Topic: Black Arrow 28nm 100Ghash Bitcoin ASIC from $0.49/GH/s - page 173. (Read 787053 times)

http://www.richgibson.com/blog/wp-content/uploads/2011/09/When_Did_Ignorance_Become_A_Point_Of_View__Cover.jpg

Delays are one thing, blatantly lying to your backers another.

Go away.

I really believe it's going to turn out to be an a dressing problem.  Without more details from BA though it's hard to help them.

I wonder if by adding the second connection to the SPI if it is losing power?  It should be in third state I imagine.  Difficult to make any guesses without having any idea of what is going on. 

Preordered 10 X1 from BA back in December.

When u first failed to deliver in February i asked for the possibility for a refund in march.

After really slow support on the tickets you asked for my wire details etc and made me believe that you would finally proceed refunds.

Then again weeks later you tell me that refunds aren't possible till May, which also was the next promised shipping date.

Prefering hardware over late cash i choose to stick with you folks in good faith and opted for the hardware in May.

Now, again almost a month late you still have no working & complete hardware and you telling your backers that a refund isn't possible anymore. This is ridiculous and i smell noone ever got a refund from you guys :/

So Black Arrow it is clearly your duty to provide your backers with fully assembled and working miners and a fair compensation.
Offering compensation in the form of paperweight hardware is a joke. 50% hashingpower after this long delay is a joke too. And to top all this, making your backers pay for additional parts and shipping to get their compensation units work is total bulls!?#

Deliver what your customers payed for, FULLY WORKING AND ASSEMBLED MINERS.
Give us a real compensation, 200% of ordered hashingpower, while still not that great would be acceptable.

Ofc free shipping and with PSUs…
I'd rather get one or two less compensation units then having to send you folks more cash.

Don't they have Baidu, then?  Their major search engine says:

http://baike.baidu.com/link?url=4CMzx7Iab7004XcqzRhYxfRnPd1Bms7HQjyvRL2-vKpnuc8h5t5gUNww93ScPP9a

I know, shittier and censored, compared to Google, but sometimes you gotta make due with what you got when the angry mob's at your door.

Sounds like there will be a big shipment of x1 to minersource next week. Anyone have a date for the arrival of the cases and the date of shipping to minersource?  Does minersource already have the ceaseless x1 units?

The PICE connector has 98 pins but many of them are bussed together for power. In the photos on your website I count 18 data pins on one side of the card and the rest - bussed together for power. An 18 conductor ribbon cable isn't that hard to come by, and then it would just be a matter of getting enough power to the board. I could be wrong - and I can't say I still believe this is possible because I don't have a full understanding of the PCB layout - only what I've seen in scattered photos.

A few questions before I "book my ticket":

1) What is the pinout of the hashboard?

2) Does the board create 1 SPI link between the raspberry pi when it is working, or 2 (one for each chip)?

3) Does the backplane now bus together all the SPI links from the hashboards, or is there a single isolated link between each hashboard and the controller?

4) Does the SPI "Daisy Chain" - that is to say go from the first hashboard to the backplane and into the second hashboard then back out to the backplane so on and so forth. If so has the ability of the hashboard to support this daisy chain link been verified?


I'm honestly trying to help, not trolling, and I'm a little dismayed by your dismissive response to my sincere suggestion - especially considering you guys openly asked for help.  




Oh, one more thing. Even if it was all 98 conductors active there are commercially available parts for that.
http://www.frozencpu.com/products/20753/cab-1908/PCI-Express_PCI-E_x16_Extension_Cable_Riser_-_7.html

Prospero - China trip update, Stuart from Logic Ethos Ltd

"Hello,

I made it to Black Arrows offices.  Heathrow, Moscow, Hong Kong and crossed the Chinese border on foot.

I have meet BA's management several times before, but this is the first time I have seen the place where it all happens.

The offices are full of people. A buzz of activity, from not only their own engineers, but those of their suppliers too, trying to resolve a problem with the SPI bus.  They showed me what’s happening. 
A single 2 chip board works fine.  Add more chips, and the SPI bus fails. This effects X3 & datacenter hashing.

I spoke to BA's management last night.  They said they will make good with compensation, but they are so busy trying to fix the problems at the moment to speculate exactly what that will be.

More details here http://www.blackarro...2-may-2014.html

I'm going to stick around for a bit, at least for as long as my visa lasts, and see what happens. So if there is any news, I will post it.

Stuart"

Surely most of those 96 pins are power and the backplane doesn't have any problem with power does it?
So why not cutting the spi traces and running wires only for the spi?
It might not be elegant but depending on the pcb it might be workable (ie, desoldering something and putting an spi only connector glued to the board) or at least it works to test it.

that's like an apology, but different. 

Dear Jonny,

The PCIE connector has 98 pins and there are 5 connectors in each case.

If you still believe that this can be done even on one X3 (not to talk about mass production) you are welcome to come to our office and try to do this yourself.

Regards!

Unfortunately, I think that BA really did try to create multiple SPI control signals on a single SPI connection across the backplane.  There really is no other explanation for why "SPI communication" fails with more than one hash board.  If they used independent control wiring (one to one) rather than an SPI bus approach, then what BA describes would be impossible:


I cannot think of any explanation for that behavior unless they attempted to have SPI "one to many" control to slaves.  As the beginner's guide to SPI points out, that way lies ruin.

I'll say it again:
Replace The Backplane with Wires
As in 1 (one) SPI channel PER HASHBOARD going direct to a controller. I hope to god you didn't have all the hashboards sharing a common bus on the backplane before - because I can't imagine trying to get that to work.

They would have to ship 40 X-1's for every X-3 that was ordered to meet their compensation promise ... even if they "only" shipped 35 (to meet the promise they made already) that is a LOT of extra hardware, shipping, etc.  And personally I don't know what I would do with 40 X-1's except maybe sell them.  It would cost an extra rack just to mount them versus what I have. 

According to the last update the X1 is ready to ship as soon as the cases arrive next week. Maybe they should refocus their efforts to get all batch 1 orders filled with those instead. It's not the prettiest solution but what choices do they have left?

To put a finer point on it, from the article above (beginner's guide to SPI lol):


So, Black Arrow, as it appears your backplane did not adequately address the lack of an addressing mechanism (surprisingly, but that is water under the bridge .. though I think at least one engineer ought to walk the plank for that f-up)... It is impossible, even for a very sophisticated customer, to address the feasibility of your second option (getting hashing boards for a DIY solution) without some further information.  To wit:

(1) We need ALL technical specifications and interface protocols for the hashing board.  Is SPI the only control mechanism?  Can the boards be used with USB instead or would an SPI - USB bridge need to be implemented by us poor customers?

(2) Are you supplying power through the backplane, or does each board (hopefully) instead have its own power connection?  Either way (individual power through board, or backplane) we need detailed specs on the types of connectors available on the hashing boards to supply power, and the power requirements, tolerances, etc.

(3) For a customer that might be interested in the DIY solution, it is essential to have a ready to go mounting solution.  The X-3 cases would provide such, so why not send us the cases and power supplies together with the extra hashing boards.  Ideally, you would provide an 5 Prospero cases with power supplies for each X-3 order (enough to mount and power the full 4.5 Th/s of boards you are proposing to send).  At a minimum, you should provide 4 cases (for 4 Th/s solution) and the extra hashing boards can be left to our creativity to mount and power.  Without cases and power supplies, any engineer working on solving your problems for you would be hobbled by first having to rig up power and mounting solutions .. that would delay any solution by at least a week.  How can you expect your customers to pitch in our time and creativity to solve **your** problem if you don't at least give us what we need (and for that matter, what we paid for)?

(4) You must release the board specs and some warranty that the boards hash if you are going to offer such a degraded solution.  If I blow a board by providing too much power or damage the boards in some way, its my fault.  But if I power them according your spec, mount them in the nice X-3 case that you should provide, and use your power supplies, they should hash at least invidually at the .7 watts/Gh/s that you announced earlier.  Right?

(5)  When the backplane is fixed, you should provide it to us free of charge.  We paid for it, and now we are helping you fix your engineering problems... why wouldn't we get the backplanes when fixed???  The "extra" .5 Th/s in hash boards does not even begin compensate for the pain of having to assemble the system, much less debug and provide communication and control mechanisms ourselves ... and you are asking is to do much more than that.  So the backplanes are part of what we paid for and you have to provide those when ready.  It is offensive that you would not offer that to early adopters who might try to fix your issues.  (Oh, and yes I know you have only announced compensation of 3 Th/s for each X-3, but as addressed previously the "current" market price puts required comp at 4 Th/s to be equal with Spondoolies .. and it will get much higher in comp when you are really ready to ship so the offer of 4.5 Th/s in hash boards is hardly enough even by itself).

Black Arrow, if you can't at least offer the foregoing (and preferably a lot more), how can you expect us to help you?  I (and everyone else) will be **forced** to ask for a refund.  Or, if we do elect to wait it out, your required compensation will skyrocket... by August Spondoolies will be shipping SP30's at $1/Ghs and under .7 watts/Gh/s - and those in the know say Spondoolies will beat their August ship date.  So do you (Black Arrow) really want to chance that?  If you cannot ship by the time Spondoolies starts shipping the Sp30, you will owe EACH customer 6 Th/s in a complete, power supply included, system.  Or refunds now.  So what I am asking (demanding really) for those of us willing to try to fix your issues, is very minimal.  Others may want even more.  But you AT LEAST have to offer something that is workable .. and you haven't done that.

You think the guys working on the hardware have anything to do with forums, questions, whiners, or decisions about compensation?

You need to get up close and personal with Morse code, that is the only way you will be able to communicate, that and frickin laser beams!

Seriously everyone needs to chill out, STFU and wait, no more stupid compensation questions, whining, and boo hoo sob stories. You'll only delay things further. The miners will come, these guys ain't jumping ship, its all hands on deck now for the final push. And just in case you wanna playa hate, I got 30k on the line here too, I'm sure BA will make it right, Jesus told me so.

BA, thanks for this update.

However, I hope there will also be improved compensation for the non-technically skilled customers, as we are nearing June!


That being said, is there any possibility to actually mine with only the PCB and chips without much technical knowledge? How are we supposed to communicate with the PCB?