Topic: KnC news - page 2. (Read 18065 times)

-------------
*edit*
RESOLVED

I'd hazard to guess, January at the earliest.

You could probably pinpoint it more accurately if you knew when any of the competition will start shipping.

Cointerra and Hashfast appear to be just vaporware at the moment, which I think is wierd because I felt sure they would deliver.

The only news I care about is.. WHEN IS THE NEXT BATCH FOR SALE! 

Because Corsair don't have any "engineers" they don't make their own PSU's they get other companies to make them. So perhaps Channel Well who make the Corsair HX850 Gold, couldn't/wouldn't send an engineer over to a small company who aren't their customer in the first place?

Come on guys, as I said before Bitcoinorama is just the conduit of information back and forth, as he said don't shoot the messenger.

Bitcoinorama, on the other hand, don't take it personally if people question the process, just pass the information back to the team and then wait for more feedback. Don't bite back, or even worse feed trolls.

Please. No.

Look i'm reiterating what i was told. Do not shoot the messenger. I didn't have to ask. I did it because people were concerned and wanted to know. This isn't my armchair engineering is better than yours, academically trained I'm a design engineer and if I design an electrical, or electronic device I'd consult an expert in that field. In this case this is precisely what I did as, and when the opportunity presented. So quit the abuse, or I quit bothering to share. The fact is unless we can differentiate between the start up characteristics of each individual PSU on the market after they have all been tripped - and bear in mind in every case they had ALL had their fail-safes tripped and required restarts -then we have no scientific ground for recommending one power supply over another. The only thing we do have to go on is the 'long term reliability proving exercise' w.r.t. the Coolermaster V850 being used throughout the entire hosting facility and working faultlessly the entire time. Thus it has the most assured recommendation as the PSU to use. KnC were actually going to buy the HX850, but couldn't secure enough of a volume in time, and taht appears to have purely been by good fortune.

In any case the number of boards we had fail in this manner is insignificant, less than 10 boards out of the thousands manufactured and in the wild, so this is a mute point, but based upon the statistical significance of the corsair hx850's involvement in the occurrences, and the fact the PSU is known to behave in the manner described, it'd be wise to consider avoiding that for a PSU for another known to behave faultlessly...

Exactly my thought.

I think the question was rhetorical...

Because the Corsair engineers designed and built the HX line... 

+1

You're almost as bad as dogie.

Let me get this straight. A GE engineer claimed the problem wasn't caused by GE hardware, but by some other company's product? Do you see why people would question the results of this "test"? How about you fly in a Corsair engineer and get his opinion?

Actually I sat in a room one Saturday with General Electric's Critical Power Specialist who flew in specifically to address the issue, i'd wager he knows what he's talking about and the HX has fried components at a with a statistically significant reason for concern according to other posts on the internet. It's not solely limited to KnC boards.

When the PSU is tripped and you have to replace the ATX lead, unlike with other brands, or even other ranges of Corsair PSUs, the HX series restarts with current apparently at full load. If the reason that caused the fail safe to cut the PSU still exists that component is then place under greater stress then beforehand, where as typically other PSUs ramp the current after being reset.

If you't care to believe what i'm saying, then fine. I didn't have to ask further questions in the first place. We could have just bluntly said do not use the Corsair HX PSU and be done with it. At the time the thread was so full of competitors trolls attempting to discredit the company for not checking every PSU on the planet despite purposefully recommending one in particular as it had been thoroughly tested.

I'm not attempting to be smarter than an armchair engineer, i'm relaying what I remember from a conversation with an industry expert.

In any case there were less than ten boards out of however many thousand made that suffered from this. It's a total non-issue, but to err on the side of caution we suggest you refrain from using that PSU.

Go through his forum posts with regards to the Corsair HX series PSUs.  He drew the conclusion from a sampling of two which were run from wall AC with no overcurrent protection and googling for the HX850 and motherboard problems, which as I stated you can find for every PSU.

I'm not saying he's lying, I'm saying he's not technical enough to make the statements he is with any validity and folks on this forum should listen to what KNC says officially.  Further KNC should acknowledge the lack of grounding as the source of problems.  I emailed them about this shortly after receiving my machines and have yet to get a response.

I think you are taking it too personally mate, Bitcoinorama just passes on information that he is given on the ground, in the KnC office where he is sitting.

He gets asked questions here and other places, he then goes to the people who are there and asks them, or is in meeting where things get brought up, he then relays information back to us here, all with good intentions.

To accuse him of making it up or lying or falsifying truths is just paranoid-conspiracy-theory-nutso.

I doubt you because you made a non-technical correlation between two KNC customers and an unrelated forum post saying that the same brand had popped a motherboard.  News flash, every manufactured PSU on the planet has popped a motherboard, usually to no fault of its own.  Further the HX850 is a very stable PSU, we use them for 12V supplies in the lab and its what I have run on a dozen Jupiters for over a month.
One thing I did notice, which I overlooked in the product pictures during ordering is that KNC did not properly ground each ASIC board.   The PCI SIG clearly states that 6 pin auxiliary connectors are for 75W draw and 8 pin added two additional grounds, making it _SAFE_ to draw up to 150W.  Note: don't be a moron and read the spec incorrectly, the PCIe edge connector itself can deliver 75W, the 6 pin aux gives an _additional_ 75W, for a total of 150W.

The reason people have been popping boards is the combination of sloppy AC and the boards drawing too much power with not enough ground.  It has little to do with choice of power supply.  I altered my boards by scratching off a portion of the solder mask and soldering the additional two grounds.  Then I sampled the VRM output voltage for two hours.  End result less jitter.  And I'm running off of 208V synthesized AC from a large bank of Liebert equipment.

Long story short, if it didn't come straight from KNC's mouth, don't believe Orama's non-nonsensical, uneducated babbling.  KNC never recommended one brand over another, the only thing they indicated was a minimum capacity, which was 850W.   Until they post somehting on their website, ignore Orama.

Hi all,
650+ Ghash/s is a great, great plus, it doesn't justify the delay of 5-10 days, but compensate
 what  i aspected from nov-shipment.
Go go KNC, let competitors burn in their jealousy, you rock, and your philosophy is pretty closed to BitCoin mentality.

I have 2x V850 +2x V100 that I bought last night, so hopefully these will be enough for 19 boards (2x October Jupiters, 3 Upgrade modules and 2x November Jupiters).

Haha no worries bro. I admire anyone who is forced to speak English as a second language, and does so as well as yourself. I've grown up learning other languages I'm not required to speak, which is just as well, as I'm appalling at those!