Well, ran out of cp2102 usbs
edit: now stable at 252 GHs
315 Watt / 252 GHs = 1.25 W/GH
Your prototype will be posted on wednesday. tomorrow will be bank holidays for end of WW1.
I had a look at a few things regarding the hardware of the blades.
Each chip have an address defined on 8 pins.
each board will have the following addresses:
00-08
10-18
20-28
...
F0-F8
Linking both boards in serial shouldn't work since there will be 2 chips with the same address.
And I'm not sure parallel would work for the same reason.
If I understand correctly, the chips aren't pushing the results but are waiting for the result to be read from them.
if you ask for chip A8 result, you will have 2 results returned, and I'm no sure it will be something readable by software or some random garbage mixed from both chips.
Initializing the chips might not work for the same reason, since you are expecting an answer from the chips.
I've just received a few parts to move on next step and work on the 8 hashing boards controller.
I'll also soon be able to make my own fine pitch mylar stencils. It will make the soldering job way easier.
Next on my worklist, something else related to the antminers. (sub 0.9w/GH at the wall on first gen chips)