Topic: GekkoScience BM1384 Project Development Discussion - page 3. (Read 146713 times)

I see your point about a back plane cooling, if its not all a common ground plane, then:- "magic smoke"
Some plastic standoffs and a secondary "case fan" will be  a safer base plate.

No long, wall of text from me sidehack and friends...

Just a quick note to say I keep following this thread and things look very nice!

I hope you all had a great Christmas, and if you are as excited as me for the pods, then that is when Christmas is coming for us

Yes, you can see in the first picture the temp sensor is the little square guy in the center of the ring of ASICs.

I've also thought about doing a plate of aluminum across the back. The bottom layer has a few signal traces in it for the temp sensor, but otherwise is for the most part absolute ground plane - which means chips aren't belly-pad-sinked into it directly like on the S5, except on the bottommost node. This thing is designed for top cooling. I can change that layer around if necessary to give it a bit more heat flow into the rear face of the board without a lot of trouble, but then you have to make sure your heatsink is electrically isolated which tends to increase thermal resistance. That's a big reason most string boards are top-coolers - having 15 different voltage-potential planes all rubbing up against the same block of aluminum can get fairly risky.

If I retool for a different chip whose dimensions are any larger than the BM1384, I'll probably have to either go with fewer chips or give up on the CPU cooler. It was darn difficult to fit 8x 8x8mm ASICs in a 30x30mm space with room for routing. Any bigger would be impossible. But I could just build for a basic heatsink, maybe 80mm square with a case fan mounted, and make the chip field any size I wanted to fit underneath it. Of course I'd keep the same size and screw positions on future boards which could mount up to the same heatsink.

Is the board flat on the back?
I'm thinking of a single block aluminium heatsink on the back, then the top cooler will just screw to it through the mounting holes.

lookie at that gorgeous thing! its a 1st step, once you get the buck bolted down, and get the controller talking to the chips and auxiliaries (temp sensor in the middle yes?) all you need to do is retool for the next line of chipsets to work with.

Care totry with resitor based divider

I cannot wait to get my hands on 1 of those  - looks great and keep up the good work. 

Merry Christmas everyone!  Excited for these to see the light of day.  Nice work. 

Love reading through this thread, great stuff!

Looks the goods mate and I have to say I'm excited! But take it easy, it's Christmas, spend it with family!


Merry Christmas everyone

Holes for LGA115x cooler, use whatever you want. But I got a couple cases of those Freezer7 coolers which were supposed to be mounted on about 40 of Technobit's minion boards about 17 months ago for hosting customers; said boards have yet to be delivered though I'm fairly certain they were all paid for about 17 months ago. So I'm using one to test on this board. If I end up making enough pods to sell, and folks want me to provide coolers also, I have permission from the owner of half the coolers sitting idle on my shelf to sell them. But in general, I won't be providing coolers. If I end up on a future build getting away from CPU coolers, I may just rig up a simple heatsink and fan that bolts right to the board; if I do that, I'll keep the same size and screw positions on future boards because not reusing the heavy mechanical parts which never go obsolete and greatly increase your shipping cost is kinda dumb.

I should have a few hours tomorrow I might go by the shop and see if I can't get the onboard buck working better. There's nothing I can do about the controller until we mount up the micro and get firmware and drivers going, but I'd like to have everything besides that working since everything besides that is specifically my job. So maybe I'll do that tomorrow. If not I'll do it next week. And hopefully somewhere in there it decides to start hashing more reliably because it's averaging something like 86GH instead of the 132 I should be seeing.

Glad to see the sexy miner photos again ^_^

So are you guys going to use a cooler like this, and keep it low profile like the pods?
Or was it just a cooler available in the shop?


Hope you guys also have a Merry Christmas and keep producing cool mining gear

looking good there sidehack im really liking the pictures of the project

Thanks for sharing!  It's great to see a pod.   Exciting to hear the future of these.   I like the idea of more chips on a board and going from 1 to 8 definitely does that.

Have a Merry Christmas aswell!

since Bitcoin is a Proof of Work coin, here's some proof of work.





Here you can see the new thing with 8 chips. I haven't debugged the buck yet so it's still running off the hacked '53 but that's okay. Shouldn't take long to work it out. Comm lines are run straight into my hacked S5 controller. I'm not sure what's up with this guy, if it's comm noise or what, but it doesn't get full speed. Ran overnight at 150MHz (should have seen 66GH) but only regsitered 52GH or so. Now it's set at 300MHz and I should see 132GH (or at the 80% metric set by the previous test, 106GH) but I'm only seeing about 87GH. It broke 100GH earlier but I think it's getting tired. Not sure what's up. But it's running. I have found one error on the board, I forgot to level-shift the chip reset coming out of the microcontroller, but it's hashing which tells me all the worrisome and complex parts are pretty much right. My level shifters are all good. Still need to test dividing PLL LDO voltage but I'll worry about that after the local buck is working.

Merry Christmas, chumps.

very cool sidehack cant wait to see the pictures of them

So I took a few pictures but it's almost 10PM and I don't really want to still be at work. I'll just post some info, and probably put up pictures tomorrow.

Right now I have an 8-chip pod that works. Kinda. It's on life support. It works by being attached to a couple extra boards. For one, my buck isn't working quite right. Volt setting is good, it lights up when it's supposed to. But I think current limiting isn't quite right. So I borrowed one of my hacked '53 boards with an appropriate voltage range and tied it in instead. Additionally, since Novak's still writing the firmware and I don't know if he's started on cgminer driver code yet (probably not), we don't have a controller onboard. So I tied in my hacked-up S5 controller.

The thing had a bit of problem starting at 150MHz 650mV, probably because of string imbalance. But it lit up at 150MHz and 700mV, once operating stably I turned it down to 650mV and it's running fine. I say "fine" but it's still behaving a bit weird. After about 17 minutes I'm seeing 52GH average reported, when it should be at 66GH. Not sure what's up. But I am seeing only 0.4% errors and it's registering 8 chips properly. Power draw is 12V 2.5A, so about 30W, which ain't bad for 66GH if it were doing that right. That's also higher than it should be, for one reason because I have an extra buck circuit (sitting idle but still drawing power); for another reason, because the S5 controller pulls a heck of a lot more power than our 48-pin LPC micro will.

Eligius and its high diff were giving me fits, so I switched over to slush pool, sidehack.worker1 in case anyone figures out how to view stats on that. I'll keep it running overnight and see what it looks like in the morning.

I figure tomorrow I'll try and get the onboard buck working. If I can figure that out, I'll probably build a second one so I can keep playing with one and have a complete unit for Novak to finish up.

oh well. Also have 15 s3's in the process of being retired.

Buck parts maybe, but I'm after BM1384 rather than 1380. I won't take in anything until I get a design proven and working (which the hardware I hope to have proven this week, software is on Novak) and even then I'll need basically a threshold minimum number of sales committed before I move to manufacture. It's possible that I won't get them, because the price might have to go up and as diff increases and new stuff appears these are becoming less and less desirable so my prototypes may be the only things that exist. But it's still worth doing in that it gets me about 80-90% of the way toward a similar design with new chips if new chips become available.

Faboo!
Along those lines I currently have 1 s2 off-line and another soon to be replaced with a s7 so if ya need chips yer welcome to the blades.. Also have 4 s1 hash boards sans heatsinks from doing two s1>s3 upgrades.