GekkoScience BM1384 Project Development Discussion

[quakefiend420]

Last thing before i turn back to my other endeavors.

I don't think that you should consider any compensation to anyone for either burger or shirt donations.
simple "thanks" is sufficient-and you already have done that on multiple occasions.
donations were just there to help with the project that most people cannot do themselves
keep calm and sidehack on!

Agreed, I didn't donate to the burger fund expecting to get some preferential treatment or any of it back later, I just like seeing cool stuff happen in the community.

[1714214697]

1714214697

[1714214697]

1714214697

[PatSNL]

Agreed, I didn't donate to the burger fund expecting to get some preferential treatment or any of it back later, I just like seeing cool stuff happen in the community.

Hear, Hear.  Some call it karma, others pay if forward, and some reaping what you sow.  I love all of the positive aspects of this project and the developers.  I believe that is really the bigger part of what makes this so attractive to many, including myself.

I'm looking for a a long-shot investment of mine to pay out handsomely soon.  When it does, I will donate more than enough to cover the materials and wages for Sidehack and Novak on this project.

Keep up the great work!

[sidehack]

Assuming we let you. Remember what I said about not liking getting something for nothing?

[armedmilitia]

I've been following this thread a while. This is some cool stuff you guys got going here, I think its awesome that there's still the 'little guy' competing with the big hardware corps. Also, this has got to be the most transparent development I've ever seen. Get yourself a deluxe burger with these bits! 

I'm definitely in to pick some of these up when you're done.

[vapourminer]

Though the burger money thrown in so far could cover about 60% of our materials costs. There's no wage associated with this dev work, so it's almost been self-supporting as far as billed project costs. Novak's lucky, someone's actually paying him to do his job.

that is nice.

I donated about 10 bucks to the burger fund and I DO NOT want my future order to be treated as anything but a normal order..  after all it was freely given on my part with no expectations.

[chiguireitor]

Also, peeps, remember, these guys sell amazing PSUs that are a heck a lot better than shitty "80 Plus Whatever" crap you get on amazon.

[electerium]

I sent a mcdouble.

goooooo sidehack

[valkir]

http://teespring.com/sidehack-support

100% of the profit will go to Sidehack and Novak!

We are now at 5 shirts!

And just to clarify this idea is really jsut to help sidehack and novak and to have some fun!  

I do really like the spirit of people that support this project!

[sidehack]

So, general testing update.

I was worried about comms - I've never been fully satisfied with the UART level shifter I had on there. And based on I think on vs3's comment that USB lines could be finicky, I was a bit skeptical about comms. So I peeled off the CP2102 on a stick I had working moderately, and ran it off the Novak USB adapter because that had proven itself on the breakout boards. Same performance. Rigging up a diode-drop level shifter for the RX line (to bypass the actual level shifter chip) gave me a bit better logic level hysteresis (and I might end up using it on the final since it's simpler, cheaper and more resilient than the level shifter chip) but the same overall performance. So then I says, well what if it's regulator noise hosing up the chip? I was seeing about 200mA draw being pulled by the whole hub and occasionally around 400Mh reported speed with sporadic shares reported, so obviously something wasn't right - I should have seen 600mA draw at least at the numbers I was running. But the breakout board, with identical circuitry, worked just fine - except its regulator was more distant than and also orthogonal to the ASIC, which was itself under a large aluminum heatsink.

So I yanked an inductor on a moderately-functional board, wired up some pins to drive it off my test schfifty-three board (the one rigged up for the dual-chip testing with the original breakout boards) an inch or two away (on the other side of a ground plane from the ASIC) and BOOM within a second the chip was too hot to touch, current meter was pegged and I saw about 8GH reported on cgminer.

So it looks like I need to do some shifting around with the regulator circuit and see what might be done about isolating RFI from the ASIC. An initial test will probably be to stick a grounded metal shield between the inductor and the ASIC. Past that, maybe drill out the inductor pads and mount it on the back of the board, to see if putting it on the other side of a ground plane from the ASIC will do the trick. If necessary, we can make a double-sided board with the ASIC and its heatsink on the back. This might help make it a bit more compact, and though double-siding sucks we need to get good at it anyway for the TypeZero boards. Hopefully sandwiching the ASIC between ground planes and a heatsink will be enough to RF-isolate it from neighboring sticks' inductors in a fairly dense hub setting.

It's getting on quitting time now, but I might have more news tomorrow.

Edit - so shielding added a bit more stability, not a lot. I went ahead and stuck on a heatsink all proper, and it basically did things just as well as the shield. We noticed a lot of ground noise on the board that wasn't present with the external regulator attached. Tomorrow I'll probably drop on a higher inductor and see if the reduced ripple current cuts down on interference. If something like that won't solve the problem, we gotta go for a board redesign with better ground plane islanding - probably in conjunction with a larger inductor. The U2 inductor is 4.7uH to our 0.47uH and it's pretty much right up against the ASIC with no performance problems. We couldn't do with a 4.7uH inductor though, since we need to be able to handle probably 10A there aren't gonna be many options which are power-efficient and also not huge or expensive. But I'll test with one anyway, stolen off a roasted Block Erupter since we have a bunch of those. Should be good for 3A, which will do 600mV 125MHz testing.
The thing is running off the schfifty-three power board right now, and its own stock heatsink, on Eligius at the 1BURGER address. Should be at 125MHz churning out around 7GH. I'm really hoping I don't have to do another board design, because that'll shift things back another two weeks or so.

[vs3]

So, general testing update.

But the breakout board, with identical circuitry, worked just fine - except its regulator was more distant than and also orthogonal to the ASIC, which was itself under a large aluminum heatsink.

So I yanked an inductor on a moderately-functional board, wired up some pins to drive it off my test schfifty-three board (the one rigged up for the dual-chip testing with the original breakout boards) an inch or two away (on the other side of a ground plane from the ASIC) and BOOM within a second the chip was too hot to touch, current meter was pegged and I saw about 8GH reported on cgminer.

Yeah - I was lucky that I didn't have too hard time with my power design ... And a good catch on your side!

Past that, maybe drill out the inductor pads and mount it on the back of the board, ...

If you can make it work without going to double-side components this will save you a lot on production cost (and not to mention testing headaches). But it is what it is and if you have to - then by all means go for it. There is always the option for another improved version later.

I'm really hoping I don't have to do another board design, because that'll shift things back another two weeks or so.

Crossing fingers on that! And let me know if you go past version 7
And actually I think v5 is something like the statistical median for me - first NF1 was 0.6 , NF2 was v5, NF6 was the only lucky one with v3.

[sidehack]

Production costs should be relatively unchanged - it'd just mean another pass on the pick-and-place and another bake, with a different thermal profile and a different solder paste with a different melting point... yeah, it'd suck.

I've already lost two weeks to the previous regulator design, and one week to figuring out what's going wrong with this board. Though admittedly I did kinda waste a few of those days not being terribly productive.

The breakout boards will probably allow me to skip some official revisions. I'm hoping the Amita will be pretty much a hole in one, if I can iron out the main details on the new multi-chip breakout and build it as an extension of the Compac with the same power and comms.

[elrippo]

So, general testing update.

I was worried about comms - I've never been fully satisfied with the UART level shifter I had on there. And based on I think on vs3's comment that USB lines could be finicky, I was a bit skeptical about comms. So I peeled off the CP2102 on a stick I had working moderately, and ran it off the Novak USB adapter because that had proven itself on the breakout boards. Same performance. Rigging up a diode-drop level shifter for the RX line (to bypass the actual level shifter chip) gave me a bit better logic level hysteresis (and I might end up using it on the final since it's simpler, cheaper and more resilient than the level shifter chip) but the same overall performance. So then I says, well what if it's regulator noise hosing up the chip? I was seeing about 200mA draw being pulled by the whole hub and occasionally around 400Mh reported speed with sporadic shares reported, so obviously something wasn't right - I should have seen 600mA draw at least at the numbers I was running. But the breakout board, with identical circuitry, worked just fine - except its regulator was more distant than and also orthogonal to the ASIC, which was itself under a large aluminum heatsink.

So I yanked an inductor on a moderately-functional board, wired up some pins to drive it off my test schfifty-three board (the one rigged up for the dual-chip testing with the original breakout boards) an inch or two away (on the other side of a ground plane from the ASIC) and BOOM within a second the chip was too hot to touch, current meter was pegged and I saw about 8GH reported on cgminer.

So it looks like I need to do some shifting around with the regulator circuit and see what might be done about isolating RFI from the ASIC. An initial test will probably be to stick a grounded metal shield between the inductor and the ASIC. Past that, maybe drill out the inductor pads and mount it on the back of the board, to see if putting it on the other side of a ground plane from the ASIC will do the trick. If necessary, we can make a double-sided board with the ASIC and its heatsink on the back. This might help make it a bit more compact, and though double-siding sucks we need to get good at it anyway for the TypeZero boards. Hopefully sandwiching the ASIC between ground planes and a heatsink will be enough to RF-isolate it from neighboring sticks' inductors in a fairly dense hub setting.

It's getting on quitting time now, but I might have more news tomorrow.

Edit - so shielding added a bit more stability, not a lot. I went ahead and stuck on a heatsink all proper, and it basically did things just as well as the shield. We noticed a lot of ground noise on the board that wasn't present with the external regulator attached. Tomorrow I'll probably drop on a higher inductor and see if the reduced ripple current cuts down on interference. If something like that won't solve the problem, we gotta go for a board redesign with better ground plane islanding - probably in conjunction with a larger inductor. The U2 inductor is 4.7uH to our 0.47uH and it's pretty much right up against the ASIC with no performance problems. We couldn't do with a 4.7uH inductor though, since we need to be able to handle probably 10A there aren't gonna be many options which are power-efficient and also not huge or expensive. But I'll test with one anyway, stolen off a roasted Block Erupter since we have a bunch of those. Should be good for 3A, which will do 600mV 125MHz testing.
The thing is running off the schfifty-three power board right now, and its own stock heatsink, on Eligius at the 1BURGER address. Should be at 125MHz churning out around 7GH. I'm really hoping I don't have to do another board design, because that'll shift things back another two weeks or so.

WOW!!! Things are heating up 
I personally don't mind if it lasts a month longer or not and i believe most contributors in here share my voice 
Keep it on sidehack yeeeaaahhhh!!!

[AJRGale]

Production costs should be relatively unchanged - it'd just mean another pass on the pick-and-place and another bake, with a different thermal profile and a different solder paste with a different melting point... yeah, it'd suck.

I've already lost two weeks to the previous regulator design, and one week to figuring out what's going wrong with this board. Though admittedly I did kinda waste a few of those days not being terribly productive.

The breakout boards will probably allow me to skip some official revisions. I'm hoping the Amita will be pretty much a hole in one, if I can iron out the main details on the new multi-chip breakout and build it as an extension of the Compac with the same power and comms.

man didn't build Rome in a day, or 2 weeks. nether did they get to the moon with apollo 1. you live to learn, we are enjoying your stories of your adventure you're taking. i don't thing people take you kidneys or knees if you fail this concept. you have kept up on what's going on, we appreciate this.

[philipma1957]

I get a t-shirt at worst.  So Just relax and sidehack on!

[sidehack]

That's actually the primary reason we're out of pocket on this project, because I'm saving my organs for a special occasion and I'd rather not mortgage them away for a stick miner.

Though tell you what, anyone wants to support the business in a more direct way, feel free to buy power supplies. I don't know what hardware is still shipping what folks need supplies for, but we have a few 750W and 2000W still around and another batch of 750W board parts inbound. I've been focusing on this miner a lot lately and haven't really tried to drum up any sales, but we really like sales since that sorta also pays the bills. And then there's the 12KW of hosting space we have open right now. That doesn't really pay the bills but it is self-sustaining.

[chiguireitor]

That's actually the primary reason we're out of pocket on this project, because I'm saving my organs for a special occasion and I'd rather not mortgage them away for a stick miner.

Though tell you what, anyone wants to support the business in a more direct way, feel free to buy power supplies. I don't know what hardware is still shipping what folks need supplies for, but we have a few 750W and 2000W still around and another batch of 750W board parts inbound. I've been focusing on this miner a lot lately and haven't really tried to drum up any sales, but we really like sales since that sorta also pays the bills. And then there's the 12KW of hosting space we have open right now. That doesn't really pay the bills but it is self-sustaining.

I'll probably buy one or two 2000w PSUs in a week, the 750 is chugging nicely and i'm getting 3 more S5s and would like to cut the clutter down with that orgy of PSUs i have. Also, do you by any chance sell them with extra large PCIe cables?

[philipma1957]

That's actually the primary reason we're out of pocket on this project, because I'm saving my organs for a special occasion and I'd rather not mortgage them away for a stick miner.

Though tell you what, anyone wants to support the business in a more direct way, feel free to buy power supplies. I don't know what hardware is still shipping what folks need supplies for, but we have a few 750W and 2000W still around and another batch of 750W board parts inbound. I've been focusing on this miner a lot lately and haven't really tried to drum up any sales, but we really like sales since that sorta also pays the bills. And then there's the 12KW of hosting space we have open right now. That doesn't really pay the bills but it is self-sustaining.

looking for the pdf  data sheet on the DPS-2000BB

any one have it?

1 board         = 50
12 pcie cables  = 36 or 48 length matters
1 DPS-2000BB = 35

total price = 133 + shipping  with long cables   for a 2000 watt psu that will drive

3 s-5's

this is a nice psu if you have 240 volt power.

[sidehack]

The best I could find was the stats on the DPS2500.

When you say "extra large" what do you mean? The wire is always 16AWG but I can vary the length.

Also, initial tests with the new inductor are less responsive than the barely-functional standard inductor from yesterday. Not sure why yet.

[chiguireitor]

When you say "extra large" what do you mean? The wire is always 16AWG but I can vary the length.

Longer, not thicker... sorry for the bad translation 

I'm going to put the PSUs about 0.5m above the S5s and would like to have a tidy rack.

[TheAnalogKid]

That's actually the primary reason we're out of pocket on this project, because I'm saving my organs for a special occasion and I'd rather not mortgage them away for a stick miner.

Though tell you what, anyone wants to support the business in a more direct way, feel free to buy power supplies. I don't know what hardware is still shipping what folks need supplies for, but we have a few 750W and 2000W still around and another batch of 750W board parts inbound. I've been focusing on this miner a lot lately and haven't really tried to drum up any sales, but we really like sales since that sorta also pays the bills. And then there's the 12KW of hosting space we have open right now. That doesn't really pay the bills but it is self-sustaining.

looking for the pdf  data sheet on the DPS-2000BB

any one have it?

1 board         = 50
12 pcie cables  = 36 or 48 length matters
1 DPS-2000BB = 35

total price = 133 + shipping  with long cables   for a 2000 watt psu that will drive

3 s-5's

this is a nice psu if you have 240 volt power.

Will power 4 S-5's if you have proper cooling... these things are rock solid workhorses.  I've had over 20 of them running for almost a year, first powering my S-3's, then the S-5's.  Blew only 2 of them in that time span.