Is this likely a bad caps issue?

[einhverfr]

Hi all... This is my first post here.....

I run a computer consulting and repair firm in rural Washington State. I have seen a very weird issue and I am wondering whether people here attribute it to bad caps.

I have a customer who has a system (ASROCK motherboard) which worked fine for a short period of time (three months). Then winter came. The computer would randomly power off several times a day. All other computers in this area would work properly (though the print server could not handle being on the same outlet as a laser printer which may indicate the quality of electrical wiring in the building). Anyway when these black screens occur, the main fans stop running and the computer would not power on until it has been disconnected and reconnected from the power.

Replacing the power supply did not help. If I take the unit to my office the problem does not occur. So the current situation (for the last eleven months) has been that we have used a UPS to stabilize the electrical supply for this unit. However, something appears to be making this unit extremely sensitive to electrical issues-- far more than other units. I am wondering if faulty capacitors might be responsible for this extreme sensitivity to such issues.

[willawake]

Re: Is this likely a bad caps issue?

the answer is that yes caps in the motherboard and psu are there to filter the power. What motherboard is it exactly? what caps are on the board 470uf and above? ASROCK is the cheaper line of asus if i remember correct.

(though the print server could not handle being on the same outlet as a laser printer which may indicate the quality of electrical wiring in the building).

i dont like that. what does it do? does it trip the breaker or does it just not work?

3 months is kinda early for probs to start happening with the caps though. You say winter came so are they using electrical heating now? perhaps even the utility power has become worse due to increased consumption.

you might consider recommending to have an electrican check out their system. i can imagine how that suggestion will go down. what you really need is to rent borrow or steal a fluke ac meter and have it monitoring the outlet for a while.

the problems stopped with the ups right? i guess you must be using a smart-ups or similar cos many dont do much power conditioning. if you were using a smart ups you could get some voltage monitoring going.

offices should have all pcs on ups anyway. if it always works fine on the ups and not off it then that gives you an idea.

can you get another pc working fine on that outlet?

anwyay tell us the brand of caps on the board 470uf and above and we can tell you if they are crap or not. you could recap the board and see how that goes. gonna be difficult to get an rma on that board. you could check the caps with an oscilloscope also.

[linuxguru]

Re: Is this likely a bad caps issue?

A checklist of things to check:

1) The quality of the earthing (grounding), including the mains wiring, socket, AC cable, etc.
2) The PSU (SMPS) on the system - they often have bad caps.
3) The motherboard capacitors (a quick visual inspection can reveal some problems, but not all
bad caps actually bulge or leak before they fail).

Note that 2) and 3) are not exclusive - both can contribute to the problem.

[MD Willington]

Re: Is this likely a bad caps issue?

Asrock mainboards do not work well with cheap power supplies, some also have cold boot issues, what is the exact model of the mainboard?

For the most part Asrock boards are a great value and have very few issues.

A really good source of Asrock information is RHCF:

64 bit asrock: http://www.rhcf.com/sisubb/ultimatebb.php/forum/20.html
32 bit asrock: http://www.rhcf.com/sisubb/ultimatebb.php/forum/36.html

MD


PS what part of WA?

[einhverfr]

Re: Is this likely a bad caps issue?

i dont like that. what does it do? does it trip the breaker or does it just not work?


3 months is kinda early for probs to start happening with the caps though. You say winter came so are they using electrical heating now? perhaps even the utility power has become worse due to increased consumption.

you might consider recommending to have an electrican check out their system. i can imagine how that suggestion will go down. what you really need is to rent borrow or steal a fluke ac meter and have it monitoring the outlet for a while.

I have recommended an electrician check out the wiring. The UPS did resolve the issue at least temporarily but I am worried about simply delaying a more serious problem. I have also done the following tests in this area:

1) checked out the AC voltage with my multimeter with the unit plugged into the other socket on the same outlet set and watched while the problem occurred (nothing showed up).

2) Checked out the DC output of the 12V and 5V power outlets on the PSU while the problem occurred (nothing showed up).

3) Checked out earth ground issues and tried plugging the system into an outlet with a known good earth ground and nothing else on that breaker (did not fix problem).

4) Replaced the 400 watt cheap PSU with a Startech 450 watt PSU. This did not solve the issue either.

5) Plugged the system into the outlet on my office and let it run all night (no problems during a time when we would get problems in their office every 10-15 minutes).

My working theory is that there is a degraded quality of AC coming from the PUD's transformer and that in the winter, they may get some sort of dirty power and/or line sags of a sort that is making it past the PSU (even with a new PSU though). My working theory is that these line sags or other anomilies are fast enough that the capacitors on the other units are able to compensate but this one is not due to reduced capacitance of the capacitors on the motherboard. However, this seems to be possibly incomplete. It is clearly electrical though as the UPS did resolve the issue.

As for the file and print print server, it was originally plugged into the same power strip as the HP Laserjet 5, and would kernel panic when the laser printer would warm up. Moving them to outlets on separate lines solved that problem. Of course, although laser printers draw a lot of electricity when they warm up, the electrical wires are supposed to handle it :-)

As for utility power issues, we know that these are a factor too. I have another customer two doors down who had to put line conditioners on all their POS terminals because they would randomly crash despite having had the building recently rewired for commercial kitchen equipment. But it is clear that something is making this specific unit more sensitive to these problems than the other units around it. Thanks for confirming that my suspician might be accurate.

I will open up the unit tonight and get the name of the motherboard model and the brand of capacitors and post more.

[einhverfr]

Re: Is this likely a bad caps issue?

Brands of caps seem to be (I think) something like KQS (not sure about this one) and RLP. But I left the model number in the office so I will have to get that tomorrow.

FWIW, I don't think it is a standard case of bad electrolyte recipes. I am thinking that it is more likely that it might be some other manufacturing issue that has rendered the capacitors less effective than normal.

[linuxguru]

Re: Is this likely a bad caps issue?

RLP is probably a particular series from OST (the manufacturer of the caps). OST are a mixed lot - they're definitely implicated in a lot of the early failures from 2001/02, but their recent caps are probably slightly better. In order of reliability (from bad -> not so bad):

YEC, Gloria -> GSC, Wendell -> Choyo -> OST (RLP, RLX) -> G-Luxon, Capxon

I'd replace the OSTs only if they're visibly bulging - otherwise, I'd look for other causes first.

[einhverfr]

Re: Is this likely a bad caps issue?

Well, they are not bulging. a fairly careful evaluation today, I have decided that the caps don't seem in any danger of blowing and that the current state is probably worth allowing to continue (things are stable with the UPS).

As I said, this doesn't look like a typical failure but looks like something capacitor-related. In other words, combined with noisy power input, these caps are not working properly to stabilize the power but they don't appear to be leaking or bulging. In other words maybe they just don't have as much capacitance as they are supposed to due to a manufacturing defect or something.

So the next question is, given that the symptoms appear to be under control and that we don't see bulging or leaking, is it safe to run the system in this state or are we asking for trouble later?

[willawake]

Re: Is this likely a bad caps issue?

if you are happy leaving it then ok. personally i would swap it out with another mobo perhaps similar model to what the other comps are using. better to deal with probs finally than have repeated failures and reputation goes downhill.

i read your power troubleshooting but what can i say......perhaps it is the caps, perhaps they cheaped out on some other board component, perhaps they left out too many caps and other components from the original design that the board is now only stable with optimal power.

you wont know if it the caps until you recap it or check it out with a scope

[einhverfr]

Re: Is this likely a bad caps issue?

Thanks everyone for the help. This is likely to be an issue that I will continue to watch. But in the end I think that since it has been stable with a UPS for almost a year, I will leave it as it is.

At this time, I am concerned that it might not be electrolytic caps but could be a short across other caps, so the only way to really be sure it is not a cap is to replace all of them (not just the electrolytic ones). This seems to be more work than I want to put into it at the moment given that everything has been stable for about a year. In the mean time, I will watch and wait and re-evaluate from time to time.

Again, thanks for the help.