Graywolves
04-26-2009, 04:22 AM
Since this has more to do with PSUs, this comment is more at home in the power supply section.
Mods...feel free to move it to another place if you feel it is appropriate. Just want to give others a huge heads up.
Recently, my parents, who live a couple hours away from me, had a hardware failure and called a local repair guy who supposedly knew what he was doing. Normally they would call me, but as I can be enormously busy most times, it was easier for them to call this other guy. He has apparently got them out of a few messes ok.
The basic symptom of the problem was a pc that crashed (blue screened I suspect from a diagnostic report when I did get it up and running), and failed to reboot. No POST beep. All fans running, hard drive initiallized. Upon close visual of the components, some capacitors were bulged...this was what the other tech zeroed in on as the culprit and immediately recommended a new mainboard, stating the capacitors could not be replaced because they were rare and unavailable. Of course, the repercussions of that also required a new video card, memory and cpu and PSU. My mom was worried about the cost and decided to get dad to call me and consult.
I am an electronics engineer by schooling. I am well aware of the bulging capacitor issues from experience. I am also aware of how to fix them and work around the issue of special rated cap unavailability. So, yesterday I visited to take a look. And indeed found a dead pc with bulging caps...though I have to admit, I've seen worse, and those were still operating. The PC itself was active, power was making it to the mainboard, all the fans were operating, just no POST beep to tell me what was up. The speaker was attached, so it should have worked. A visit to a local components supplier, I came out with a dozen caps...
Note for those of you not sure. In the case of the caps I needed, those were 6.3v at 3300 microfarads. Since I could not immediately source those without waiting several weeks, I took some 10V caps at 2200 microfarads, soldered two in parallel together, feeding one of the two into the other holes of the damaged caps after I removed those and the solder. Caps in parallel, the farad rating is added, giving me 4400 in this case. So long as I maintain a value above that of the originals, there should be no problems. After some creative placement of the new caps (these are the ones around the CPU, very little play room regarding the HSF), I placed the board back in.
Still no boot.
Upon closer inspection and reading the user manual, I found that the sole memory stick was in the wrong slot (a 3 slot dual channel design). It was in slot 3, when it should have been slot 1. I suspect the other tech swapped that...trouble shooting maybe, though I am suspicious. After that, I got a POST beep. I put everything back together, and booted the computer.
It would not boot. I reset the CMOS. I got it to boot and reset the CMOS settings, and finally got into the OS. A few initial tests, I found the machine really flakey. I thought it might be the cap repair, but continued my inspection. On some of my reboots, the PC would hang and not get into the first POST display screen and no beep. On a couple that were successful, I had hard drive failure after the initial startup POST and components screen displays.
Again, I am thinking maybe I made a poor repair with the caps. Undeterred, I decided to boot into the OS and installed SISoft Sandra. Most people use this program for benchmarking, but I find it an invaluable tool for diagnostics.
In this case, since power seemed to be the issue, I wanted to view what was happening. SISoft has an Environment Monitor. Starting that, and giving it a 1 second refresh, I could monitor directly what was happening. I saw two things after a few moments of observing the displayed results.
One, the +5V SB (standby voltage) was ALL over the place, dipping as low as 1.3V on up to 7V peak. The second, the +12V rail was averaging 11.2V, with a minimum of 10.9V.
The standby voltage is the voltage that is always live on a mainboard. It provides power to key sections, like the main power switch on the front of the tower, or for components that make use of the wakeup feature, such as a LAN card, or mouse/keyboard activity for power saving. So THIS solved my issues with regard to why the PC was not behaving properly on reboots.
The 12V rail was also partially responsible for that but primarily also responsible for the hard drive failing to boot. Most PSUs have a tolerance of 5% +- the rated voltage along any of the rails. This one was well above that...7-10% deviance below.
The solution: Replace the PSU which was barely sufficient a supply for the needs of the whole PC...it had several new additions installed that sucked up the breathing room of the PSU wattage, and it had finally decided it had enough.
So, the lesson here is, don't immediately assume a dead or malfunctioning PC is because of the Caps. Though I have no doubt the caps were on their way to failing, I do not believe they actually failed. Had that other 'tech' taken the time to use diagnostic approaches like resetting the BIOS and removing major components, he would have been able to reboot, and perhaps go from there to using a diagnostics program like SiSoft to check how the hardware is actually performing. I had taken the assumption he had done those measures and (sort of) wasted my time replacing the caps.
Anyway, I hope that this information can be used by others in this forum down the road to keep them from needlessly wasting money on an upgrade. Upgrading is ok, but ONLY when ya need it for the sake of needing more performance. My parents did not in this case.
Mods...feel free to move it to another place if you feel it is appropriate. Just want to give others a huge heads up.
Recently, my parents, who live a couple hours away from me, had a hardware failure and called a local repair guy who supposedly knew what he was doing. Normally they would call me, but as I can be enormously busy most times, it was easier for them to call this other guy. He has apparently got them out of a few messes ok.
The basic symptom of the problem was a pc that crashed (blue screened I suspect from a diagnostic report when I did get it up and running), and failed to reboot. No POST beep. All fans running, hard drive initiallized. Upon close visual of the components, some capacitors were bulged...this was what the other tech zeroed in on as the culprit and immediately recommended a new mainboard, stating the capacitors could not be replaced because they were rare and unavailable. Of course, the repercussions of that also required a new video card, memory and cpu and PSU. My mom was worried about the cost and decided to get dad to call me and consult.
I am an electronics engineer by schooling. I am well aware of the bulging capacitor issues from experience. I am also aware of how to fix them and work around the issue of special rated cap unavailability. So, yesterday I visited to take a look. And indeed found a dead pc with bulging caps...though I have to admit, I've seen worse, and those were still operating. The PC itself was active, power was making it to the mainboard, all the fans were operating, just no POST beep to tell me what was up. The speaker was attached, so it should have worked. A visit to a local components supplier, I came out with a dozen caps...
Note for those of you not sure. In the case of the caps I needed, those were 6.3v at 3300 microfarads. Since I could not immediately source those without waiting several weeks, I took some 10V caps at 2200 microfarads, soldered two in parallel together, feeding one of the two into the other holes of the damaged caps after I removed those and the solder. Caps in parallel, the farad rating is added, giving me 4400 in this case. So long as I maintain a value above that of the originals, there should be no problems. After some creative placement of the new caps (these are the ones around the CPU, very little play room regarding the HSF), I placed the board back in.
Still no boot.
Upon closer inspection and reading the user manual, I found that the sole memory stick was in the wrong slot (a 3 slot dual channel design). It was in slot 3, when it should have been slot 1. I suspect the other tech swapped that...trouble shooting maybe, though I am suspicious. After that, I got a POST beep. I put everything back together, and booted the computer.
It would not boot. I reset the CMOS. I got it to boot and reset the CMOS settings, and finally got into the OS. A few initial tests, I found the machine really flakey. I thought it might be the cap repair, but continued my inspection. On some of my reboots, the PC would hang and not get into the first POST display screen and no beep. On a couple that were successful, I had hard drive failure after the initial startup POST and components screen displays.
Again, I am thinking maybe I made a poor repair with the caps. Undeterred, I decided to boot into the OS and installed SISoft Sandra. Most people use this program for benchmarking, but I find it an invaluable tool for diagnostics.
In this case, since power seemed to be the issue, I wanted to view what was happening. SISoft has an Environment Monitor. Starting that, and giving it a 1 second refresh, I could monitor directly what was happening. I saw two things after a few moments of observing the displayed results.
One, the +5V SB (standby voltage) was ALL over the place, dipping as low as 1.3V on up to 7V peak. The second, the +12V rail was averaging 11.2V, with a minimum of 10.9V.
The standby voltage is the voltage that is always live on a mainboard. It provides power to key sections, like the main power switch on the front of the tower, or for components that make use of the wakeup feature, such as a LAN card, or mouse/keyboard activity for power saving. So THIS solved my issues with regard to why the PC was not behaving properly on reboots.
The 12V rail was also partially responsible for that but primarily also responsible for the hard drive failing to boot. Most PSUs have a tolerance of 5% +- the rated voltage along any of the rails. This one was well above that...7-10% deviance below.
The solution: Replace the PSU which was barely sufficient a supply for the needs of the whole PC...it had several new additions installed that sucked up the breathing room of the PSU wattage, and it had finally decided it had enough.
So, the lesson here is, don't immediately assume a dead or malfunctioning PC is because of the Caps. Though I have no doubt the caps were on their way to failing, I do not believe they actually failed. Had that other 'tech' taken the time to use diagnostic approaches like resetting the BIOS and removing major components, he would have been able to reboot, and perhaps go from there to using a diagnostics program like SiSoft to check how the hardware is actually performing. I had taken the assumption he had done those measures and (sort of) wasted my time replacing the caps.
Anyway, I hope that this information can be used by others in this forum down the road to keep them from needlessly wasting money on an upgrade. Upgrading is ok, but ONLY when ya need it for the sake of needing more performance. My parents did not in this case.