Capacitors across rectifiers

[spuffock]

In valve amplifiers, sometimes there are capacitors across the HT bridge rectifier diodes, and sometimes not. Removing them seems to have no effect. I doubt that they have anything to do with transient suppression since the diodes take ages to turn off anyway. All they seem to confer on the amplifier is a propensity to burst into flame. Can anyone think of a good reason to replace them?

 

[chemelec]

I assume you mean a capacitor and bridge rectifier on the tube filaments?
(Valve Amplifiers = Tube Amplifiers)

Depending on the amplifier, the can reduce Hum.

Most tube amplifiers just use AC on the filaments. No rectifier.

Some also use a balanced AC. A Potentiometer, Center-tapped to Ground to help balance ou hum.

 

[spuffock]

No, I mean the anode(plate) supply. 4 capacitors, one across each diode in the bridge. Values can be up to 0.1uf. Often with big black holes in and a nasty stink of burnt.

 

[Nigel Goodwin]

The capacitors are simply to help protect the diodes, they absorb short term high voltage spikes, which could otherwise cause the diodes to fail.

Personally I've never fitted them in anything I've ever built, and I've never had a rectifier fail - but I generally use over specified diodes. It could also be though that the mains supply in the UK is of such a high standard?, and that prevents any problems?.

 

[JimB]

I have seen circuits with capacitors across the diodes, particularly where several diodes are wired in series to increase the PIV rating. Comments with the circuits indicate that they are there to "suppress transients".

I have no experience to comment whether capacitors are a good thing or a bad thing, but my gut feel tells me that the idea is a "left over" from the earlier days of silicon diodes, when they were not as robust as they are now.

If I were building such a circuit I would just put 470k resistors across the diodes to share the reverse voltage evenly between the diodes.

If the capacitors are giving you grief, I think that you can safely remove them.

JimB

 

[spuffock]

It seems,then, that they are there for "historical" reasons, and can be discarded as being more trouble than they are worth. Thanks all.

 

[Styx]

caps on the rectifier had their uses.

1) turn-off/turn-on voltage sharing for series-connected devices (to increase blocking voltage)

2) snubber cct to try to move the losses out of the diode (you need an R in there as well)

3) spike suppresion


There was a time when these three effects were major problems, now they are relegated to high-voltage/high-power environement's.
They were mostly needed to get over problems with early rectifier units.

These days not an issue

 

[Dean Huster]

I'd say that the application is hardly for historical reasons only. Amateur radio linear amplifiers often used bridge rectifiers having three and four diodes in each leg. Kilowatt linear amps might have had supplies pushing 1500 volts and 2 amps. Although diodes may have been available to handle the PIV as a single unit, they may not have been able to handle the current. In any case, when diodes are in series and reverse-biased, their resistance is very high, but not necessarily equal. If one diode is exceptionally high in reverse resistance, it can take the brunt of all of the reverse voltage being applied, whether as the normal reverse voltage or as a transient spike. High-value resistors across the individual diodes equalized the normal reverse voltage across all the diodes while ceramic caps (10nF or so) across each diode took care of the faster transients.

Dean

 

[Styx]

I'd say that the application is hardly for historical reasons only. Amateur radio linear amplifiers often used bridge rectifiers having three and four diodes in each leg. Kilowatt linear amps might have had supplies pushing 1500 volts and 2 amps. Although diodes may have been available to handle the PIV as a single unit, they may not have been able to handle the current. In any case, when diodes are in series and reverse-biased, their resistance is very high, but not necessarily equal. If one diode is exceptionally high in reverse resistance, it can take the brunt of all of the reverse voltage being applied, whether as the normal reverse voltage or as a transient spike. High-value resistors across the individual diodes equalized the normal reverse voltage across all the diodes while ceramic caps (10nF or so) across each diode took care of the faster transients.

Dean

if you note what I wrote, I said such use of caps (for voltage ballancing) has been relegated to high-voltage and high power.

 

[spuffock]

I can see the reasoning in the case of chains of diodes in series, for the purpose of sharing the pain equally, but not in the case of single diodes.
Maybe it does protect the diode from transients, but it's likely to go short itself, then it takes the opposite diode out, then the fuse if you're lucky.
Since the diodes are cheaper than the capacitors, and the wreckage much the same, I think I'll stick with no caps on single diodes.