at for asking. From the magic chip fairy, of course! The honest answer though is that they just place another order with the fab, and make their paying customers wait another 100+ days for the equipment they ordered over a year ago. I don't see what the problem is?

The answer to that question will just anger you further.

Looking to get rid of my 48 chip order with Mr. Teal. Paid my deposit already. Just looking to get back what I paid for the chips and chip credits. $1664 for the 48 chips. Bitcoin accepted here.  ;)

Sorry for the late update, we've hit a bit of a snag that set us back quite a bit. The PCB supplier sent boards that had a few internal shorts, which resulted in a couple boards that don't work. New PCBs are in house now, so they should be getting populated today to test.

Was it faulty manufacture or a mistake in the design (if their mistake then tut-tut, if your mistake, then "oh well, these things happen :))

Do they not offer electrical testing on boards? Having said that, I never tick the option anyway as they charge for it and almost never make mistakes, so I can understand why you wouldn't go for it...

Hope all goes well with the respin anyway, good luck!

No, it was their problem with the board. That being said though, it's my bad for not getting the e-test done. For pretty much every board I make I get the e-test done since the client pays for it anyway and I've never had a board fail it, so I didn't bother spending the extra this time. Lesson learned.

Thanks for the update MrTeal! How does that old saying regarding hind-sight go? Oh yeah :)

*Once you guys have finished hammering out all of your remaining issues...do you think you guys will be able to produce anywhere near 400 finished units per day :P (couldn't resist)

**BTW, have you guys decided what your going to be naming your 'baby'?

The "fastest" Butterfly is the Hawkmoth that travels at 30mph

That is my suggestion.

Yeah, that was my attitude, I'd never had one fail so I stopped spending the money on the testing. On the other hand, maybe they failed quite often and they just don't tell you and they respin the board quietly and just give it to you like it worked all along... That would explain it why we've both made the same mistake!

Did you get my email by the way?

+1 :)

My suggestion would be the "Dr. Mrs. The Monarch" or the "Swallowtail".   :D

Amazing! :D

MrTeal I sent you a PM, got it?

There have been, but on the BFL forum. It's worth keeping an eye on the parallel thread over there too:

https://forums.butterflylabs.com/hardware-discussion/3901-bfl-asic-mining-board-project.html

Hope that helps.

Sorry for the lack of updates, I've been way up North with work and don't have internet when I'm stuck underground for 12+hrs a day.
An 8 chip board is with ChipGeek now and he's working on it, though there appears to be a few problems with communications to the chips. It worked prior to shipping, so hopefully it's nothing serious. I'm waiting on getting home on (hopefully) Wednesday when I can build another one and post up some videos assuming all goes well.

Alright, I am back in town and back to work.

Good news today, I managed to salvage two of the boards that were defective from the PCB house. Thank god too, because I really didn't want to lose the one with 8 ASICs on it. It's definitely a kludge, one board has Vdd shorted to ground but a good MCU, the other has problems with the JTAG interface that prevent me from programming but a good VRM and set of ASICs.
http://i43.tinypic.com/2n0qbft.jpg
There were a couple issues, chiefly that the new build of the firmware meant for the 8 chip boards is having some issues and wasn't hashing so I had to hack together some code on the old 4 chip test board to get it to work. It behaves really strangely too, cgminer reports twice the hashrate that it's actually putting out. If I had to hazard a guess I'd say it has something to do with having ASICs 0-3 on SPI1 and 4-7 on SPI2, and my hacked code just has it looking for 8 on each bus. Bitminter reports it nicely at 27GH/s though.
http://i43.tinypic.com/2qdtvet.png
The other issue is that the early code doesn't do a good job up keeping the ASICs feed with work. You can clearly see the dips in output current, which is quite different from the straight line you get when all 8 engines are loaded up at the same time and hashing test cases during startup. Once the issues with the new code base are resolved, that should get resolved. It will be interesting to see what kind of hashrates are achievable with grade A chips and some more optimized code.

The other problem is that the ASICs actually get quite hot, even with the closed loop water cooler. The only TIM pads I had laying around was a roll of 3M 5590H, and while it's decent stuff for most applications the 0.5mm thick 3.0W/mK really isn't effective enough for the kind of heat load these chips put out. I'll have to order some 0.5mm 12-17W/mK material, and I might be adventurous and try it with just grease.

I cut the other stuff even though it's awesome. I'm curious to see how level the chips are, if you're using paste on an 8 chip design there's a pretty good chance at least one of those chips is not going to get adequate cooling.

I was thinking that regular paste might work with a 4 chip design. I'm somewhat disappointed that a 4 chip board hasn't been explored/tested a little more seriously.

OTOH, if you can get optimum results from an 8 chip board, then that is -of course- better..  :)

That being said, I have great confidence in this project, pending the actual delivery of chips. Thanks for all your hard work.

http://www.frozencpu.com/products/17499/thr-181/Fujipoly_Ultra_Extreme_System_Builder_Thermal_Pad_-_60_x_50_x_05_-_Thermal_Conductivity_170_WmK.html

Will the grease work if you put enough to fill the gaps effectively? I was under the impression that grease has to be as thin as possible for best effect, and we know that 8 chips wont be perfectly level.

Depends how much variation in height and angle there is, I suppose. A pad is a better conductor of heat than grease, but grease is (hopefully) much thinner. It's not quite straightforward since the properties (and thickness) of a pad will change depending on pressure, but for a basic estimate you can just use thermal conductivity / thickness. Whichever number is bigger wins.

I'm doing some more testing tonight and actually having a lot better success with cooling. The backplate that came with my Water 2.0 Extreme is basically useless, so I just wrapped a small LGA1156 heatsink in duct tape and mounted it to the bottom of the board to act as a proper back plate. I can get it a lot tighter without worrying about board flex now, so that helps a lot.

Just to let everyone hire know good news:

Congrats! nice to see this project get off the ground.

This in not my order but I would just like anyone to know that impossible looks to be possible. BFL on time :) So you know that you are not wait for noting (like in Avalon case) and you will need MrTeal or any other(my :) ) board...

Uh, you do realize that customer are getting chips that "do" about 2.4 GH/s correct?

If this were a good thing, then no one in the community would be chasing [a single] 28nm ASIC that do about 200Gh/s to 500Gh/s.

The customer should ?*rejoice*? about what again? Getting obsolete hardware at a very high price??

Explain it to me...

Oh sweet Jesus. There is no thread that is Troll free.

PuertoLibre should be banned from the forums, if I were an admin he would be for trolling, he is obvious spreading FUD, on one hand he is critical about a chips that has shipped 100/s of TH/s worth, on the other hand he praises a chip that has shipped none.

I would just like to point out that this still counts as feeding... He's obviously hungry, just starve him out and he'll find a different thread for snacking...

EDIT: God it's hard though, I'm already struggling to take my own advice!

Erk, until recently there were more than 10 people advocating you be removed from the forums. So I would not speak too loudly.

As for the 100Th/s, because those 100Th/s [multiplied by 9] have shipped, the chip itself has become obsolete. I would have said the same thing of an ASICMiner chip or an Avalon chip. Gen 1 is obsolete in the current platform designs. (Obsolete = Nearing or has passed the line of Not being Profitable to run)

Could you revamp the platform to scale these chips massively? Probably. But you would always run into limits of the chips cost vs the electricity cost vs the ROI.

Notice I am talking about substance. Not just a wild claim. You often used to cheerlead on no substance.

=========================

Edit: I am not trying to "start anything" I just read a stupid comment...and reacted to it. Apparently you didn't like that.

People were plenty angry that a 300Mh/s chip sold at ~0.08btc (Avalon) was delivered "too late" just weeks ago. The same kind of logic applies to a BFL chip that was just about equally priced on a per-Gh/s to dollar ratio.

I am only guilty of pointing out the obvious.

It's not that what you're saying is wrong, it's that this is not the place for it.

Many people here are painfully aware of what a shitty situation they are in, but it is what it is, and it's not clear why you feel the need to rub it in the faces of those suffering the consequences of poor decisions they've made.

This is a thread about a board that is being designed to run chips that have already, at least in part, been paid for. Either contribute to this discussion in a useful way or don't contribute at all, but coming on here to rub people's faces in their losses and de-rail a useful thread about hardware R&D is not on.

Alright, then you could have said so.

I'll take it elsewhere, as long as people like erk aren't too dumbfounded by the obvious.

I just reacted to a very dumb comment made by someone else.

Thanks, I will take my point and place it elsewhere.

Fair enough, and it's appreciated!  :)

I, for one, cant wait to see a full production board doing its thing! Its going to be SEXY!

I can't wait either. A small quantity of production boards (50) will be arriving here on Wednesday and will get a couple built up. There's a couple improvements in the production boards (fixing a couple ECOs, improved cooling for the VRMs, new GPU cooler mounting holes) in addition to the new (and better) form factor, but functionally they're basically the same as the existing test board so we're very confident that it will work out of the box.

In testing, things have been going really well. The DC/DC mosfets are a little hotter than I'd like, they're probably going to require airflow over them or heatsinks on the bottom side. We'll see once the production boards get here; there are layout improvements and a move from 1 oz to 2oz copper vs the prototype, so we might be able to avoid needing to heatsink them. With a couple small RAM heatsinks from an old VGA cooler I have on the underside, I'm seeing ~65C MOSFET temps, so cooling temp without deafening fan noise should be pretty trivial.

Other things are coming along. We've had a lot of peripheral code disabled as we debugged some job queue issues, so now we're adding that back in. The quality of sample chips is still hit or miss; I've had one board that's happy to hash away are 32.5GH/s on 113 engines at stock (1V) voltage and staying nice and cool, while another board had chips that simply wouldn't produce valid results from half their cores even at 200MHz.

I took a video of some hashing at 32.5GH/s @ 168W (AC at the wall, random OEM HEC PSU, including the 9W from my cooler) but I'll retake it with a tripod and post it up tomorrow. Oscar quality it isn't.

For now we have a couple pics of less aggressively clocked testing I ran overnight; my normal test board and the aforementioned bad chip board.
http://i41.tinypic.com/724ys7.jpg

Ah, and here's a quick one of the outline of the boards that will appear on Wednesday.
http://i41.tinypic.com/fnakjo.png

Tomorrow I'll post up a quick video of the unit hashing.

Not sure if you might wanna test the MOSFETs you're getting to make sure they actually meet the data sheet spec. Reading between the lines of BFLs (mis)adventures, it seemed like they had that issue. I've also heard rumors of counterfeit parts, meaning you think you got a big name 20A part and it's a remarked POS burning up at 5A.

As I recall the issue BFL had was that the boards traces were not hefty enough to handle the unexpected increased power requirements.

Some videos. The first one is kind of a nominal load for the board. 107 hashing engines, setpoint of 283MHz per engine to give 30GH/s. Everything stays nice and cool, the board draws 130W at the wall. The cooler is a Hyper 212 Evo, and there's a couple little stick on GPU RAM heatsinks on the bottom under the mosfets, with a little quiet 80mm fan blowing towards them. The mosfets are 125F.
http://youtu.be/VOOGvlpNwws (http://youtu.be/VOOGvlpNwws)

This one is what I'd consider an overclock. 1.15V on the chips, with the same 283MHz nominal setpoint. 38GH/s and the mosfets got up to 175F. If you plan to do that constantly you'd probably want to invest in some proper cooling and airflow for them. Pulling 220W from the wall, so a big power penalty for those extra 8 GH/s
http://youtu.be/cCqiCodE12Y (http://youtu.be/cCqiCodE12Y)

I've found  1.1V on the ASIC to be a nice compromise for the future, with 35GH/s and 180W draw and things staying much cooler.

For people who are interested in MrTeal's soon to be ready boards and want to get the product soon, I have 96 chips worth of chip credits invested with original group board/chip buy, 1st and 2nd half of the payments paid. I am willing to sell my share for less than the price I paid in BTC. Basically you get the hardware for 1/4the price of what BFL sells so it is a big win.

I paid ~48BTC for 1st and 2nd half for 96 chip. I am willing to transfer my boards for 30BTC if you want all 96 chips... (this is hashing power equal to 6x 50gh units that BFL is selling for $2500 each, it is a bargain!)

If you are interested in using escrow (although it is quite trivial for MrTeal to transfer my purchase to you), I have put my slot on sale on Bitmit!
https://www.bitmit.net/en/item/61563-380gh-bitcoin-asic-miner-slot-for-35btc

You can PM me if you are interested (I can go even lower on the price but not much more) or buy from BitMit.

THEY ARE SOLD, THANKS FOR LOOKING

http://i39.tinypic.com/x2otp2.png
Since I posted up those videos the board's been going strong for two days at 35GH/s, drawing 180W at the wall.

More details on how purchases will happen will be released by tomorrow.

This new Revision of BFL chips w/ the metal layers seems to be able to hit higher Gh/s but still just as power hungry and hot! 

@MrTeal, how was the board running w/ that Watercooling loop you had on it?  Just wondering if it's worth picking up a decent closed loop system over a $30 120mm HS&Fan setup.  It it lets me get an extra 5Gh/s while being able to use a 80-92mm Fan to cool off the vregs and others circuits then I'd probably go for it on atleast one of my boards. 

Are the Sample chips you're using the new Revision chips by chance?  Do you notice a big difference in what Coolers you're using since they output so much heat?  And last question I promise, have you had any luck using a TIM as opposed to the heat pads?  Hope the newer chips are more uniform as that'd make cooling much more effective.

So far I've only used the rev A chips on boards. When the production boards arrive I will be mounting one with some Rev B samples. Efficiency actually isn't terrible right now; while 130W for 30GH/s isn't great that's a little under the sweet spot for my 850W power supply. I'll probably pick up a proper gold or platinum PSU once I get a moment, it should perform better than the plain brown box HEC PSU I'm running them on right now.

The closed loop cooler seemed to work ok, but I've only tried it with decent (as opposed to great) thermal pads. I took it off when I took the videos because I didn't want to give the impression that you need to use water cooling. Far more important than the cooler itself is the thermal interface between the two surfaces. While specs don't tell the whole story, the difference between a 0.25mm thermal adhesive tape like 3M's 8810 at 7.74°C*cm^2/W, versus the initial pads I was using (3.0°C*cm^2/W), vs something like the Fujipoly Ultra Extreme pads (0.4C*cm^2/W) absolutely dwarfs any differences between a good air cooler and a closed loop water cooler. I just really liked the water cooler since it moves the mass of cooler away from the board and makes probing test points easier.

I've actually had some good conversations with an applications engineer at Arctic Silver who has provided me some literature and samples on their products for my idea to bond a heatspreader to the ASICs so I can use a proper minimum bond line grease interface between heat spreader and cooler. Once the first production boards get built here in the next couple days I'll try that on the existing board I have to see what happens. The per chip temperature sensing is now being output out the debug UART (cg/bfgminer integration is still on the to-do list), so I'll be able to directly compare results between TIM+Hyper212 Evo, TIM+Water, and bonded heatspreader on the same board.

Very exciting stuff thanks!  I'm sure i'm not the only one looking forward to the incoming boards and your testing results.  I can't get enough of this good stuff and the GH/s is just bonus  ;)

But, thanks again for your time and your work and Chipgeek as well.

I have a few decent HS&F I picked up for a great deal and may just go ahead and pickup a closed loop water-cooler if I can find a good deal.  Will be fun to see what they can do for sure.

I am selling my 96chip slot for 30BTC right now:
https://www.bitmit.net/en/item/61563-380gh-bitcoin-asic-miner-slot-for-30btc

As bonus, I can bundle 256 chip credits (although it is not mentioned on BitMit, after purchase you can send me your address to transfer the credits to).

THEY ARE SOLD, THANKS FOR LOOKING

If anyone else is looking to buy - I have up to 248 chips on order that I would be willing to part with - PM with offers if interested.

I haven't seen it mentioned in this thread but have you tried going for just paste and see what you get?

My BFL little single when it originally arrived would clock in at 31 GH/s  (which is pretty sweet for rev A chips) but after a few weeks only seemed to run at 30.6 GH/s.

I said screw it and cleaned off all the thermal pad gunk and applied Arctic Cooling MX3. It dropped temps by about 4C and has ratcheted up to 31.2 GH/s :)

I want to wait until the new boards arrive (should be tomorrow) and get populated before messing with the hardware I have. I don't think it would ruin anything, but I don't want to risk frying chips.

I would have thought that just as a new batch of boards is about to be made up would be the perfect time to fry the old ones! ;)

Joking aside though, perhaps just after the new ones are confirmed to be working would be the ideal time to try something with the old batch... :P

Just an FYI for those looking for decent and Cheap cooling for these boards, Newegg has this Zalman for $10 after $20 rebate http://www.newegg.com/Product/Product.aspx?Item=N82E16835118099 (http://www.newegg.com/Product/Product.aspx?Item=N82E16835118099)

MrTeal, Im still waiting for your reply..

Dang ppl, this project wont give you ROI.

Mr Teal, please looking into having bitfury mining board. Thats what i would buy. Dont waste time (money) into this. The longer delay in having bitfury the more btc you will not mine.

If you're interested in this project and need chips at a great price, I have 100 available (https://bitcointalk.org/index.php?topic=300455.msg3220985#msg3220985) (shipping from BFL this week). The order is from 6/20/2013, so it should be much faster than buying from BFL directly at this point ... which means a lot, given recent difficulty increases.

PM ASAP with interest.

Can I mail in the chips and get them mounted on the boards?

Likewise, but with 248 chips. Again, PM if interested.

Sweet MrTeal!  Who is gonna be able to get boards from first batch?  Gonna be in order of order?   :P   

BTW, waiting for mined BTC to finish confirms and will get the last 50% Chip payment out on Thurday.

Awesome!  :)

Presumably the production run of the boards will be done by the time the chips arrive so it won't make any difference?

True, true   :D

Gonna be fun to have some open firmware boards to overclock w/ aftermarket coolers and watercoolers.  Makes for a very enjoyable project IMO.

Sorry but that photo deserves to be shown again. It's truly beautiful. I can't wait to see it fully populated. (man that sounds dirty  ;D )

Whose design work is that? is that you? or chipgeek? or both??

Also, are you planning to have the airflow moving from the chips towards the vrms? or the other way around? I'm curious to see which need more cooling, seeing as the hashrate seems to go up wrt voltage almost linearly, whereas it seems like you haven't even explored changing the clockrate, but now I'm babbling. 

Very exciting project.

Just curious about if there's a reference design of how to mount 12 8-chip board and setup the cooling solution/power supply in an elegant way?

And since its hash power fall on about the power of BFL's 'Single', are you considering to give it a code name? maybe 'Singular' or 'Singleton'??

MrTeal did all of the schematic and board layout.  I helped with MCU selection (different than BFL's MCU) and the general concept of what we wanted in the hardware.  MrTeal did 100% of the voltage regulator selection and design.  I wrote all of the firmware and I'm still making some major improvements over the next few days.  (Details are expected to be announced next week.)

We have experimented extensively with clock settings and changing voltage.  We have a very good understanding of how that works.  The VRM does generate quite a bit of heat.  These newest boards have heavier copper layers to assist in cooling but some airflow will likely be a good idea.

BTW - That photo is for 2 boards in a "panel".  The assembly house prefers to make two boards at once.  Then, after all components are mounted, the boards are snapped apart.

We do not have an explicit "reference design" exactly.  But since these are able to plug into any PCI Express slot (x1 to x16), you should be able to find PC or server cases that are appropriate for your needs.  MrTeal has located some interesting ones but I'm NOT going to bother him tonight since he's making new boards right now.  :)   ;D

Our board is only 8 chips so it's more like BFL's little single.  We do have a project name picked out.  Look for it coming soon.

not sure about 12, and not sure about elegant  :P  but I'm thinking of building an enclosure something like this.

__________________________
|                   |                   |
|                   |                   |
|                   |                   |
|                   |                   |
|                   |                   |
__________________________
|                   |                   |
|                   |                   |
|                   |                   |
|                   |                   |
|                   |                   |
__________________________

open on both ends with 1 unit inside each compartment, then putting a box fan on the front and just leave the back open for all the wires and PSU and what not.

I'm assuming you can run 4 easy with a 1200W PSU? or is that pushing it?

depending on the size, I guess you could maybe make it 3X3 squares, which would fit 9, but I'd probably block off the middle box and only load 8.

We don't have any hard numbers yet but we believe our 8 chips board will deliver 30GH/s (+/-) at about 150W (+/-). 

Remember, this is more like a BFL Little Single and not a full Single.

Looking forward to getting my 16 chips mounted to boards!!!

Willing to sell my 128 chips for BTC20.

There is $3200 to pay to BFL.  The chips will be shipped before Sept 30.

https://bitcointalk.org/index.php?topic=303224.0

Need to find a buyer today or tomorrow.

Start laughing.  :D

My hats off to MrTeal and ChipGeek both, that is one sexy looking board I must say.

Hats off indeed! Way to go you two!


Any idea when boards will be available for chip mounting?

Thanks guys.  Look for an announcement from MrTeal very soon.  Maybe tomorrow.

And yes, those boards look fantastic.  MrTeal did a great job.  I can't wait to get one in the mail tomorrow.

fianl is board for 8 chips or 16 chips?

Yes, a big thank you to both MrTeal and ChipGeek for making this a reality, incredibly smooth and professional!

Question: how are people going to house their boards? They have PCI brackets on them, but I have a lot of boards coming and would quite like to get some sort of casing that can hold rows and rows of PCI cards, but don't have a stack of old motherboards lying around, which would have worked.

Does anyone know of any cases for holding rows of PCI cards? Or have any other thoughts on how to handle this? (and obviously space needs to be left for the coolers in between)

8.

I'll bet you could buy some old GPU miners' used frames for not too much.

Final board is 8 chips.  We thought about 16 and decided the size, power and cooling would be issues.  However, our architecture fully supports 16 and we COULD make a 16 chip board in the future.  I'm just not sure either of us are willing to take on that much work and stress at this point.  :D

Also, is there any point really? Are there any upsides to not just running two smaller boards?

My GPU mining wasn't at that kind of scale so I don't really know what sort of solutions they used - plus that required PCI risers and connections to motherboards, so you were always limited in density as v. few mobos would allow more than ~3-5 connections.

I'm just looking for a physical structure to allow the cards to be stacked and nothing more, as they obviously connect via USB rather than PCI. Anyone know of anything like that where you can bolt a whole bunch of PCI cards in? I had a google and couldn't find much (as it's not a very common use case as most things with PCI brackets also need a mobo!)

I've used just a quick frame I made from back in the GPU days, which is basically something I threw together in 15 minutes out of 1" wood and a piece of aluminum angle.
http://i42.tinypic.com/am5pwy.jpg
For my boards I'm planning on using a proper 20 slot PCIe expansion chassis that's rack mountable.

The Chili is born.
https://bitcointalk.org/index.php?topic=304250.0

Very awesome post!! :P

Only 48 chips left. Can be sold in batches of 16 for BTC5.5

Chips sold:

16 to Choadmeyer
64 to dat.le79

Expected delivery: next week.

I am selling my 4chip slot for 1BTC right now, 1st and 2nd half of the payments paid.
This is hashing power equal to 4 x 4 GH/s = 16 GH/s

https://www.bitmit.net/es/item/63445-16gh-bitcoin-asic-miner-slot-for-1btc

You can PM me if you are interested.



 

PM sent

Looking to buy 8 Chips for 2BTC from someone in one of MrTeals Groupbuys.

Send PM Please

How do you rate the performance of that dual-white-fan in the middle? <edit> How do you rate the performance of your various cooling methods? The best and quietest one for cooling the chips would seem to me to be the tower style one in front but I notice you've got a little fan blowing at the vrm's. Have you tried putting a second fan on the back of the tower in a push-pull formation to try and get more airflow over the vrm's? would heatsinks like the ones they have on motherboards be helpful? </edit>

Are there any reference type coolers (the ones with the blower on one end) that would be feasible on this platform?

(btw, awesome job with this you guys. BFL should just send their "monarch" chips directly to you so this can be done right!

I don't think there's enough of a difference between the closed loop water cooler, the Hyper212 Evo and the Arctic Cooling Twin Turbo II to make a difference at least with my setup. You are probably better off spending the money on improving the thermal interface of your gap filler pad.
I would say this one is good (http://www.frozencpu.com/products/16878/thr-164/Fujipoly_Extreme_System_Builder_Thermal_Pad_-_14_Sheet_-_150_x_100_x_05_-_Thermal_Conductivity_110_WmK.html?tl=g8c487s1288) (note that with careful cutting that's enough to do 12 boards) while this one  (http://www.frozencpu.com/products/17467/thr-178/Fujipoly_Ultra_Extreme_System_Builder_Thermal_Pad_-_Full_Sheet_-_200_x_150_x_05_-_Thermal_Conductivity_170_WmK.html?tl=g8c487s1797) (enough for 24 boards) is the best. They make smaller sizes too.
I don't recommend greases; while I'm pretty confident the pro assembly house will have more consistent ASIC heights than me doing the test boards by hand I am still not confident enough myself that I will get good contact on 8 chips that way.
Also note that whatever cooler you have really needs to have a backplate. You just won't be able to get adequate pressure on the ASICs without introducing too much board flex if you don't.

I would say that the Hyper 212 Evo is a good reference cooler based on price and performance, my only reservation about it is that it can be kind of fiddly to mount since the ASICs are much lower than a CPU would sit. The included standoffs are a little too long, but replacing them with four of these (http://ca.mouser.com/ProductDetail/RAF-Electronic-Hardware/M1253-3005-AL/?qs=sGAEpiMZZMtrde5aJd3qw8BUoqM2q3%252bQuI%2fOHhCRKy4%3d) fixes the problem nicely.
The GPU coolers are easier to use, since they're designed to have a die that's ~1mm off the PCB. The biggest thing is finding one that has a large enough mounting pad to cover all 8 ASICs. I know both the Arctic Cooling Twin Turbo II and the Mono Plus both work. I would guess some of the other AC ones might as well, but they're more money and you really don't need that much cooling.

Also, some people have asked me where you can get PCI brackets. I got mine at Mouser, they're Keystone item 9203
http://ca.mouser.com/ProductDetail/Keystone-Electronics/9203/?qs=sGAEpiMZZMtzcnMBgC2bs6gC270mlsjAcmixu2SULFc%3d

Wow, the price difference between those two is crazy!  :o

Is it worth it (in your opinion)?

I am going to order the "Arctic Cooling Twin Turbo II" units for my boards.  What would you recommend for the backplate?  Does this unit come with one?

-Brim

Will this board work with anything less than 8 chips?

Doesn't look like the Arctic comes with a backplate - I too am looking for suggestions for sourcing the correct backplate.

I'm not sure if any GPU coolers come with a backplate. I just used some of the ATI HD5xxx style heatsink retention brackets since I have a few laying around from my GPU mining days temporarily, and I might eventually just use a piece of 2.5" aluminum flat with some kapton tape on it and then thermal grease. I figure if you're going to use a backplate it might as well move some heat.

For other cooler ideas, I was wanting to buy a few of these to test, my quick measurements make it seem like they should work
http://store.antec.com/Product/clearance_items/kuhler-shelf/0-761345-77081-1.aspx
They won't sell them to me in Canada though, which is unfortunate. Performance is actually pretty poor for what was at the time a $50 heatsink, but a lot is forgiven when you're selling a heatpiped cooler with a backplate and a good assortment of mounting hardware for $10.

Anyone know where to source these since I do not have any extras around?

I don't, I'm sorry. You can buy CPU backplates at Frozencpu if your cooler doesn't include one.
http://www.frozencpu.com/products/12547/cpu-pro-07/Prolimatech_LGA_Series_Replacement_Backplate_LGA_775_115x_1366.html

It's too bad, it's so close to the half brick DC/DC hole pattern that you could almost use those.
http://www.digikey.ca/product-detail/en/VHS-45/102-1488-ND/1016696
Out of curiosity, if I were to mill those so they have 53.2mm holes and you could use them as a backplate, would anyone be interested in that? They'd probably take up an extra slot so it's not desirable if you're looking at high density, but it might be interesting if you aren't trying to pack them in tight.

Just picked up a 1/4 sheet of the Fujipoly 11W/mk which should be sufficient for pads.

Are the sample chips you tested with Rev 2 BFL Chips MrTeal?  Interested to see if these will make any differences as Josh said they are more willing to reach higher mhz in the Singles.

Also, is there a specific date we should have our BTC into you by?  So I/We don't get delayed? 

Josh says a lot of things, don't get your hopes up... :P

Too true, haven't seen him pop his head out anywhere this week either lol

To be fair, he must be having a pretty rough time of it. Cleaning that much egg off of a face must virtually be a full-time job.  :-\

 :D  I stand corrected, he's posted on BFL forum so he's not missing anymore

Was it grossly unrealistic promises about the next generation? Or unsolicited rudeness to genuinely concerned and despairing current customers? Or covering his back/arguing about something he was wrong about?

EDIT: I checked - it was option 3.

God, I haven't read those forums in ages, I just follow one or two pages over there. I didn't realise quite how hard they are getting trolled these days. Soooo many posts and threads being deleted all over the place, and so much butt-hurt in general. Some of the posts that have been deleted now did make me laugh though... :)

Found some, if anyone's interested. Arctic Cooling makes one
http://www.arctic.ac/en/p/spare-parts/536/eva-foam-und-gpu-back-plate.html#

Thanks for that.
I think I am going to try the CPU backplates from FrozenCPU you had linked.

did you get the chips?

Mr. Teal got the chips and the boards should be in production now in Texas. The last word is that the boards would be all finished by early next week with some available Friday-ish.

what thermal tape i must buy 0.5mm or 1.0mm???

This is what I received from MrTeal

You mean we must put first thermal grease
And then put thermal pad and then the CPU cooler??
If put only thermal pad is not ok???

No, I can see how the wording might be confusing.

You only need the Thermal Pad and then whatever cooler you are using on top of that.

Ok thanks under stand


Standard cooling for CPU is ok?

You can not use most CPU coolers.  There are only a few that work correctly.  This is because of 2 factors:
1) A CPU is higher (taller) than our chips are on the board
2) Some CPU coolers are too small to cover all of the ASIC chips.

Please choose one of the recommended coolers. 
If you do not, you could damage or destroy your ASIC chips.

which is the correct cooler
any model to tell me
thanks

Read the post two posts above yours. Also:
https://bitcointalk.org/index.php?topic=304250.msg3318180#msg3318180