Ripple

Check if a diode in the bridge rectifier has failed.

Good idea! I must admit I'd overlooked that possibility. I should have checked the frequeny of the ripple whilst I was at it.

The shape of the ripple will tell you if that has happened or if one of the diodes haas gone open circuit. Another possibility is a fault in some other part of the circuit that is drawing too much current - you might smell something getting hot.

Usually there will either be a passive filter after the initial large capacitor, such as an inductor in series and a further large capacitor to earth, or an active voltage regulator circuit involving an amplifier and series semiconductor, or both. A simple diode plus capacitor is always going to have a huge amount of ripple, not hard to calculate if you know the current load. But no electronic instrument is going to be powered just by a rectifier capacitor combination. You need to look further into the circuit.

In addition to the other suggestions, you might check to see if a snubber capacitor exists across the diodes.

There's nothing getting hot, but the ripple had a definite saw-tooth shape to it, which could be indicative of abnormal operation. Shame there are no schematics available anywhere on the net; always makes troubleshooting much harder.

Is the rectifier circuit half-wave or full-wave? If the latter, is the ripple also full wave?

Some designers allowed a larger amount of ripple on the first capacitor to reduce the heating of the mains transformer (the copper losses are proportional to I squared, so by allowing a longer charging time per cycle at a lower current, they achieved lower overall resistance losses). As long as the lowest part of the ripple was sufficiently far above the stabilised output, a series stabiliser could cope with it.

Is the 'stabilised' line actually stable or is there ripple on it in the form of a series of little notches?

Marconi have used these things (not these exact components of course, but the same footprint/pinout:

So full wave. Is the ripple full wave? Well, I did mean to check the frequency of it, but other stuff got in the way. I'm going to have to re-investigate a bit more thoroughly later in the week.

Not sure what you mean by stabilised line. There are two secondary windings on the mains transformer which go into separate bridge rectifiiers of the type linked to above and then on downstream to the first tank caps. One line has a ton of ripple on it, the other very little indeed. Again, I'll need to go back and pay more attention to what's going on here. I've missed out too many important details.

Older practice, when active devices were expensive and bulky was to have an active stabiliser circuit (which also got rid of the ripple) directly after the power supply. Nowadays they tend to have voltage regulators for each module or circuit, local to that module. If you have the latter arrangement I would be looking for ripple on the output of each local regulator rather than the main power supply. Or the high ripple could be due to excessive current being drawn from one power supply. Anything getting excessively hot?

I was assuming there would have been some sort of stabiliser after the capacitor, so that the line downstream of it was reasonably stable and quiet. Alternatively they might have used a series choke (or resistor) and further capacitor to smoothout the ripple without actually stabilising the voltage

If this is made by Marconi, the circuits for a lot of Marconi instruments are available online but you may need to do a bit of searching-out. If you can get hold of a handbook, that should always have the circuit in it. What is the "TF" number of this equipment?

["TF" is Marconispeak for "Test Fixture" i.e. any sort of test gear.]

I don't think anything's getting unduly warm, let alone hot. But I can't be sure as access is impeded and they've used the most

*enormous* finned cast aluminium heatsink for only four TO-220 package regulators. This sig gen is the 10k to 5.4Ghz one and is about 25 years old I would guess. I agree with your approach on checking each regulator individually but the access issue is a real pain here. They clearly haven't given much thought to servicing with this model - and that's unusual for a company like Marconi/Aeroflex (or whatever they call themselves now). Plus the fact that they've never released schematics doesn't help much either.

Prior to the regulator circuit, the ripple voltage depends upon current drawn for a given supply at a given frequency. The higher the current drawn, the greater the ripple for your supply. So one possibility in your generator is that too much current is being drawn i.e. a defect in some circuit "downstream" of the bridge/cap.

Depending on what follows the diode/capacitor supply you may have a regulator failure. Regulators hold the voltage at some specific value, thus they remove ripple. Voltage measurements usually reveal which regulator is failing to provide the proper output voltage. Again a defect downstream may pull the voltage down. Without a manual/schematic its harder to diagnose - but it is still possible. And I read somewhere that manuals are available. See the url posted below - it may help.

Next - perhaps the ripple really isn't excessive. By that I mean it may be within spec for your generator. You may find this site helpful:

Good luck. Let us know whe you fix it! Ed

Many thanks for that, Ed. :)