PDA

View Full Version : A visit at my friend



carlsson
August 19th, 2006, 09:35 AM
Today after our marching band gig, I drove a friend home and was invited to his little apartment. Compact living gets a new meaning, I can tell you. Things everywhere, baskets mounted above the bed with all kinds of stuff including a "standalone" VCR. Among what amazed me most was his "collection" of vintage and semi-vintage PCs in a wooden rack. One disconnected IBM XT, one AT. One PS/2 about to be dumped. One noname 386 in a tower. At least three small 486 desktops and one Pentium 166 without top cover. The thing is that he has linked together everything from the 386 and upwards in some kind of Linux network, using one of the 486:s as a firewall/router and the other computers to balance load. He recently acquired an 1.1 GHz Athlon, so maybe he'll take some of the other machines out of use. Everything is in pretty dull physical condition. I wish I had brought a camera, it was quite a sight.

I donated some small (20 x 5.5 x 3.5 cm) power supplies from Mac LC computers many years ago, and I got one back. My idea was to use it to power a Commodore disk drive. While the voltages and maximum currents are enough, it doesn't seem to be stabilised, so maybe not suitable to be used. 5V differs between 5.11 to 5.27V, 12V between 12.02 to 12.45V. It also makes a faint clicking noise while running. Could be some weird fan.

Maybe after I told him those old IBM PCs could be worth something, he'll look it up and try to sell rather than dump it. Of course, if things are broken or scavenged for parts, it may not be so interesting anymore.

carlsson
August 19th, 2006, 10:00 AM
It also makes a faint clicking noise while running.
It is a fan-less design, so that's not it. I clocked the ticking, and it seems to be about 230 bpm. Is it a co-incidence that the wall outlet is at nominal 230V AC, or could the electronics inside (some capacitor constantly loading and emptying?) actually tick along at a speed equal in beats per minute to the input voltage? :? :stupid:

Chris2005
August 19th, 2006, 11:12 AM
crikeys dunno mate. Could have something to do with the voltage I guess, but more often then not noises have more to do with the line frequency then the voltage. Anything is possible though. The Electronics_101 yahoo group is an excellent resource.

modem7
August 19th, 2006, 06:29 PM
I'll bet my 5150 that the 230bpm and the 230VAC figures are a coincidence.
The 230bpm will probably be the a harmonic or subharmonic of some frequency happening inside the PSU. For example, if the PSU is a switch-mode type, it could be running at around 230Khz with some component 'vibrating' at the 230Hz subharmonic.

dreddnott
August 19th, 2006, 08:13 PM
Read that again, it's two hundred and thirty beats per minute, rather than per second. 230Hz would be around Bb below middle C, a pleasant tone rather than a clicking sound.

If Carlsson's BPM measurement technique is very accurate, we can assume about 3.8 HZ for the sound he is hearing (and therefore probably nothing to do with 50/60Hz AC frequency).

It would probably also have nothing to do with a switching PSU, as the frequency is indeed very high (typically over 20kHz) and subharmonics (also known as 'resultant tones') cannot be generated by a single tone, rather requiring a combination of harmonics to be perceived by the ear.

Typically resultant tone effects are not perceptible below 16Hz, let alone 3.8Hz.

Unfortunately I don't have any more positive suggestions of what it might actually be.

Terry Yager
August 19th, 2006, 09:09 PM
Soundz like a pretty kewl beat, if ya could amplify it and lay down a melody track to go with it...

--T

NathanAllan
August 19th, 2006, 10:27 PM
If you think it's the power coming in try a line filter, or even simply using a different cord. Ferrite around the end of the plug is a thing to do. But I'm with modem7, I don't think that it's the power going in but more shielding is never a bad thing. Aluminum tape around the power cord is an idea. And it may need a new cap like you say. You could quick-fix the cap if that's the problem) by putting one in the line leading to the plug but you'd still hear the cap in the supply, and you'd have a bulge in the wire.

Interesting that you should come across this type of problem, right when I'm learning about shielding and stuff.

modem7
August 19th, 2006, 10:50 PM
it's two hundred and thirty beats per minute, rather than per second.
Yes, that was a horrible misread by me.


It would probably also have nothing to do with a switching PSU, as the frequency is indeed very high (typically over 20kHz) and subharmonics (also known as 'resultant tones') cannot be generated by a single tone, rather requiring a combination of harmonics to be perceived by the ear.
A switchmode PSU does not run in a sinusoidal fashion (ie. single tone). The switching section switches from one state to the other - close to a square wave (a very corrupted square wave). Thus there are plenty of harmonics. And the frequency of operation (the frequency or repetition rate of the 'corrupted square' wave) is constantly varying. If you look at that switching waveform on a frequency spectrum analyser, you'll see that switchmodes are electrically very noisy.

And so generation of subharmonic currents/voltages in a switchmode PSU is a given. Of course, for someone to hear those, there has to be a faulty electrical component somewhere to convert those to a sound waves. Transformers are the main cuplrits.

What could be causing a clicking sound at such a low repitition rate of 4Hz?

Puting my mind to it (difficult to do on a Sunday), I can think of a scenario. Arcing. It's possible that an electrical component in the PSU is breaking down. Arcing can sound like a click. Either that or a click beetle has crawled into the PSU!

dreddnott
August 20th, 2006, 09:16 PM
That's true, a capacitor could be building up enough charge to arc across something at around 4 times per second.

From what I read, carlsson, you haven't yet plugged the disk drive into the clicking power supply. When you check the DC output with your voltmeter, do the voltages vary rapidly, or is it kind of random?