Corsair deny all knowledge - AX1200i software broken (probably applies to AX860i and AX760i as well)

Just a warning for others considering purchasing the AX1200i. This issue almost certainly applies to the AX860i and AX760i too since they use the same software (Corsair Link). The monitoring and control software is pretty much useless.

The PCIe rails get stuck, and give wildly ridiculous current indications that have no relationship to reality.

I RMA'd my first AX1200i because of it, here's the thread: http://forum.corsair.com/v3/showthread.php?t=111282

Then it seemed OK for a few days, until I noticed the replacement was doing it again: http://forum.corsair.com/v3/showthread.php?p=608540

Then I asked in http://forum.corsair.com/v3/showthread.php?p=612174 if Corsair could reproduce the issue I was seeing.

A Corsair employee told me:

So they say they can't reproduce the issue and I am the only user reporting it, but I found this video (a Corsair promo with George Makris from Corsair demo'ing the AX1200i to a reviewer this June) where it's happening: http://www.youtube.com/watch?v=Yv7oLWF4dZU&feature=player_embedded

Screenshot:



Note PCIe 1 is reading 56.1 Amps = 673 Watts, while the total system power consumption is reported at 152 Watts.

I then did some more digging and I found a second Corsair employee, PR director Rick Allen, pointing at the faulty power reading while describing how you can monitor the PCIe currents:

http://www.youtube.com/watch?v=ZD8u7-vMSzs#t=2m37s



In that same video, at At 2:28, it's doing it on PCIe1, PCIe2 and PCIe3 with 55.8 Amps reported on each (a total of over 2 kilowatts) while total power output reads 174 Watts.



So supposedly 2 different Corsair employees have demo'd this PSU while it was doing this, and neither of them noticed the software reporting that a single 8-pin connector is carrying over 50 Amps.

Supposedly their testing hasn't revealed this issue before they started selling these PSUs, even though the issue shows up on a system they are using to demo and market the product.

More here: http://forum.corsair.com/v3/showthread.php?p=612174&posted=1#post612174
Edited by BorisTheSpider - 5/13/13 at 2:38pm

Hi,

Yeah, it also makes the overcurrent protection (the selectable multi-rail mode) useless - if you can't set a trip-point because the values falsely report it being exceeded (and hence trip the OCP) then that feature becomes useless too.

Generally, the monitoring seems bad in other ways too - the PCIe current values jump around a lot instead of smoothly varying, the reported efficiency actually differs from what you'd calculate from the reported power-in and power-out values, often reading 100% and the PCIe rails often read zero at idle - that might be correct however, since I don't know if that's because GPUs actually draw nothing at idle (instead drawing from the 75 watt capacity of the PCIe slot itself), or whether it's another problem in the monitoring system.

I don't know whether it's fixable in software - at the moment, it's impossible to say whether the problem is originating in the PSU itself (in the DSP and monitoring hardware) or is being introduced in the link dongle, or PC.

I will update as more details emerge.
Edited by BorisTheSpider - 11/9/12 at 3:32pm

Thanks Murlocke, but I already PM'd Corsair George (if that's who you're referring to) over 3 weeks ago asking him if he could look into it. I didn't get a reply.