GekkoScience BM1384 Project Development Discussion

[sidehack]

My pair at 300MHz have been running beautifully for a week. Note that yesterday ambient air temperatures reached about 100F

Actually Friday morning the ambient was close to 60F, so you can see a pretty wide temp range had basically no statistical effect on the hashing output.

http://eligius.st/~wizkid057/newstats/userstats.php/1BURGERAXHH6Yi6LRybRJK7ybEm5m5HwTr

[chiguireitor]

According to the eligius' stats on the 1BURGER address, ROI is reached on 287 days, without including postage... not too shabby for a stick miner may i say...

Here's hoping TypeZero doesn't bugs you too much with the power supply issues you had with the Compac... 

[sidehack]

Honestly, I think the original Compac regulator would probably have worked alright if we had better init code. I've mentioned several times that one remaining issue is the chips briefly pull a large burst power on init that's handled differently in S5 code (which ramps them up slowly, helping maintain node-level power balance in the string) which overcurrented the monolithic regulator. If I'd started out doing the Chuckwagon-style regulator I ended up using (which is built for enough current that, if the inductor didn't melt first, the USB jack would) I'd have saved weeks jacking around with power systems.

I'll be using the same regulator design on the Amita. If we can get drivers worked out, it'll allow me to substantially shrink node-level capacitors from my test units. The same code will be used on the TypeZero, so no burst issues there. Today I'm working on designing power for the stud hub we're working on (Project Aardwolf) which the same driver could probably power the TypeZero but I want to use dual-phase instead.

At top clock, TypeZero boards can pull something like 37A off the regulator. Conduction losses in inductors and FETs increase with current squared; switch losses are pretty flat across the load range but double with double the number of switching elements. Going to a dual-phase buck decreases the current per switch node to 1/2 the total, so 1/4 the power loss per node (meaning half the total conduction loss). This doubles the switching losses, but since that's going to be something like one fifth the total power loss I'm okay with it. The overall reduction is substantial. Multiphase systems also simultaneously increase the ripple current frequency while decreasing the ripple current amplitude, which means input capacitors aren't having to handle as much of the immediate current needs so power loss due to ESR is reduced. Similarly, output current ripple decreases (while output ripple frequency increases) so output caps don't have to work as hard either to keep a very flat output voltage. Basically, switching losses double but every other system operates more efficiently, the net result being a substantial overall increase in system efficiency. I'm sure you all agree this is a very good thing.

So it'll take some additional finagling to make it all work perfectly but I highly doubt the process will be as chewed-up as Compac dev.

[AJRGale]

Frankly, no. It was good, but not amazing. I think the universe was conspiring on my side, because during the course of the film I got a drop-in visitor, three phone calls and some diarrhea.

Also, everyone posting and PMing me about picking up a couple sticks, don't. Sales aren't open yet. We still have no public review or opinion from half the testers.

Shows you how impatient humans are when it comes to new toys eh?
you mentioned when you may ship, if it all goes to plan (which it is showing not to atm) and now a price, you wanted to stir this wasp nest

none the less, I am just putting feelers out atm to see who in aus wants some, so far it hasn't pushed the 10/pp limit (i want 4 to 6, 2 other people want 2 each, so till then), who ever else wants some can PM me, i'll tally up and go from there.

I really wish i didn't choose woodworking at High-school over Electronic engineering, id so help you with your power plain setup..
I understand what you are saying there, but i have nothing to input

[chiguireitor]

[...]
I really wish i didn't choose woodworking at High-school over Electronic engineering, id so help you with your power plain setup..
I understand what you are saying there, but i have nothing to input

Same here, and i'm a Computing Engineer here... so yeah, pretty much understood, but man... wish i could be as proefficient with electronics :/

I'll bet some people here down south will want some stickies of those... will be sure to grab a mob here to buy it when it is ready.

[sidehack]

My college education consists of completing coursework for Computer Science, Computer Engineering and Electrical Engineering degrees (with most courses required for minors in math and writing). Work experience, however, was mostly limited to construction before Novak and I teamed up to form GekkoScience. That being the case, though I have all the bookwork knowledge my practical knowledge is less than a lot of folks on account of I haven't really done a lot yet. Part of not having done a lot is the availability of resources - I've never really had any money. But it's okay, because I'd rather have my current job and be broke than work on boring stuff for some corporation. And this way I get to hang out and do stuff for all y'all regular folks instead of making rich guys richer by selling useless toys to hipsters or whatever.

[PlanetCrypto]

Any word on the chip order?

[WBF1]

Any word on the chip order?

From three posts and 36 minutes farther up the page:

Also, everyone posting and PMing me about picking up a couple sticks, don't. Sales aren't open yet. We still have no public review or opinion from half the testers.

Welp, i'm sure that gets annoying!

It'll be ready when it's ready. I'm sure when they have news it will be posted. No need for daily inquiries.

[sidehack]

Chips are already here. I'm ready to start this thing any time, but my stated preconditions regarding posted public approval from reviewers have not yet been met so now we're waiting on other people (namely MrTeal, vs3, CanaryInTheMine and ManeBjorn) to report in. I know at least three of those guys have received their sticks, and I know at least one of them has run the stick and talked to me about it, and one of them told me a week ago he was starting a review and also had other stuff to send me (none of which has been seen). So with the exception of chips in hand, we're no closer to starting the sales and manufacturing process than we were this time last week.

[PlanetCrypto]

Any word on the chip order?

From three posts and 36 minutes farther up the page:

Also, everyone posting and PMing me about picking up a couple sticks, don't. Sales aren't open yet. We still have no public review or opinion from half the testers.

Welp, i'm sure that gets annoying!

It'll be ready when it's ready. I'm sure when they have news it will be posted. No need for daily inquiries.

1) My last post was on July 09, 2015, 04:17:43 PM (6 days ago).
2) It was a statement not an inquiry and thus required no response.
3) My question was NOT about stickminer availability. RTMFP
4) If sidehack finds me/us annoying he can ban us post haste at his leisure.
5) You are not privy to the back channel communications sidehack/Novak and we have had pertaining to this critical (and might I add brilliant) development effort.
6) By posting in the public forumn versus a PM was complying with his request to NOT be PM'ed about trivial matters (completed product availability).

The question was posted because very recently (yesterday) Bitmain has indicated to us that there are no chips (BM1384's) and S5 hash boards available for sale as they have all been used up making complete S5's (The price of which has risen ~$100 in the last 2 weeks). They are also unwilling to sell broken hash boards. It seems they are even having (creating) issues with supplying warranty based replacements. We have 2 boards that need to be replaced under warranty and they are giving us "the run around".

The above is a "sea change" in attitude with respect to our relationship with Bitmain.
Who in the past has been more than willing to supply us with "raw" components for our immersion cooling endeavors.
I suspect that they are "cleaning out" inventory of their old designs and at the same time gouging the ecosystem due to the lack of mining H/W availability.

Anyway, back to lurking.

[sidehack]

They also have not responded to my last two or three inquiries regarding chip availability and/or new-gen chip information. The last email I got from them was 12 days ago with a tracking number.

[WBF1]

Any word on the chip order?

From three posts and 36 minutes farther up the page:

Also, everyone posting and PMing me about picking up a couple sticks, don't. Sales aren't open yet. We still have no public review or opinion from half the testers.

Welp, i'm sure that gets annoying!

It'll be ready when it's ready. I'm sure when they have news it will be posted. No need for daily inquiries.

1) My last post was on July 09, 2015, 04:17:43 PM (6 days ago).
2) It was a statement not an inquiry and thus required no response.
3) My question was NOT about stickminer availability. RTMFP
4) If sidehack finds me/us annoying he can ban us post haste at his leisure.
5) You are not privy to the back channel communications sidehack/Novak and we have had pertaining to this critical (and might I add brilliant) development effort.
6) By posting in the public forumn versus a PM was complying with his request to NOT be PM'ed about trivial matters (completed product availability).

The question was posted because very recently (yesterday) Bitmain has indicated to us that there are no chips (BM1384's) and S5 hash boards available for sale as they have all been used up making complete S5's (The price of which has risen ~$100 in the last 2 weeks). They are also unwilling to sell broken hash boards. It seems they are even having (creating) issues with supplying warranty based replacements. We have 2 boards that need to be replaced under warranty and they are giving us "the run around".

The above is a "sea change" in attitude with respect to our relationship with Bitmain.
Who in the past has been more than willing to supply us with "raw" components for our immersion cooling endeavors.
I suspect that they are "cleaning out" inventory of their old designs and at the same time gouging the ecosystem due to the lack of mining H/W availability.

Anyway, back to lurking.

Sorry. I did RTMFP but did not fully understand TMFP.

[PlanetCrypto]

They also have not responded to my last two or three inquiries regarding chip availability and/or new-gen chip information. The last email I got from them was 12 days ago with a tracking number.

Relieved to hear you have a miniscule amount of "raw material" to work with, else a lot of your "time, hate, and discontent" would have been for naught.

Have been doing some research on the Avalon4 and the sfards.
Not really impressed with the sfards as I'm not an alt crypto kinda' guy.
Now an Avalon4 remasked and shrunk to 12nm-16nm, hhhhmmmmmm.
Really like the ability to tweak bunches of stuff. Understand that's not so much of the hash chip itself but the ecosystem it lives in.

IMHO, what this community needs is to find a chip designer and get something designed and manufactured in 12nm to 16nm (isn't that what AM started out to be?)

Anyway I promised I'd STFU . . . .

[notlist3d]

They also have not responded to my last two or three inquiries regarding chip availability and/or new-gen chip information. The last email I got from them was 12 days ago with a tracking number.

Relieved to hear you have a miniscule amount of "raw material" to work with, else a lot of your "time, hate, and discontent" would have been for naught.

Have been doing some research on the Avalon4 and the sfards.
Not really impressed with the sfards as I'm not an alt crypto kinda' guy.
Now an Avalon4 remasked and shrunk to 12nm-16nm, hhhhmmmmmm.
Really like the ability to tweak bunches of stuff. Understand that's not so much of the hash chip itself but the ecosystem it lives in.

IMHO, what this community needs is to find a chip designer and get something designed and manufactured in 12nm to 16nm (isn't that what AM started out to be?)

Anyway I promised I'd STFU . . . .

With AM doing what it did.  Would be intresting to see someone take that chip and make it and see if you can get better results.  But I'm sure to get the right's would be a mess with AM and cant really consider them reliable.

It's interesting about not responding on chips.   I hope Bitmain keeps supporting things like GekkoScience as it really is neat to see a design from a community member and it go all the way to production.  And not only that they succeed and make a great product on their usb miner.  I am still playing with my review one, and can say they did a great job.

[vs3]

Chips are already here. I'm ready to start this thing any time, but my stated preconditions regarding posted public approval from reviewers have not yet been met so now we're waiting on other people (namely MrTeal, vs3, CanaryInTheMine and ManeBjorn) to report in. I know at least three of those guys have received their sticks, and I know at least one of them has run the stick and talked to me about it, and one of them told me a week ago he was starting a review and also had other stuff to send me (none of which has been seen). So with the exception of chips in hand, we're no closer to starting the sales and manufacturing process than we were this time last week.

btw over the last week I've shared my thoughts and feecback with sidehack. philipma1957's review thread already covers 99% of the rest and I don't see much of a benefit to repeat the same.
Also, I'm not sure if any of my feedback is worth posting on this thread so I'll leave that to sidehack and I don't mind if he wants to repost it here.

In a nutshell - a very nicely put device, good design job and I'm sure it will make a lot of "overclockers" and all other bitcoin enthusiasts happy!

One more time - good job sidehack!

[PlanetCrypto]

"With AM doing what it did.  Would be intresting to see someone take that chip and make it and see if you can get better results.  But I'm sure to get the right's would be a mess with AM and cant really consider them reliable."

We have given thought to purchasing the IP (as licensing is out of the question) for the BE300 (assuming one could find the legally clear IP owner). From what we know, it would be a nightmare with the possibility of actors coming out of the woodwork after the fact demanding compensation. If an entity were to obtain the IP for the BE300, the design would need to be "tweaked" to make it a derivative work (~20%) to stay clear of IP infringement. Likely after that expense(s) it would be cheaper to "start from scratch".

And yeah, I'd like to get my hands on qty 100 BE300's.

[sidehack]

So, Novak and I did some discussing yesterday on priorities.

We have 1200 chips from Bitmain, and between their out-of-stock statement and that they've not addressed the question I've asked three times regarding more chips, I have to assume 1200 chips is all we'll see.

Demand for Compacs seems to be pretty high. We were planning on making about 600 Compacs and 200 Amitas with 200 chips left over for dev on larger boards, but I think what we'll end up doing is make 1000 Compacs and zero Amitas, with 200 chips left over for dev on larger boards.

I've got one thing left to test and Amita hardware would be done; a few steps past that and a 4-chip pod board would be done. A few steps past that and 18- and 30-chip TypeZero boards would be done. I'll probably make prototypes of these things with some of the dev chips, but will not manufacture any batches for sale.

If Bitmain (or Avalon, or BitFury or whoever else with a decent chip) will work with us in the near future, it shouldn't be too difficult to adapt the completed designs to a new chip. So it looks like, for this round at least, we're stopping production with the Compac.

Unless someone wants to put up a quarter million dollars for a full batch of chips, but we probably wouldn't get 'em in hand until October anyway and by then Bitmain, Avalon, Spondoolies and SFARDS will be rolling on better tech.




PC, if you do something like that I'll build boards for you. The BE300 was what got us started on this project in the first place.

[chiguireitor]

[...] Unless someone wants to put up a quarter million dollars for a full batch of chips [...]

I would want, however, i don't have the quarter million

Too bad good ideas are usually underfunded

[sidehack]

This idea could have been more well-funded but I'm unwilling to compromise. I don't like borrowing, and I despise investing. Borrowing to fetch the chips we did was a necessary evil, because if I'd waited until we had the cash in hand (which is to say, sometime next week) we'd have been sorely out of luck.

[philipma1957]

Well I am an optimist.  So I see the glass as half full.  Working tested reviewed usb sticks.  The best ever built in the history of the human race.  Pretty fucking good job on your part .

Sidehack, Novak and Bitmaintech.

I think you will be able to do more business with Bitmaintech. Cultural differences aside  they are business men and I think they see a future in doing business with you.

This is a funny industry we are playing in.  Could end up being very big  or fully busted out in the next ten years.