Some musings on leaky capacitors, and capacitors with a high effective series resistance’ (ESR):
Confusion often seems to arise with regard to these quite separate, and arguably opposite, fault conditions in aging capacitors. Basically, leakage arises in non-electrolytic capacitors – typically paper wax covered ones, known as ‘waxies’.
The term ‘leakage’ is in the electrical sense – it does not refer to the soft often melted wax covering of waxy caps. What happens, is that instead of the capacitor only passing AC signals - as for example, the anode of an IF valve to the grid of the output valve – often known as ‘that’ capacitor, the capacitor acts as though it has a resistor in parallel, and thus, allows DC current to pass. This can damage the output valve and other components such as the output transformer, by causing excessive current to be drawn.
Not only will a leaky cap cause DC voltage to appear in places where you don’t want it, such as on the grid of a valve, it can also reduce the DC Voltage where we do want it - for example on the cathode of a valve.
There are some places where leakage is less critical, and where the capacitor isn’t subject to high voltage, but it must be stressed that the only effective way of checking if a cap is leaky, is at its working voltage. You can’t measure this parallel resistance in a passive test, for example, on an Ohm-meter.
It’s for this reason that many restorers - particularly on sets that aren’t rare, such as the Bush DAC90A, routinely replace all waxy caps, when it's likely in any event that that they've become leaky. They are after all, at least sixty years old. Another candidate for exchange would be in test gear such as a res/cap bridge, where the accuracy of the equipment may be affected or a signal generator where that too could happen, or the waveform degraded.
However, if the aim is to ‘conserve’ rather than to ‘restore’ equipment then clearly, rather than to yank out paper caps willy nilly, which – however old they are - may nevertheless be in good condition, some means of checking the leakage to see if it’s within acceptable limits at its working voltage must be devised. A google search on ‘capacitor leakage’ will provide more information and home-brew designs and leakage testers.
Effective Series Resistance (ESR):
This only applies to electrolytic capacitors. The effect is the opposite of leakage, in that rather than appearing as a parallel resistor, it appears as a series resistor. Again, as with leakage on non-electrolytic caps, this undesirable phenomenon cannot be checked with normal test gear such as a multi-meter or a capacitance meter which doesn’t have an ESR function. Indeed, a cap with a high ESR might seem to be perfectly healthy and of the right capacitance when checked on a normal capacitance meter.
The usual method of checking the ESR of an electrolytic capacitor is - not surprisingly – on an ESR meter. They almost all work on the same principle, by generating a small AC signal of a very low voltage in an oscillator at a frequency of 100kHz – 150kHz. The actual frequency is unimportant. If the capacitor under test has a high ESR it will be shown on the meter in Ohms. By ‘high ESR’, we mean anything from perhaps 5 Ohms upwards.
An Australian – Bob Parker, is the acknowledged expert on ESR and has designed several ESR meters in commercial production. I’ve previously posted on this forum a homebrew meter that I’ve built, many others of which have been successfully built, based on a design by a Canadian radio amateur, VE7IT. Commercial meters measure ESR to several decimal places, but an accurate reading isn’t necessary, so a homebrew meter is fine. Whether a cap has an ESR of about 10 Ohms on a homebrew meter, or 9.758 on a commercial digital meter, the ESR is far too high and the cap is a dead as Monty Python’s parrot, so needs a dignified burial.
Bob Parker has produced a helpful table of acceptable ‘worst case’ ESR levels for capacitors of different values, which can be found at this link:
http://www.your-book.co.uk/design/esrchart.htm
I should say right away, that ESR meters have little application in valve radios, most of which use only two electrolytics for smoothing, and one for the output valve cathode by-pass, usually 25V 25uF, and usually duff, so best to replace on sight. However, solid state radios TVs and other equipment such as ‘scopes, often use scores of electrolytics, often physically small, sometimes subjected to heat, which can and often do, develop a high ESR over time. It is often the cause of displays in VCRs and microwaves going dim. Radios such as the Roberts R200 and similar, use several electrolytics, as do Bush TR82Cs, and whilst a high ESR may not stop the set from working, it will degrade its performance.
These notes aren’t exhaustive, and the main point I wanted to clear any confusion and to get across is the essential difference between leakage in paper caps, and a high ESR in eletrolytic caps – two very different fault conditions.
Hope these notes are of use and interest.
David
Confusion often seems to arise with regard to these quite separate, and arguably opposite, fault conditions in aging capacitors. Basically, leakage arises in non-electrolytic capacitors – typically paper wax covered ones, known as ‘waxies’.
The term ‘leakage’ is in the electrical sense – it does not refer to the soft often melted wax covering of waxy caps. What happens, is that instead of the capacitor only passing AC signals - as for example, the anode of an IF valve to the grid of the output valve – often known as ‘that’ capacitor, the capacitor acts as though it has a resistor in parallel, and thus, allows DC current to pass. This can damage the output valve and other components such as the output transformer, by causing excessive current to be drawn.
Not only will a leaky cap cause DC voltage to appear in places where you don’t want it, such as on the grid of a valve, it can also reduce the DC Voltage where we do want it - for example on the cathode of a valve.
There are some places where leakage is less critical, and where the capacitor isn’t subject to high voltage, but it must be stressed that the only effective way of checking if a cap is leaky, is at its working voltage. You can’t measure this parallel resistance in a passive test, for example, on an Ohm-meter.
It’s for this reason that many restorers - particularly on sets that aren’t rare, such as the Bush DAC90A, routinely replace all waxy caps, when it's likely in any event that that they've become leaky. They are after all, at least sixty years old. Another candidate for exchange would be in test gear such as a res/cap bridge, where the accuracy of the equipment may be affected or a signal generator where that too could happen, or the waveform degraded.
However, if the aim is to ‘conserve’ rather than to ‘restore’ equipment then clearly, rather than to yank out paper caps willy nilly, which – however old they are - may nevertheless be in good condition, some means of checking the leakage to see if it’s within acceptable limits at its working voltage must be devised. A google search on ‘capacitor leakage’ will provide more information and home-brew designs and leakage testers.
Effective Series Resistance (ESR):
This only applies to electrolytic capacitors. The effect is the opposite of leakage, in that rather than appearing as a parallel resistor, it appears as a series resistor. Again, as with leakage on non-electrolytic caps, this undesirable phenomenon cannot be checked with normal test gear such as a multi-meter or a capacitance meter which doesn’t have an ESR function. Indeed, a cap with a high ESR might seem to be perfectly healthy and of the right capacitance when checked on a normal capacitance meter.
The usual method of checking the ESR of an electrolytic capacitor is - not surprisingly – on an ESR meter. They almost all work on the same principle, by generating a small AC signal of a very low voltage in an oscillator at a frequency of 100kHz – 150kHz. The actual frequency is unimportant. If the capacitor under test has a high ESR it will be shown on the meter in Ohms. By ‘high ESR’, we mean anything from perhaps 5 Ohms upwards.
An Australian – Bob Parker, is the acknowledged expert on ESR and has designed several ESR meters in commercial production. I’ve previously posted on this forum a homebrew meter that I’ve built, many others of which have been successfully built, based on a design by a Canadian radio amateur, VE7IT. Commercial meters measure ESR to several decimal places, but an accurate reading isn’t necessary, so a homebrew meter is fine. Whether a cap has an ESR of about 10 Ohms on a homebrew meter, or 9.758 on a commercial digital meter, the ESR is far too high and the cap is a dead as Monty Python’s parrot, so needs a dignified burial.
Bob Parker has produced a helpful table of acceptable ‘worst case’ ESR levels for capacitors of different values, which can be found at this link:
http://www.your-book.co.uk/design/esrchart.htm
I should say right away, that ESR meters have little application in valve radios, most of which use only two electrolytics for smoothing, and one for the output valve cathode by-pass, usually 25V 25uF, and usually duff, so best to replace on sight. However, solid state radios TVs and other equipment such as ‘scopes, often use scores of electrolytics, often physically small, sometimes subjected to heat, which can and often do, develop a high ESR over time. It is often the cause of displays in VCRs and microwaves going dim. Radios such as the Roberts R200 and similar, use several electrolytics, as do Bush TR82Cs, and whilst a high ESR may not stop the set from working, it will degrade its performance.
These notes aren’t exhaustive, and the main point I wanted to clear any confusion and to get across is the essential difference between leakage in paper caps, and a high ESR in eletrolytic caps – two very different fault conditions.
Hope these notes are of use and interest.
David
