Replacing smoothing capacitors

[PaulR]

Sometimes when replacing smoothing capacitors where there were originally two in a can it is more convenient to connect the replacements somewhere else under the chassis, say to a valve base. This means that the earth lead of the replacement may not be connected to the point where the original was connected and may mean that two separate points are used when the original can used one.

Is it likely or possible that this may introduce hum problems due to the currents flowing in the chassis or are they too small for this?

[Nuvistor]

I think it’s a case of try it and check the results, yes it’s possible to introduce problems but each model will react differently.
I have always tried to replace any components as near to original location as possible.

[boxdoctor]

In conventional (consumer) valve circuitry, the circuit impedances and the signal levels are comparitively high as compared with any hum currents likely to be caused by earthing a smoothing cap at an alternative point on the chassis, so I would think it is not a problem providing the earth is electrically sound.
The same may not be true on solid-state apparatus , however, especially if inductive components are in the vicinity of the modified current paths caused by the revised layout. Tony.

[PaulR]

Thanks for the replies.

My question was brought about because of a Pye Music Maid clock radio which I recently brought back into use after several months. It worked for a while then started to hum badly because one of the original smoothing caps failed. I replaced them both in different positions to the originals and there is a bit of a hum remaining and I was not sure whether it was always there. It does have a UL41 output valve but I replaced that not long before another fault caused me to stop using it. The hum isn't too bad and I now think it must have been there all along but it made me think about putting replacement smoothing caps in a different place

[G8HQP Dave]

The critical path is the one from the transformer/rectifier arrangements to the reservoir capacitor negative, because this carries charging pulses. Ideally this path should not include the chassis, but in days of old people expected some hum and worried if it was not present.

[PaulR]

So residual hum might be improved from the original performance by connecting the negative directly to the transformer?

[turretslug]

It's worth connecting the new reservoir negative directly to the transformer secondary HT centre-tap (assuming a bi-phase, full-wave rectifier circuit) if it is a type with insulating sleeve as this completes the charging pulse circuit without sending ripple current through the chassis. The traditional layout would have a short wire link between centre-tap and chassis with the reservoir can clamped to chassis possibly quite a few inches away, thus ripple current flows through the chassis and can modulate the signal circuits with hum. The smoother capacitor negative can generally be connected to the chassis, as its ripple current will be of much lower amplitude than that of the reservoir. In the case of a bridge rectifier, the reservoir negative would go straight onto the bridge's negative tag, again with the aim of completing the ripple current circuit without a chassis path.

Of course, there are 101 things that can cause loudspeaker hum, but this is one of the quickest, cheapest and easiest to get right from the outset- it's puzzling that so many manufacturers omitted this precaution for so long.

[ms660]

It's also worth checking as to whether or not the -ve(s) of the reservoir/filter capacitor(s) should be at chassis potential, they should be in the OP's receiver but in many receiver designs they're not, due to valve biasing arrangements.

Lawrence.

[PaulR]

Thank you for the thoughts. This was a general question prompted by replacing the two caps in the Philips which is actually an AC/DC radio. The hum is not bad at all but is discernible at the low volumes I use as a bed side radio. It has an iffy valve holder and so will probably need looking at again before long and I might experiment with the negative connection of the reservoir capacitor then.

I usually find that the smoothing caps on 50s and later sets reform quite well. I connect them to a variable HT power supply with an ammeter in line and slowly crank up the voltage over several hours as the current decreases. This is the first time I have had one fail after several year's use and it made me think about the negative lead connections as the most convenient places for the replacements are not where the original was. I used to restuff electrolytics for this reason but cannot really be bothered now.

Something to think about for future restorations - if I can ever squeeze another set into the house past my wife!!

[G8HQP Dave]

On an AC/DC radio I guess the reservoir cap negative should go to the incoming neutral. Then connect the cap negative to the rest of the circuit, probably via the chassis. This should minimise hum.

However, you need to be careful: as soon as you modify safety-related things like mains wiring you lose the 'grandfather rights' which allow 1950s/60s AC/DC sets to be bought and sold today even though they come nowhere near modern safety standards.

[paulsherwin]

Don't forget that 1950s domestic radios do have some hum even when working normally. This may be more noticeable with a bedside radio at a low volume in the dead of night.

[Nuvistor]

Many AC/DC radio’s fed the output valve from the reservoir capacitor. To help reduce hum the primary winding on the output transformer was tapped to try and “ buck” the hum. Replacing with a standard transformer would increase the hum.

[ms660]

I notice that Philips receiver uses a 0.0033uF for the AF output grid coupler and it's also fitted with a hum cancelling output transformer which also has a negative feed back winding, the output valve's cathode bias resistor has no bypass capacitor.

Lawrence.

[kalee20]

Originally Posted by Nuvistor

Many AC/DC radio’s fed the output valve from the reservoir capacitor. To help reduce hum the primary winding on the output transformer was tapped to try and “ buck” the hum. Replacing with a standard transformer would increase the hum.

I don't think it would!

The tapped primary winding on the OPT is there to neutralise hum, as seen ACROSS the OPT. It works as a sort of bridge arrangement.

If there is slightly less ripple on the HT reservoir capacitor, then the input to the bridge is slightly less - and (assuming imperfect unbalance in the bridge) then the neutralised output is slightly less too. If you got lucky and the bridge was balanced to begin with, then there'd be zero hum for all HT ripple amplitudes.

[ms660]

An article here:

https://www.radiomuseum.org/forum/gr...?language_id=4

Below is a scan of another reference from one of my books.

Lawrence.

[ms660]

The tapped primary transformer reduces the hum developed in the secondary so far as I can make out, no point in using one otherwise.

Lawrence.

[Nuvistor]

Originally Posted by kalee20

I don't think it would!

The tapped primary winding on the OPT is there to neutralise hum, as seen ACROSS the OPT. It works as a sort of bridge arrangement.

Hi Kaylee,
I am not sure what you mean, are you saying a standard transformer would make no difference or is it my explanation that is wrong but it does reduce hum?