Plese DO! I'd be a blessing to see if these mythical beasties exist and perform well. Chip cound is right, 40 chips @a comfortable 30GH eah = 1.2TH. Kept in eco-mode, 1TH in the nose. fits perfect with as-reported in the dev thread.

In that Dragon thread ^^ mention is made of Innosilicon having rights to also sell BitMine.ch's A1 chip. Now, unless one company is part ot the other - bull. We deal with Asian semiconductor design/production/packaging companies and that kind of agreement is unheard of - Asia or anywhere else in the world.

These folks claim to be selling TH unis as well. http://ntekcomputers.com/products/1ths-asic-bitcoin-miner-1000gh

As for BitMine.ch's own rigs using the chip: http://bitmine.ch/?p=5404
"delivery expected 2-3 weeks' and that's for their own dog food...

There*is* hope they will take care of their resellers/OEM's and in turn take care of us. Question is - when & how? Do what BFL is doing and offer a Free Imperial with your original order AND a free 600GH card as well by years end?
Highly doubt it... Frankly if I don't get my lil'520 I wants some Ant's as compensation instead.

and bet they are running in either low speed or very minimally OC'd to get units out the door. Means a lot of chips vs what was expected. Might hefty dual PSU's as well.
Last I saw, yesterday the Dragon is 'sold out' and now call-for sales.

Have you read the dev thread? I doubt it. https://bitcointalk.org/index.php?topic=294235.380

Page 20, Feb. 18 by the way... is when tester/reviewer folks JUST got the chips up and going to any extent. Those lucky soles had received first runs around mid-late Dec. 2013. Continue reading from there.
Original promised Turbo speed per chip: 40GH/s
Now lowered to 33. On a good day. When the moon is just right...
Power usabe at 25GH/s, eccellent at 0.65W/GH/s. in turbo mode... over 1.5w.
The list goes on.

As I said above - AMT has royally blown this in their actions and business'Plan'. If like they claim are people 'experienced' in this kind of business then as soon as the delay from BitMine.ch started they in turn should have started a back-up plan - like STOP selling systems. Since the mining community was more than willing to pay in advance, then just take a deposit - not the full amount - with the clear statement of these being not-yet in production. Having those advance orders with a deposit would have given investors/banks a nice warm fuzzy feeling to provide working capital until things get moving right. Gotta say it does not look like they or other companies did. They just saw $$$ in their eyes and prayed.

As it is, if the pictures are to be believed, AMT & others are now sitting on a mess'o under-performing chips with not even a decent reference design board to use with them... Betcha BM.ch made them and the other companies pay up-front like we did... Can you say screwed, blued, & tattoo'd?

<snip>
I am inclined to believe AMT is legit.
<snip>
[/quote]
I do as well and unfortunately for them, they are as much caught in the problems BitMine's A1 chip has as we are. Actually. more so simply because while individually we buyers are currently out xxxx$, look at their position - they have scores of customers that they are obligated to. EACH who have paid them said monies. Aside from hiding & cowering in a corner like a puppy that got caught peeing on the floor AND shitting a brick in panic (making things worse) they need to communicate. Are they possibly business idiots that thought they could easily get into the OEM miner biz? Apparently so.

A. From day-1 of the chips actually getting into the hands of people to test in the real world - eg, actually connected to hardware - AMT should have at least pointed people to the A1 dev thread. All the chips trials and tribulations are there to be read.

B. They damn well had better be holding BitMine.ch's feet to the fire to do something to make this right with customers. If v1 of the chip is really destined to be an albatross then it is up to BitMine.ch to perhaps setup a program to give Antminers of equal or better total hashing speed to us.

In the end I lay this all at the feet of BitMine.ch. They are the ones who blew it. The chips were late - fine. But these were not pre-production test sample chips sent to a very limited # of testers/build companies to be used to find if they actual work right in the Real World. BM.ch hideously oversold the chip to companies and a lot like AMT & others jumped in head first. And hit a rock...

These seem to me to be pre-production quality chips that were expected to work like the computer sims say they should. Well, as their IP design company Innosilicon and BitMine.ch themselves should know better! I have never ever once heard of a complex chip right enough to be a salable item from the very first spin. At least BitMine.ch is getting the same 'roid rage from their own miner rig customers!

Again, I am NOT defending AMT as a business. They need to look at how GM is handling their recall woes and step up to the plate if they really want to get through this. Hiding is just making things worse. Will it cost them? Yep but they should have planned for something like this and not blindly believed this brand-new design would be plug & play easy.

I *do* understand the mess they and other companies are in over this but: They need to TALK! Heck, at least for smaller miners AMT is a reseller - not builder. Soooo. Get some Ants to give those will take them and also start selling them for cryin'out loud. At least they could bring in some money and have a salable item...

They want to pay my going field rate of $2k US or Bitcoinn-eqiv per-day and we'll talk. Cash or verified BC xfr of course. Paid per-day, 1/2 at noon 1/2 at end of max 14hr day.

Ain't a-gonna help the root problem though: BitMine and their Coincraft A1 rev-1 (I say rev-0.92) chip. My personal and only my personal opinion on the chip based soley on what I have been able to find out and was verified a little earlier on is that the A1 was oversold based solely on unverified - in the real world - design expectations. Being rather familiar with the semi biz I know all too well, unexpected things come up in the 1st revs. Sometimes out and out fatal, others, well call them problematic enough to force a respin based on things learned.

That is, learned from pre-production chips released only to testers...
Not ones sold with the full knowledge that they would be expected to work more or less as spec'd and intended along with the vendor (BitMine) atleast providing proven reference boards to use with them. With the full knowledge that were (are) pre-paid orders expected to be filled with those chips and boards, they have not done this.

Again, just my humble observation on this.

I've already a while ago sent an email to Jim at ATM with pretty much that exact offer regarding my 520 unit. More than happy to call it even with 3 (and given their current cost 4 would be better) S1's to match the hashing rate I paid for. I get my hashing, they even still get to pocket ~1.5k$ profit from it. Case closed.

And folks - check out the Official Bitmine Coincraft miner thread here as well. (you go find it like I did). We are not alone and 'roid rage abounds there as well...

heh heh heh... too bad the old Freon liquid heat xfr fluids aren't allowed any more. Could just plunk the whole boards in small tanks of it circulating around like the old Cray computers did. God they were beautiful sitting in their clear sealed temples with their Freon fountains running...

Then again, there HAS to be similar fluids available today. Hmmm...

Oh, and if you happen to have a dead chip around and can decap it, we have an in-house SEM  and high power optical video inspection scope available to see what they did...   If needed under full NDA of couse.

Interesting. What they call the SIP substrate is what we call an interposer and again, if you want to move heat, make it ceramic. Either ZTA or AlN and the thermal resistance drops to about 1/3-1/2 what an organic substrate gives. Also allows for an even finer pitch for the flip chip bumps that's another story. Putting coms & memory on top may seem bad but they no doubt based the decision on thermal modeling to 'see' where the heat actually goes. Seems they miscalculated a tad.

Also have to wonder how said com and memory circuits appreciate all that EMI coming from directly underneath them..Lots of possible/probable coupling mode there...

Also, while I have you here, "layering issues"? As perhaps in heat dissipation through an interposer if the chips use one? I might be able to help immensely there if it is an organic fiberglass) substrate that is used.

Oh good Lord... ^bump^ Do you realize what kind of firestorm ye will have just unleashed here?

And thank you for confirming what I garnered reading the entire dev thread. These are known/suspected issues and AMT shoulda been up front from day-1 about it. (also, why had no one else found that dev thread? After all, it *is* on the Bitmine support pages...) Woulda
 taken care of this mess long ago. Don'tch just love shoulda & woulda's... Li'sigh

Maybe add another option - give us all Antminer S1's. Enough to at least match the as-advertised hashing speeds folks were sold... AMT still might even still make money depending on out-of-pocket commitments to suppliers/contract assemblers so far...

1st, +10 on the humor! much needed here...
Um, since you are one of the regulars in the A1 dev thread, any thoughts on how goings on there affect us here in regards to AMT? I know you might have to tread a fine line with comments in regards to how other vendors/resellers may act but any insight as to what the delays are or might be from would be of immense value!

Er, also about power staging - if possible have the chips/boards sequentially turn off or step down to lower speeds to keep the rails from bouncing up and tripping any OV protection you might have. At the currents and speeds they can change that the chips can be pulling, even the at 1st glance minimal wiring inductance present gives a nice little kick back not to mention whatever response the regulators have... I design & work with high power systems (fiber lasers) and that can be a vexing problem as much as dips are.

Again, note the solution to the PSU issue - use 2 of them and scrap the current case design if ya dinna plan on this from the git-go. From what is being said in the A1 dev forum I rather assume all of these are based on what I would call development boards similar to these that Bitfury offer for their chips. http://www.bitfurystrikesback.com/shop/

To which Google returns as a condo,,, https://www.google.com/search?q=40.013228%2C-75.295929&oq=40.013228%2C-75.295929&aqs=chrome..69i57&sourceid=chrome&espv=210&es_sm=93&ie=UTF-8
I'd assume they are referring to postit & empty tower pic.

I agree 100%.

Frankly, since I'm 'only' on the hook for a 520GH unit,  if my per-order end of March deadline is not met I'd settle for them sending me 3-4 S1 ants instead ... S1's are currently cheap enough that even sending me 4, they'd still make money off me but I also get more hashing for the money I consider spent.

first let me start by saying that I am NOT out and out supporting AMT. Their lack of communications is reprehensible not so say their shall we say. less than realistic responses when they respond at all to some folks here.

That said - has anybody bothered to check out the A1 chip dev thread here?
https://bitcointalk.org/index.php?topic=294235.460
I rather doubt it. It is current up to today and very pertinent to the matter of AMT.

In a nutshell, unless I am massively wrong in reading it there are still problems regarding power supply stability and com bus timing sensitivities. Lots of other issues were there earlier on since the chips were first released to some folks to start tinkering with to see how they really work. Does not bode well for us in the short term but who knows...

Having a rather good knowledge of what it takes and happens during new chip rollouts. I understand what is happening at Bitmine.ch with the Coincraft A1 chip. If anyone, blame them for apparently giving way too many folks high hopes that for the first time I know of a brand new chip design would work as expected from the very first release-for-testing wafer(s) produced. Yes some things might be able to be helped by the foundry tweaking their processes but - unless there is ample firmware space and redundancy designed in to allow firmware tweaking to bypass gen-0 design mistakes more often than not it takes a physical design revision. Unfortunately, too many ravenous companies saw a chance to get ahead of the game (vs Bitfury 55nm chips) and Bitmine did nothing to rein them in a bit. There are what I would prefer to call development kits/boards out there and yes some working systems based on them. But...

That said, the mess AMT is in in that respect is understandable. Other companies relying on the A1 are in the same boat. Difference is they (well at least BFL) are being transparent about at - and offering some mighty fine incentives to hang in there. So far from AMT <crickets chirping>

As for the power supply issues - easy fix: just use 2 of them like these folks... http://coinhash.co/products/1-th-s-dragonminer-1000gh-s-coincraft-a1-28nm-asic-chip-bitcoin-miner-not-knc  As for is that a real working system? Dunna know but it *does* address the PSU problem.

btw: again, read the dev posts. The unexpected power usage is documented there as well.
EDIT: just read the latest post in the dev thread... up to 10% dead cores per chip? OUCH!