Topic: GekkoScience is now dabbling with 16nm ASICs for new designs - page 30. (Read 77011 times)

He also got me tentative pricing for samples and batch chips, which is good so now I can start getting total cost ideas.

vh is hoping to have a functional multi-chip cgminer driver for BM1384 by this time next week. As soon as I have news that he's able to talk to two chips even in testing, I'll be sending off for sample PCBs of the two-chip stick and, if I have it laid out yet, the 15-chip pod.

Money raised by sales of these things will be funnelled into finishing dev on a 2-chip Bitfury stick, which since the controls on that are mostly identical with controls on the 11-chip pod and 33-chip S1 refit (intentionally), once the stick works fully we've done about 90% of the work required to make the rest.

The stick will probably roll out first, as the lowest cost to manufacture, and proceeds from that will be funnelled into pod and refit boards.

As my plan always is, I won't take in preorder money on a questionable product. As far as hardware goes, the 2-chip Compac is technically something I prototyped over a year ago and I had most of the 15-chip concept built (with 8 chips) and functional a year ago. So the hardware isn't a question. But in order to provide the best guarantee I can to my customers, that's why I'm waiting for vh to confirm the code. Once that's known it'll be time to open the gates.

BM1384 Compac manufacture, as some of you buyers in the past found out, wasn't always as time-efficient as I would have liked. You gotta remember, the first Compacs were literally the first things I ran out on my pick-and-place and I was just learning how to do it. Things go a lot smoother now, and I intend to bring in more equipment to help streamline things further - as well as some assistance to speed up a lot of the tedious work.

So, some good news today and I hope to have a lot more good news within the next week.

Great news!

Got an email this morning from Punin.


So, looks like it's Game ON.

(unless you own a Molex pin-crimping press and make PCIe cables as part of your job, then it's not easier at all)

Sent some burger or espresso BTC, I hope it leads to upgrade of my  S1 pile someday.

I would like to see connectors like in miners and ditch the melting PCI-E connectors. So much ATX powers floating around that it propably won't happen but this type of connectors would make wires much easier to make.

I still like the loan address system, where anything paid in will be used for an initial batch and paid back once it's been earned from sales. Once that's set up, anyone who wants to help can do so freely.

Lead times probably won't be too bad. PCBs and heatsinks will be the worst of course; depending where I buy the PCBs and in what quantity, it'd be 2-4 weeks.

I'm definitely happy to have a preorder system whwre we put a deposit down so you can use that to buy chips/materials. Not fussed how long the lead time is (obviously once the chips are confirmed).

Matt, if you need the funding for a pre-sale of sorts towards a potential future pod; just ask me.  I will consider it a donation with hopes of a future product.


I am sure a few others will be on this boat as well... But you have my support for what I can afford


I want to see your business thrive.  It would mean the world to the community.

Punin already knows this,hope he comes thru.

Everyone knows Bitmain & Avalon aren't selling any chips sooooo,he is the only option & has been since....11 months ago 

I couldn't even get Bitmain to sell me more BM1384 last fall. They straight-up told me no with BM1385, and the 87 they didn't even publish a datasheet. I don't think they care at all.


Offering $50 per chip is the same as just saying no, and anyone with half a mind would figure that out.


Oh hey also, vh is onboard with the BM1384 project so that's good. It's probably gonna take some packet sniffing and analysis to figure out exactly what we need to do to talk to multiple chips since S5/BM1384 protocol documentation is basically nonexistent, but I'm already working on a testing setup for that. It's fortunate I forgot to include my first BM1384 testing PCBs in the prototyping stuff auction a few weeks ago because now I'm using them again.

So, with vh in the game working on the code, I'm okay putting more resources into the hardware end of the project. Already got a layout for the 2pac, and I spent a couple hours today working on 15pac layout. I pretty much know where everything is going to go, but for that many chips there's a lot of connections and a lot of inter-node level shifters and bypass caps and stuff to worry about so it might be a while.

And hopefully before long we hear good news from Bitfury, and then I'll be super busy forever doubleteaming manufacture of 2pac and 15pac while doing dev for Bitfury boards. But that's the good kind of busy.

So getting Bitmain to sell S7 and S9 chips is out of the question?

You made yourself sound so desperate for the chips. Bad move in my opinion. Now he knows he has you at his mercy. $50/chip 

Nice letter. If he has any soul left he will reply.

Looking forward to the new products, hopefully you wil get what you need to make this happen.

thanks!

I know these are old, but I like it when people say a thing and stick to it. Especially the one from... holy carp, 11 months ago? about his willingness to sell small quantities like they used to, and we've seen pretty much no movement toward that goal.

I sent in an email as well and this is part of response I received..

Our program has a minimum order quantity of $1m of ASICS.

Is that 1million? If so I think a group buy for chips would be needed as well to meet min requirements.

Not like I can dev fast enough to cramp their sales anyway. If I had the go-ahead for chips right now, I might be prototyping a refit by Christmas. The primary focus is the 2- and 11-chip things, since the control set will be pretty much the same for all 3 and the refit will basically be three 11-chips on a single board I really need to make that work first.

^^ Could not have put it better myself.
I am hoping that BitFury dragging their feet is due to the X5/X6 or LightningASIC project others are doing. Possibly an agreement to not step on their toes until a miner is released? As a businessman I can understand that but if so then I'd think that obstacle should be removed once an X5/6 or other one gets into the wild (soon?). If not BF should at least have the courtesy to tell you what the problem is...

If he doesn't listen to that letter which tells him exactly what we (the community) want and how many of use are willing to pay for it rather than bitmain with there dysfunctional asics and warranty, then he and bitfury are fools.

Fingers crossed!!!

I just sent this message to punin at Bitfury. Hopefully it gets his attention; the last 4 emails got no response. He's my only point of contact, so if anyone knows someone else I maybe should be talking to as well I wouldn't mind hearing about it.

--------------------------------------------

Punin,

In order to help raise money for initial funding of a Bitfury-based miner
project, and as an indicator of the community's readiness to have more
miner options, I'm working on a 2-chip USB stick miner and a 15-chip "pod"
miner using outdated BM1384 ASICs. These two miners will meet the same
market sector as two of the miners I'm designing around Bitfury chips.

I've just started talking about it in the last few days, and already have
a customer list totalling over 200 units of the USB stick; it looks like a
lot of people would rather buy the pod, as nothing on that order of
magnitude (for power consumption or noise level) has been offered in about
three ASIC generations. People are lining up for these products, in part
because they are the only things available in a niche market and in part
because they know the profits will be put toward the design and
manufacture of better miners using your ASICs.

Think about how good sales would be for the same product with three times
the performance? That's what I want to bring to market, and that's what
I'm working on.

If I build a 2-chip USB miner with Bitfury chips and it sells the same
quantity as my BM1384 Compac (which, 14 months after the initial batch,
are still in demand and being produced) would take 7000 ASICs. If I make
half as many pod miners (and I've already had requests to license the
design once it's ready) is 20,000 ASICs. And I've already had interest in
my S5 refit boards (which I know others are making, but I will integrate
desirable features others overlook) to the tune of another 12,000 ASICs -
and that's without advertising.

In the last two weeks I've had 0.66BTC donated from half a dozen sources
just to help with development expenses.

I very much appreciate that you sent me sample chips and datasheets. I
very much appreciate the shows of good faith that you've put forward so
far, especially claims to support community endeavors and
decentralization.

However, I'm stuck in limbo right now. I am unwilling to take in customer
money on an unproven product, especially one I don't know I can
manufacture in bulk. I can either continue to devote my limited resources
to a project with zero indication of a future, or I can put them into
something which, while not as good, has a known chance of bearing fruit.

I really want, and the community really wants, to see a variety of good
miners built with your chips. I really want to bring that to them. If I
can do what I'm trying to do, the result would be tens of thousands,
possibly hundreds of thousands of ASICs in additional sales and a buttload
of positive PR for you.

I would really appreciate, and the community would really appreciate, some
word back about whether you're willing to make good on the offer to sell
me some chips. Just that confirmation right there is enough to re-ignite
quite a bit of support for my project and, ultimately, your sales and
public image.

So, please let me know what's up.

Thanks.

Matt Hercules
'sidehack'
GekkoScience