Amazing! 100 Hz ripple and it still works!

[tom66]

So I got this PC from someone, found it in the back of an office.

Plug it in, boots up fine and starts XP. A little slow, but running. Oddly, the display has this kind of ripple look to it, like the sync signals are noisy.

I'm also hearing this rather loud buzzing from the PSU (fan okay), so I check the output voltages. WTH? 5V measured 4.4V, and 12V measured just 10V. Amazingly, it still ran fine, with no noticeable problems. Except that it didn't work with the first keyboard I tried (erratic key presses and flickering LEDs.)

So out of curiosity I scoped the rails and what did I find? A lot of ripple. Okay, I expected that, but the ripple frequency is what baffled me... it wasn't several kilohertz, it was about 100 Hz. (Twice mains frequency in the UK.) And it's still running fine... amazing.

Another strange thing I noticed was that the optical drive didn't work (the power LED flickered occasionally) but the HDD seemed fine.

I expect the two primary caps are gone (or going) and that it will quickly kill the motherboard caps.

The motherboard is reasonably old - just a 1.6 GHz Athlon with 1GB of ram.

And the PSU looks like another gutless wonder, no EMI filter of course so we can expect a lot of other components to be missing.

[Th3_uN1Qu3]

Re: Amazing! 100 Hz ripple and it still works!

That's not really a surprise, motherboard VRMs are quite good at eating low frequency ripple. Remember that the CPU load is continuously varying and the VRM has to keep core voltage constant in those conditions. A volt or so at 100Hz is nothing to the VRM... the caps won't be too happy tho.

Yes it does sound like bad primary caps in the PSU.

[tom66]

Re: Amazing! 100 Hz ripple and it still works!

That's not really a surprise, motherboard VRMs are quite good at eating low frequency ripple. Remember that the CPU load is continuously varying and the VRM has to keep core voltage constant in those conditions. A volt or so at 100Hz is nothing to the VRM... the caps won't be too happy tho.

Yes it does sound like bad primary caps in the PSU.

I went into the system health menu in the CMOS setup and the voltages were all over the place. The core voltage normally ran at 1.63V but occasionally spiked to 1.8V, but I'd expect the CPU to crash with that so more likely I think the voltage reference is unstable for the ADC... but still I don't want to run it like that for a long time. About the only thing that was okay was the 5Vsb at 5.1V.

One thing I forgot to mention was the ripple amplitude. It wasn't just 10mV, or 100mV. It was nearly 1Vp-p for 12V and 600mVp-p for 5V! 3.3V was relatively stable, that seemed odd but maybe it's because of the magamp circuit?

[b700029]

Re: Amazing! 100 Hz ripple and it still works!

I went into the system health menu in the CMOS setup and the voltages were all over the place. The core voltage normally ran at 1.63V but occasionally spiked to 1.8V, but I'd expect the CPU to crash with that so more likely I think the voltage reference is unstable for the ADC... but still I don't want to run it like that for a long time. About the only thing that was okay was the 5Vsb at 5.1V.

One thing I forgot to mention was the ripple amplitude. It wasn't just 10mV, or 100mV. It was nearly 1Vp-p for 12V and 600mVp-p for 5V! 3.3V was relatively stable, that seemed odd but maybe it's because of the magamp circuit?

This sounds like the very old Deer design that uses a completely unregulated self-oscillator and 7805 for the 5Vsb, and another linear regulator to take the 5V down to 3.3V.

[tom66]

Re: Amazing! 100 Hz ripple and it still works!

This sounds like the very old Deer design that uses a completely unregulated self-oscillator and 7805 for the 5Vsb, and another linear regulator to take the 5V down to 3.3V.

No, I had a look it's got a regulated two-transistor 5Vsb. One bad cap on the 5Vsb and one on the 3.3V. No sign of bulging primaries, but maybe they've just dried out.

[Newbie2]

Re: Amazing! 100 Hz ripple and it still works!

You should post pictures of the internals of this unstable power supply.

[tom66]

Re: Amazing! 100 Hz ripple and it still works!

You should post pictures of the internals of this unstable power supply.

I think it's just a standard L&C (350 marketing watts). Gutless with diodes on a bracket and four diodes instead of bridge, no EMI filter and only 330u 200V primaries. Barely anything inside it, loads of spaces for inductors and extra caps left shorted/open respectively. Will probably go to the trash soon as it's not really worth repairing. I will post a picture tomorrow.

[randtek]

Re: Amazing! 100 Hz ripple and it still works!

The 100 HZ ripple is due to the fact that a Full Wave Rectifier puts out a ripple frequency twice that of the A/C input frequency.

Full Wave Rectifier Tutorial

[tom66]

Re: Amazing! 100 Hz ripple and it still works!

The 100 HZ ripple is due to the fact that a Full Wave Rectifier puts out a ripple frequency twice that of the A/C input frequency.

Full Wave Rectifier Tutorial

Yes, I realised this when I saw the trace, I thought I was reading the time base wrong but it was definitely 100 Hz.