Topic: GekkoScience BM1384 Project Development Discussion - page 33. (Read 146665 times)

Sidehack already said that the project will follow the new chip. So I guess he will try to get some BM1385. 

Agreed, but the designprocess starts right over again. GekkoScienceBM1384 should be burried and BM1385 should see the light IF GekkoScience gets BM1385 from Bitmain...

If anything, provided the chips are good to work with, it merely changes it to "GekkoScience BM1385 Project Development Discussion" with a more certain future as to chip availability.  Or so one can hope anyway

Not yet, every one know the new chips from the start of this project.

Sorry to say, but i think this project is dead.....
You might want to read this from BITMAIN

Waiting patiently for the day when I can get upgrade kits for all my S1's. Hopefully it will be by christmas of this year. fingers are crossed.

Jabber has a prototype board where you would be able to connect up to 8 devices per card.

He is quite busy, he didn't produce more then one board as we speak. (prototype)

I've got the next two days or so to dedicate entirely to dev, before I have to get back around to manufacturing. The last two weeks were put on 750W board and cable assembly, but that order's done. Essential Compac parts shouldn't be all amassed until Thursday or Friday so I have a bit of time. I'll probably spend it all on buck design, since we have three projects all requiring buck design work and there's gonna be some overlap between at least two of them. With reports of successful S5 undervolting and some things happening on our end, looks like project Chuckwagon is back on. Project Aardwolf (the USB hub) needs to get kicked back into gear, and I need to finish ironing out power systems for the TypeZero boards. For the next month I'm going to try to spend all my time not required for Compac assembly focused primarly on TypeZero dev. I want to have a complete working 18-chip board, and a complete working 30-chip built around the rack-standard formfactor I also need to codify. Compac profit margins (once there are some) will go a long way in funding this stuff, and hopefully by the time I have working models we or PlanetCrypto have made some headway finding a new chip to work with.

That thread was about three down from the top when I looked for it. I haven't actually read almost all of it though, so I'm not really sure where they are. I think making something USB to drive an S5 controller shouldn't be too difficult hardware-wise but I'm not sure about the software.

PlanetCrypto - have you talked to Bitmain, Avalon or LK at all?

There's too much stuff on this Forum, for someone who is new. 

Sorry to have troubled you.

Best case scenario assuming everything goes according to plan, the earliest we could have a chip for prototyping is Jan 2016.

I kinda' think of the chip design project as an insurance policy.
I'm really hoping that one of the current chip manufacturers relents/reconsiders the embargo and starts selling their state-of-the-art chips and not their old tired technology.

IMHO, the release of used Bitmain's S5+'s with a 90 day warranty is a withered "olive branch".
It's a "day late and a dollar short" as it will ROI just about the time the block reward gets chopped in half (and likely the max block size goes to 20M up from 1M).
And then become a worthless piece of H/W.

It also pisses me off that while we were mining with .51J/GH/s units (or worse) Bitmain and other "large" Bitmain clients were mining with .445J/GH/s stock.
We have been predominantly a Bitmain shop since the S1 release days and one would think that the S5+'s would have offered to a loyal customer. They weren't.
Am feeling like we got stabbed in the back with a Claymore.

It's probably a good thing for the "big 4" they're not American companies else they'd be subject to Rico statutes.

Growl!

So what you're saying is a USB adapter to drive a single blade from one AntMiner or another?

https://bitcointalksearch.org/topic/diy-reward-100-antminer-s1s3-blade-on-raspberry-pi-671128

I have had this idea, may be a good idea, may be a terrible idea, may not be possible?  It's driven by the chip supply situation and an extension of philipma1957 desire for a single module S5+, and came to me in the shower this Morning.  

So I am assuming that there are a lot of similarities with driving all of the BM range of ASIC's? So how about a version of the USB board that will plug up to and drive an S1, S3, S5, S5+ S7 Hash Board?

Would of course need a separate PSU and you would have to organise mounting and cooling. It may be at some point that S5+ could be broken up into unit's or hash Boards? Even a single board is 860GH. If that was to happen the problem would be what to use as a controller? Also people may have odd S1, S3, S5 Hash Boards and it might be nice to have a way of controlling them?

So not sure that it makes entire sense but thought I would share it in case it does have some merit or triggers any other thoughts?


Rich

I'll at least get as far as a working model board based around BM1384. Between MacEntyre and myself we were only able to get 2000 chips, and probably 1800 of them will be turned into Compacs so I'll have the rest to play with. Retooling a proven working BM1384 board for a different chip would be maybe 20% of the electronics work of designing a whole board, so I'd like to go ahead and finish it and then hope we get access to a new chip I can shoehorn in. Probably driver code would take longer than altering the board itself.

So you will not make the S1 upgrade sidehack ?

Sad to ear. Did you use all the bm1384 that you had ?

Well to be honest, not much good work can get done on my end unless I have some chips to work with. I'm hoping we can either get in good with Bitmain or Avalon (for next-gen stuff if any of them would talk to me) or PC can work some sample magic with InnoSilicon. If I had a month in the clear and a couple thousand bucks, I could probably have a working model 18-chip TypeZero board with BM1384 on it but by the time we got to manufacture that chip'd be a year old and top-clock power consumption about twice what everyone else is selling.

So, if anyone has any sort of "in" with one of the current chip makers (that don't make big hot BGA things, so no KFC or Spondoolies or SFARDS) and could maybe sneak me some samples, even if it's classified and I can't talk about it, it'd still be something to work with. And if anyone can help PlanetCrypto get a chip dev enterprise ironed out and operating, please do that also. This community needs to cycle back around to folks kickin' butts and doing awesome stuff instead of suckling off the teat of a few greedy enterprises.

Yes, yes, very goods ideas and projects...

... but for the moment, we will be patient and the bitcoin mining is harder every day :-).

I like what I am hearing from sidehack and PlanetCrypto. Looking forward to the big boards. Keep up the good work guys.

Now, imagine if PlanetCrypto did a chip you can talk to with I2C (amazing if it had a DIP version), sidehack a "standard" board that could be used, say, as an RPi extension with an stackable design, then you could build a stack of chips and call it a miner

GekkoScience is interested in evaluating sample chips from Innosilicon, Bitmain, Avalon, whoever makes a decent chip to see how they'd behave in our preferred topology, and then design a miner around which one we thought was best (based on stability, efficiency, price and the likelihood of actually receiving chips in bulk).

We also wouldn't "burn in" boards for a month, keep entire batches running on our own shelves or sell dusty gear as "new". Just throwing that out there.

The best thing, which I'd say isn't terribly likely to come to fruition in this generation but I'd really like to see happen, is the community chip. Everything we've done in the bitcoin world has been aimed at helping home and small operations get closer to even stage with big farms. We started out by selling server PSUs to individuals so folks could get efficient and reliable power without having to lean on retail-priced ATX gear that was either expensive or failure-prone. We looked at trying to design a miner over a year ago but didn't have the resources. We still don't really have the resources but screw it, it needs to be done. A community-minded enterprise building affordable, easy-to-work-with chips which are competetive with the big players in the industry, and making them available to small lightweight innovative shops to build and sell for the community, is something I think we really need. Giving power back to the people will only go so far because the guys with money and power become increasingly focused on maintaining and increasing their money and power (and everyone is willing to bend over backward to do them favors, oddly enough) but it's a thing worth doing nonetheless. We'll be chipping in what resources we can for the chip dev enterprise, and in the meantime looking for other chip sources to try and build you guys the good stuff.