08-04-2011, 10:52 PM
For non-electrolytic caps, the issue is that of leakage - for electrolytic capacitors, it's generally the opposite - high effective series resistance (ESR), which tends not be be much of a problem on valve radios as apart from the smoothing and reservoir caps in the power section, mostly, the only other electrolytic will be a 25uF 25V or thereabouts cathode bypass cap in the output stage. Im my experience, those are invariably shot, and should be changed on sight. Smoothing & reservoir caps which exhibit a high leakage and won't reform will usually cause hum (not the only cause of hum of course), and will need replacing, but if they exhibit low leakage, may still have a high ESR.
Electrolytics will often test fine on a capacitance meter showing the expected capacitance, and may have low leakage, but will nevertheless be useless if they have a high effective series resistance (ESR). They must be tested under similar conditions under which they have to perform in circuit by applying a small ac signal - as often as not about 100kHz, and the effective series resistance of the cap under those conditions, checked on a meter. If the ESR is much above 3 Ohms or so - much less if switched mode power supply caps, the cap will be highly suspect, and should at the very least be compared to the ESR of a new one (by 'new' I don't mean 'NOS' - new/old stock).
ESR is much more of an issue in solid state equipment which often uses scores of small electrolytics which dry out over time, and start behaving like series resistors. (This is often the cause of dim displays in such things as VCRs and microwaves).
Last April I posted a thread in another forum on a simple homebrew ESR meter which I'd built, since which time it has proved far more popular than I could ever have imagined - I know that scores have successfully been built since then, and have helped several constructors to troubleshoot their meter when they've built it and found it won't work.
I claim no credit for the original design, which was by a Canadian radio amateur, VE17T, but over time - based on feedback from those who've built the design - I've refined the PCB layout so that the only off-board components now are an LED (optional), a switch, a meter, and a PP3 battery. It uses a 50uA meter movement, which can be simply calibrated using half a dozen low value resistors.
If forum members would find it helpful, I'll happily post the details of this project on this forum, including troubleshooting details in case of difficulty. I know that some people buy fancy (hence, expensive!) ESR meters with digital readouts to several decimal points, but to what purpose, I've no idea - the precise ESR is on no importance. Whether or not a cap has an ESR of about 7(ish) Ohms on a homebrew meter, or 6.975 Ohms on a commercial meter that cost £100 or more, the outcome is the same - the cap is duff and needs binning. It's often a truism that 'the only difference between men and boys is the price they pay for their toys'!
There are many other designs of homebrew ESR meters to be found on the net - almost invariably consisting of an oscillator running at about 100kHz, so why did I choose the one that I made? Simplicity and economy! The oscillator is as simple as an oscillator could ever be - one cap, one resistor, and one gate of a 74HC14N Schmitt trigger IC (about 20p) generating a square wave of about 100kHz (actual frequency not critical). The other five gates are used as a buffer/low pass filter, which feeds the signal to a one transistor amplifier, then a few diodes as a mater rectifier.
Apart from the meter movement (about £7.00) the rest of the components cost about £3.00 or so, so apart from a box in which to house it, it comes in at a tenner. (It can be built on stripboard if PCB making facilities aren't available).
Hope that's of interest.
David
Electrolytics will often test fine on a capacitance meter showing the expected capacitance, and may have low leakage, but will nevertheless be useless if they have a high effective series resistance (ESR). They must be tested under similar conditions under which they have to perform in circuit by applying a small ac signal - as often as not about 100kHz, and the effective series resistance of the cap under those conditions, checked on a meter. If the ESR is much above 3 Ohms or so - much less if switched mode power supply caps, the cap will be highly suspect, and should at the very least be compared to the ESR of a new one (by 'new' I don't mean 'NOS' - new/old stock).
ESR is much more of an issue in solid state equipment which often uses scores of small electrolytics which dry out over time, and start behaving like series resistors. (This is often the cause of dim displays in such things as VCRs and microwaves).
Last April I posted a thread in another forum on a simple homebrew ESR meter which I'd built, since which time it has proved far more popular than I could ever have imagined - I know that scores have successfully been built since then, and have helped several constructors to troubleshoot their meter when they've built it and found it won't work.
I claim no credit for the original design, which was by a Canadian radio amateur, VE17T, but over time - based on feedback from those who've built the design - I've refined the PCB layout so that the only off-board components now are an LED (optional), a switch, a meter, and a PP3 battery. It uses a 50uA meter movement, which can be simply calibrated using half a dozen low value resistors.
If forum members would find it helpful, I'll happily post the details of this project on this forum, including troubleshooting details in case of difficulty. I know that some people buy fancy (hence, expensive!) ESR meters with digital readouts to several decimal points, but to what purpose, I've no idea - the precise ESR is on no importance. Whether or not a cap has an ESR of about 7(ish) Ohms on a homebrew meter, or 6.975 Ohms on a commercial meter that cost £100 or more, the outcome is the same - the cap is duff and needs binning. It's often a truism that 'the only difference between men and boys is the price they pay for their toys'!
There are many other designs of homebrew ESR meters to be found on the net - almost invariably consisting of an oscillator running at about 100kHz, so why did I choose the one that I made? Simplicity and economy! The oscillator is as simple as an oscillator could ever be - one cap, one resistor, and one gate of a 74HC14N Schmitt trigger IC (about 20p) generating a square wave of about 100kHz (actual frequency not critical). The other five gates are used as a buffer/low pass filter, which feeds the signal to a one transistor amplifier, then a few diodes as a mater rectifier.
Apart from the meter movement (about £7.00) the rest of the components cost about £3.00 or so, so apart from a box in which to house it, it comes in at a tenner. (It can be built on stripboard if PCB making facilities aren't available).
Hope that's of interest.
David
