TX750 abrupt power cut.

I am experiencing abrupt shut downs similar to this reported on several other threads such as

tx850 random shutdowns

TX750 - PC occasionally crashes/fails to power on

AX850 Randomly shutting down

HX620 sudden shutdown

 

This acts like power supply tripping emergency shut off like a ground fault interrupter. The question is what causes it? Possibly tricky to answer since I have made several changes recently as I will explain below. However the power supply should be more than adequate for current configuration.

 

Suggestions appreciated.

 

More on symptoms:

- Shutdowns have all occurred when computer is idle, some while unattended, or under light load like reading a web page. None while pulling load like video or photo editing or running a stress test.

- Nothing in event log to suggest anything. However log is useful to confirm that an abrupt shutdown occurred at the occurrences when computer was unattended.

- When it occurs, in order to get response from start button on computer case I have to, at a minimum, turn off power to computer with PS switch.

- After that sometimes will boot normally. Sometimes will start but no POST beep and does not proceed further. If so front case switch usually will not respond so have to off power. Once or twice worked after another complete power off on. Other times required bios reset to get post. (Reading a couple threads, perhaps I should have waited longer rather than jumping right to bios reset.)

- Last three occurrences have rebooted normally after only turning PS off-on.

 

Answers to the questions you might ask:

- I don't have another Power Supply on hand that is adequate for this system.

- I have checked all connections, internal and external, used and unused, for any obvious problems. 12, 5 and 3.3 voltages reported in bios, and in windows by AIDA 64 and Speedfan don't vary and are slightly above nominal.

- Wall power seems OK. It has varied this morning between 119.7 and 122.3. I have not monitored until this morning. One remote possibility does occur to me. I am on a rural coop, linear Single Wire Earth Return power system. On these systems voltage can vary with total customer load. At night when load is light, line voltage can rise. Could that possibly trip the power supply's automatic supply voltage selector or trip a shut off?

 

History:

- bought power supply December 2010. No issues until now.

- no significant system changes until recently.

- recently made these changes: Moved one internal hdd to external enclosure with its own power supply. Increased memory from 8gb to 16gb using new matched set. Changed video card. Moved system drive to SSD (Corsair Neutron GTX 240). Cloned existing system, not clean install. I did them in that order letting each run stable for a couple days before doing the next. Problem began after SSD installation, but not immediately. Started 2-3 days later.

- Three notes on SSD but probably hot related to the current issue. 1. Had to rebuild MBR using windows install disk "repair my computer," after cloning to SSD. 2. After change, chkdsk on ssd found two index errors. Old system drive had checked clean with ckhdsk. Artifacts of cloning I guess. OK since. 3. The SSD SATA power connector is not as well supported as on hdds. Plug on SSD doesn't have a complete outside frame so connector can wiggle more and droops if power cable is hanging freely. (Cable is now supported by Velcro tie.)

Just use the link on the left to request an RMA and have it replaced.

Thanks. Yeah, RMA is an alternative, assuming they will ship new before I return old.

 

So you think recent increased full power load was enough to put something over the edge? (Though not at idle because the video card has very good variable power management that drops clock, voltage, everything down when not loaded.)

 

No testing comes to mind?

I submitted an RMA Thursday 2/28. Tech support replied 3/1 asking that I confirm model number (I could see the painted on model name, but couldn't see the label with full model number and s/n with PS installed in case.) I took PS out and replied with requested info same day. No further word from tech support yet.

 

In the mean time the system stopped booting completely and considering the weekend in the way, I couldn't wait for the RMA cycle time, so I bought another PS last night. (200 mile round trip to Fry's, one of the few disadvantages of rural living.) New PS fixed the problem immediately. I will use the rma replacement unit as spare.

 

I thought I'd post some more detail on this hard to diagnose slowly fading partial fail in case it might be useful for others to diagnose in the future. Until I started getting the abrupt shutdowns I hadn't considered a PS problem. Voltages reported from bios screen and from Windows never read low. Because I had just made equipment changes I looked at those first.

 

Once I got the old PS out I checked it with two PS testers.

Ultra which I've had for several years

Antec which I bought last night same time as new PS.

 

The summary is

-----------------old PS---new PS

Antec------------good-----good

paper clip-------good-----didn't check

Ultra------------fail-----good

Computer boots?--no-------yes

Detail:

With the new Antec nothing indicated that the old PS has a problem. It would start, stay on, and all voltages looked OK. No alarms.

 

Old unit also passes the paper clip test.

From Corsair PS troubleshooting page: Q: How can I test my power supply?

A: You can easily test a power supply for functionality with a simple paperclip. First, disconnect all the cables from your motherboard and other devices, but leave the power supply plugged into the wall. Next, bend the paperclip until you have a U shape. Find the 24-pin ATX connector and plug one part of the paperclip into the socket where the green wire ends, and the other end into the socket where a black wire ends. Make sure the power switch on the back of the PSU is on, (should be the I symbol) and the fan should spin up. If it doesn’t, your power supply may be bad."

 

With the Ultra tester (which has only lights, no voltage reading) the old unit would seem to start, but stop after 5 seconds. After stop, lights would fade over a few seconds except 5vsb which remained on. PS would not attempt restart until power cord was removed, 5vsb had faded and power cord replaced. It took 5-10 seconds for 5vsb to fade with power cord out. (The Antec tester can't monitor fade time because the LCD display goes off immediately.)

 

With the Ultra tester the new unit (and also another known good PS) started and stayed on, all lights on. Did not require power cord to be removed after turn off in order to restart. Interestingly with the new unit, the 5vsb light would remain on long after the power cord was removed. After several minutes I gave up watching.

 

With the Antec of course the new unit looked OK. New and old units indicated the same other than a couple readings differing by 0.1v.

 

So both the paper clip and the Antec tester fail to indicate the old unit as bad. But the old Ultra tester does, as did the computer, eventually, once the old PS got bad enough. Exactly what is failing isn't clear. The faster fading 5vsb on the old unit might or might not be a useful clue. Maybe it suggests a failing capacitor that could cause a 60 hz flutter in the DC voltages that is too fast for meters to show. But I don't have anything to display waveform, so a that's just a guess. And it's not clear why that would affect only the Ultra tester. It would be interesting to know what a lab test finds if Corsair bothers to test it.

Got rma approval Monday 3/4. Shipped failing unit UPS to Corsair Tuesday 3/5.

UPS Delivered to Corsair, per UPS tracking, Monday, 03/11/2013 at 9:22 A.M local time.

Replacement shipped 3/12/13 evening. RMA status page never did say received.

Delivered to me 3/18/13.

So 18 days from the day I submitted the RMA request. (Did not use premium shipping either way.)

 

Replacement appeared to be new unit (not rebuilt) in box with adhesive flap seals and plastic shrink wrap. Same model as original obviously, but the revised version with low power fan cut off.

 

Since I had already bought a new PS rather than wait, I only tested the RMA it with my two PS testers, not installed in computer. Testers did not indicate any problems. Neither tester does a load test, so fan never started. Put this replacement unit back in box and on the shelf as spare.

I've got the same problem! Any chance to solve it? Or I ll have to ask a new one? But my problem is i've bought it on USA and i'm in Brazil now ;/

There is not much you can do with a bad PSU but replace it.

 

There should not be a problem shipping you one to Brazil.

I've got the same problem! Any chance to solve it? Or I ll have to ask a new one? But my problem is i've bought it on USA and i'm in Brazil now ;/

In hindsight my unit was slowly getting worse and contributing to more and more mystery problems which I didn't initially recognize as caused by the power supply. If your problem is like mine, I would replace it now before it causes problems with other components or data corruption. Ideally get an RMA, but do replace.

My only problem is the shipping! Is the same price to send to USA and buy a new one haha! But I ll send it ;/ Not fair selling this product with problems and I don't do anything about it!

 

Thanks for the fast answer guys :)

JYraya, how old is your PSU?