Quote from: eightcylinders on May 22, 2014, 02:28:00 AM

Quote from: blackarrow on May 21, 2014, 08:01:38 PM

"The issue is the SPI communication fails on the current backplane as soon as you attach more than 1 hashboard."

Quote from: rvansch on May 22, 2014, 12:36:37 AM

When problem solving, sometimes it helps to start with the basics:
http://www.embedded.com/electronics-blogs/beginner-s-corner/4023908/Introduction-to-Serial-Peripheral-Interface

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

Quote

SPI's full duplex communication capability and data rates (ranging up to several megabits per second) make it, in most cases, extremely simple and efficient for single master, single slave applications. On the other hand, it can be troublesome to implement for more than one slave, due to its lack of built-in addressing; and the complexity only grows as the number of slaves increases.

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) Huh

?

(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.

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.