- Jan 4, 2001
- 41,596
- 20
- 81
My fileserver PC here has started locking up occasionally lately, after months of stability, being able to run for a week straight before I decided to reboot it. I just tried Prime 95, and in the In-place Large FFT test, I noticed that something toward the power supply and CPU area of the PC is making a high pitched warbling. I was able to record it.
There are three files to choose from, all in FLAC format, to prevent any quality loss that MP3 might introduce:
The first one is as it was recorded, with the background noise in place. The test was initiated at 5s into the recording, and the squeak starts immediately. The first two clicks you hear are from the mouse button as I prepared to start the test; I used the Enter key to actually initiate it though.
Link, 610KB
This is the same file, except that I removed the noise, and maximized the volume again. Just before the sound itself begins in this one, the additional artifacts typical of the noise reduction process begin. But the much higher frequency is then audible.
Link, 352KB
This file had the noise reduction filter used a few times. The noise itself still starts at 5s; the squeaks heard right before it are merely artifacts of the noise reduction process, amplified by the volume maximize process.
Link, 241KB
What the heck gives here? PSU dying? Or might I just tweak it? The voltage in MBM5 (yes, I know, these can be inaccurate) reads as low as 3.15v when the noise is being produced, but when it manages to make it to 3.20v, the noise stops. Might something like a power supply vmod help here?
For the record, I have done work around dangerous power levels before. Right now, there's a 27" TV near me, with its guts on display. Found it out for the trash at someone's house and I rescued it - it seems to have a loose wire inside.
Update: Got the Seasonic. News is not good. 3.3v line is at 3.10V.
:shocked:
There are three files to choose from, all in FLAC format, to prevent any quality loss that MP3 might introduce:
The first one is as it was recorded, with the background noise in place. The test was initiated at 5s into the recording, and the squeak starts immediately. The first two clicks you hear are from the mouse button as I prepared to start the test; I used the Enter key to actually initiate it though.
Link, 610KB
This is the same file, except that I removed the noise, and maximized the volume again. Just before the sound itself begins in this one, the additional artifacts typical of the noise reduction process begin. But the much higher frequency is then audible.
Link, 352KB
This file had the noise reduction filter used a few times. The noise itself still starts at 5s; the squeaks heard right before it are merely artifacts of the noise reduction process, amplified by the volume maximize process.
Link, 241KB
What the heck gives here? PSU dying? Or might I just tweak it? The voltage in MBM5 (yes, I know, these can be inaccurate) reads as low as 3.15v when the noise is being produced, but when it manages to make it to 3.20v, the noise stops. Might something like a power supply vmod help here?
For the record, I have done work around dangerous power levels before. Right now, there's a 27" TV near me, with its guts on display. Found it out for the trash at someone's house and I rescued it - it seems to have a loose wire inside.
Update: Got the Seasonic. News is not good. 3.3v line is at 3.10V.