Quote from: Bitcoinorama on November 11, 2013, 08:14:58 AM

Quote from: texaslabrat on November 11, 2013, 02:50:55 AM

Quote from: Aggrophobia on November 11, 2013, 02:46:36 AM

i'm sorry to correct you, but the initial firmware(0.90) was 890-910W

Exactly, which meant that 850W PSU's were put into an over-spec situation, and some of them handled it badly (eg the Corsair psu's that came back from the safety-shutdown mode and popped capacitors on the hashing boards). Firmware all the way to .94 was pulling 900+ at the wall.

So in answer to CYPER's question...hell yes, it was a bad recommendation. I'm glad I got a 1000W which accounted for KnC's screw-up in power estimation. Granted, later firmware has rendered the issue moot..but for the initial release it was a swing-and-a-miss by KnC on that 850W recommendation for sure.

No. There's several hundred of the V850s running in the data centre. I have 3 on my desk, and have at some point used firmwares for, and not for public release. None of those PSUs haven't had anything thrown at them so far, they cannot handle. The 1000w I mentioned was for something entirely different to the Jupiters sold.

I've never said that the V850's had problems..I am specifically referring to the corsair HX850's which were the center of the "exploding capacitor" investigation. By leaving the recommendation so open (generic "850W" recommendation) without actually testing, many customers were set up for failure in the early days of delivery as you well know. Thankfully, less power-hungry firmware was released later which fixed the issue.