PSU Ripple Update

Isn't that what you'd expect? The storage capacitor gets topped up twice per mains period. In between top-ups, it's the sole source of the output current, so its voltage drops until the next top-up.

You could measure the rate of voltage drop and check if it has the expected slope for the output current and the storage cap's value: dV/dt = -I/C.

You could also check if the pass transistors still have enough voltage across them at maximum current and just before the next top-up. Do you see ripple on the regulated output?

Jeroen Belleman

Yes, it's a fault alright. The ripple is superimposed on the output of the RF generator; it's impossible to ignore.

Not only am I seeing ripple on the regulated output, I'm seeing it on the final output of the signal generator this PSU powers. It's quite a beefy PSU, but as little as 90mA draw gives rise to an unacceptable level of ripple which permeates through the whole of the downstream circuitry.

You sure you got your probe division/scope division configured congruently?

Yes, it's on 1x @ 5ms. The only thing I can think of is some sort of issue with a rogue feedback signal which has ripple on it and is causing the PSU to (mis) regulate accordingly. But I have no schematic and there's none available online either, so I can't check for that.

OK, we need to know more. Your scope picture shows about

Next, what is the DC level on the input to the regulator, and what is the DC level on the output of the regulator? The ripple should be eliminated by the regulator if it's working properly. Can you descri9be the regulator circuit?

Ed

I'm unable to answer very much until the next time I have an hour to spare to look at the issue again (which will probably be next Sunday). In the mean time, I'll be noting any questions such as yours and will answer them all in a single post after the next inspection. It's a pity I don't have more free time available for this, but that's just the way it is at present. :-/

The question, which appears to have flown straight over your head, was: Does the scope setting match the probe you are using?

I.e., is it a 1X probe (or a switchable probe on the 1X setting)?

Is the probe 1X, 10X, or some other X?

Possible open ground connection of a 3-terminal or shunt regulator, especially if the output voltage is high,

Very light loads? all I can think of is open circuit input filter capacitors.

Did you replace the rectifiers, until something (anything) changed?

The ripple has changed since your last photo, as have your test conditions. You still don't indicate a 0V reference, so we can't tell what the % ripple IS.

This waveform shows equal phase peaks at the expected frequency.

What is your problem?

RL frequency.

Were you measuring across the filter cap terminals? No other scope ground connection?

RL

That sounds as if the linear regulators aren't regulating.

You need to compare the voltages applied to the inputs of the regulators, which can be expected to show appreciable ripple with the nominally regulated output, which shouldn't. Any residual output ripple should be a couple of orders of magnitude smaller than the input ripple.

It may be "beefy" but it doesn't seem to be working as it should.

Always good to thoroughly consider user error.

I wonder if Cursitor Doom has heard of "MF Doom"? "Do you see the perpetrator? Yeah, I'm right here.."

I fell into the same old trap as last time and the time before that and the time before that.... It was nothing to do with the PSU. I eventually tracked it down to a coax's shield in the RF section which had come adrift. When re-grounded, the ripple on the output completely vanished. Must have been somehow picking it up from the mains transformer despite all the screening and compartmentalisation in this device. All that time I wasted on the PSU - just because ripple *has* to be a PSU problem, doesn't it. Until it isn't, that is.

If the shield was open, then the return current for the RF would probably have been forced to flow back to the source via the power-supply ground paths (back to the power supply, and then through the ground wiring to the source circuitry). Any impedances along this ground path would have created some voltage drop, and variations in the power supply currents (e.g. ripple) would have created a varying voltage which would have been impressed on the output.

Good to hear you've tracked down the problem.

Actually measuring the output ripple of a functioning linear can be an issue, as you're often expected to measure microvolts on a DC level repeatedly - requiring carefull decoupling, probe technique and sometimes external amplification, if the spec is typically silly.

Suddenly you will become aware of all of the unshielded sources of low frequency interference in your work area.

RL

Indeed. Anyway, if they hadn't made it such a bitch of an area to inspect, I'd have spotted the issue at the outset.