AX1200i random shutdowns with red light

Recently I've finished building new PC which uses brand new AX1200i PSU - basic system spec below. After finishing the build I have installed OS and did some full load test (both GPUs and CPU stressed using Prime95 + FurMark). After that I've continued using this PC without any problems for about week.

I didn't change anything related to the hardware and didn't have any spare time to install more software either, the only difference I can remember is that the OS (Windows 10 64-bit Pro) got updated from version 1809 (the one I've installed) to 1903. Anyway next the the random shutdowns started. It looks like that: PC is running IDLE or near it (like browsing web or something like that, no heavy load on CPU or GPUs), and there there goes a random shutdown - status light on the PSU goes from green to red, PC power light is still on (now that's odd) but everything else like case fans are off. Power button can not be used to turn the PC on again. I have to use the power switch on the PSU to turn it off, wait few second and now I can use the PC power button to turn it on again. Yet if I do so, it runs for about 1 second, shut downs again including status light on the PSU going red. After waiting for about 10 to 25 minutes (this amount of time is a bit different every time) I can turn the PC on normally.

The PSU is connected using dongle and internal USB 2.0 hub to the motherboard 2.0 only one internal port. I've got Link software (latest version) installed and I believe I was able to notice the reason of those shutdowns.

The total power consumption while IDLE is about 200 W of which about 180 W is consumed using 12 V voltage. I've got it running in multi-rail mode with no custom over-current protection setting for each output. Only one of the outputs (CPU/GPU output #4) is reporting any current consumption (it seems that anything below 5 A is showing as 0 A so there is some other small consumption on other outputs too) of about 5 A. I believe since this equals 60 W it must be the one powering the main GPU, as it should IDLE near this power level. This current consumption changes between 5 A and 6 A from time to time, but after 10 to 20 minutes it steadily starts going up. After few seconds the reading is 8 A, next 12 A, 17 A and then it goes crazy with 3 consequent readings going something like: 27 A, 45 A, 59 A and there goes the shutdown. What is odd the total power consumption on the 12 V voltage and the total power consumption of the PSU does not go up at all. All the temperatures of CPU and GPUs are not increasing in any unusual way either.

So it seems there is a malfunctioning device connected to the CPU/GPU output #4, right? Assuming that the numbers in Link software corresponds to numbers below the PCIE cabling connections of the PSU it would seem that CPU/GPU output #4 is the 6-pin cable powering my main GPU. This GPU has also additional 8-pin plug connected to the PSU CPU/GPU output #5. To test this without removing all the cabling from the back side of my case I have connected my last unused CPU/GPU output (#7) using new cable going directly to the 6-pin socket of the main GPU. Guess what - current consumption on output #7 is 0 A while on output #4 it is still 5 A by default, next it increases slowly and the random shutdowns continues.

Any ideas? I'm not sure is this a case of faulty PSU (and there are a lot of reports here on random shutdowns of AX1200i model) or not.

Basic system specs:

- CPU -AMD Ryzen 7 3800X,

- GPUs - 2 x Sapphire Radeon 5700 XT Pulse,

- MB - Asus TUF X570 Plus,

- CPU AIO cooler - Corsair H115i Pro,

- 1 x M.2 SSD, 1 x SATA SSD, 4 x SATA HDD.

I've got it running in multi-rail mode with no custom over-current protection setting for each output. Only one of the outputs (CPU/GPU output #4) is reporting any current consumption (it seems that anything below 5 A is showing as 0 A so there is some other small consumption on other outputs too) of about 5 A. I believe since this equals 60 W it must be the one powering the main GPU, as it should IDLE near this power level. This current consumption changes between 5 A and 6 A from time to time, but after 10 to 20 minutes it steadily starts going up. After few seconds the reading is 8 A, next 12 A, 17 A and then it goes crazy with 3 consequent readings going something like: 27 A, 45 A, 59 A and there goes the shutdown. What is odd the total power consumption on the 12 V voltage and the total power consumption of the PSU does not go up at all. All the temperatures of CPU and GPUs are not increasing in any unusual way either.

 

So it seems there is a malfunctioning device connected to the CPU/GPU output #4, right? Assuming that the numbers in Link software corresponds to numbers below the PCIE cabling connections of the PSU it would seem that CPU/GPU output #4 is the 6-pin cable powering my main GPU. This GPU has also additional 8-pin plug connected to the PSU CPU/GPU output #5. To test this without removing all the cabling from the back side of my case I have connected my last unused CPU/GPU output (#7) using new cable going directly to the 6-pin socket of the main GPU. Guess what - current consumption on output #7 is 0 A while on output #4 it is still 5 A by default, next it increases slowly and the random shutdowns continues.

 

This came up in another context the other day. I always use single rail mode and thus cannot see the individual rail current levels, nor do I sit in the settings where they are displayed in iCUE. I see the same seemingly random current fluctuations, with most rails reading 0A all the time at the desktop, including the 24 pin. Loading the GPU (powered by GPU 1, 2, and 3) causes the current level to hop back and forth between 1 and 2 with 3 continuously showing zero at a 110% power level. I am not sure how much faith you can put in these readings.

However, you may want to put the PSU in single rail mode and see if this changes the shutdown behavior. It reads like a PSU issue and others in the past have had trouble with multi-rail mode, only to have the problem vanish on switching over.

Thank you for the replay. Today I've tested switching this PSU to the single-rail mode via Link software.

Just to be sure the problem persists, first I run it in multi-rail mode (so without any changes) and there was again shutdown after 17 minutes of the PC running basically at IDLE. This was preceded by current reading on rail #4 steadily increasing in the last few seconds and reaching about 55 A.

Next, after waiting to be able to turn the PC back on, just after the OS loaded I went to Link settings and found out there is a usual reading of 5 A on rail #4. I've switched the mode to single-rail and all the rails over-current protection toggles went to off state and they got greyed-out too. Immediately after doing so the reading on rail #4 went from 5 A to 50 A. It was increasing for about another 20 seconds and stopped at 69 A, and then after 5 seconds more there was the usual shutdown. This time it took 4 minutes in total from boot till shutdown.

Last try. Waiting again till it can be turned on, OS boot, checking in the Link settings if the single-rail mode is still active and it was. After 13 minutes of running at IDLE there was again increased current reading going on at rail #4 and at about 59 A there was shutdown.

You know what is really silly about that? I have nothing connected to the CPU/GPU output #4. I have moved the plug powering my main GPU to CPU/GPU output #7 for good. All this time CPU and both GPUs load was near 0 %, temperatures looking good and the total 12 V output of about 12 A with the total power consumption of the PSU near 200 W.

I was reading through the various posts related to AX1200i on this board and actually there are quite a few of them reporting high current consumption (of about 50 A) getting stuck on one of the rails and over-current protection tripping at very low power consumption.

I would assume at this point that this one PSU is faulty and go for RMA, yet:

- I'm not even 100 % sure is this all PSU fault as it was working without any problems for one week before all this started,

- even if I got it replaced I'm not sure if I'm still interested in using PSU with a price tag of 379 € (actually I couldn't buy it cheaper than 439 €) that can potentially do this again - look at the similar problems described on this board for the AX860i model, maybe all this "fully digital" stuff of the whole AXi series is flawed to some degree?,

- I've asked at the local retailer where I bought this PSU and it will be not accepted for local return without returning every piece that was shipped with it such as manual, cardboard insert and floppy cable adapters that I've all thrown away after the first week, when seeing that the PSU was not DOA so the need for replacement was very unlikely (I was using my previous PSU at my old PC - which was by the way AX1200 - for 7 years without any problems),

- even if Corsair would directly accept my return without those items, I would first need to remove all the tidily done cabling and still wait up to few weeks with my main PC laying dead.

At this point I'm actually thinking more of buying something else that could be swapped easily, such as HX1200i (I believe the cables should be compatible), testing it to be sure the current PSU was indeed faulty, trying to RMA it and if successful selling it as new for half a price or so to get the price difference between those models back.

Typically when you RMA the PSU through the manufacturer, you keep the cables and only ship the main unit and in turn only receive the main unit back. If it is shutting down in single rail mode and there is no obvious wiring issue on your end, then you likely need to take action.

I do think you should contact Corsair Tech Support through the Ticket System regardless of your plan. At least then the issue is logged.

*BTW - HX1200i cables are the flat ribbon type and newer than the AXi standard braided. They are both type 3, but I am not 100% certain they are cross-compatible.