Electrolytic capacitor values

[tin0gauge]

Hello, can someone help me on capacitor values please. How do designers arrive at the values of these ? originals can be 16mfd or 50 mfd or other values.
In valve circuits they tend to be lower than transistor circuits is that because there was a limitation on value and size at the higher voltages ?

When replacing these in decoupling circuits on valve radios how critical is the value ? for example is it better in a smoothing or decoupling circuit to always go higher value with a replacement ?

In a typical mains soothing circuit there is a electrolytic then a resistor and then another electrolytic how do these work together ?

Do these old electrolytics tend to lose value or go O/C in which case one can just use the original post as a suitable anchor for a replacement axial capacitor or do they leak ?

[ms660]

Designers arrive at the values that will do the job and be cost effective.

In the old days there was a limit in capacitance values possible at that time.

Reservoir capacitors in solid state stuff can be a lot higher in value as there is a lot less power dissipated in modern solid state rectifiers, with valve rectifiers there is in effect more dissipation due to the rectifiers internal resistance, that limits the value of a reservoir capacitor that can be used, manufactures valve data sheets usually specify the maximum value allowable for the reservoir capacitor. The value of the reservoir capacitor is quite critical but not absolute, eg: 33uF will do for 32uF, 47uF will do for 50uF etc.

In a capacitor, resistor, capacitor HT circuit the filter capacitor just finishes off what the reservoir couldn't do because of the reservoirs limited value as dictated by the valve rectifier specs.

Yes they can loose value or go O/C or leak, or develop a high effective series resistance. It's not a good idea to use the tags of a duff capacitor to connect a replacement one to, because, as you say, the original can leak.

Lawrence.

[threeseven]

I find myself in disagreement with some of the forgoing.
Capacitor values are arrived at by calculation, using circuit requirements as factors. EG, a given capacitor may be required to pass a certain frequency range and attenuate other frequencies, at the same time blocking a certain operating voltage. Relevant calculations will determine the value required.
As for capacitance value, in general, valve circuits are high voltage and resistance, so required capacitance value is lower but at a higher operating voltage. Transistor circuits are low voltage and resistance, so capacitance and operating voltage is lower. For reservoir caps, this means the load is higher current, so much higher capacitance is required.
This is a very general overview of a very complex subject.
As for how critical values are, it depends on the positon in the circuit. Reservoir caps are not that critical within say 30- 50% of original value. Feedback, coupling and bypass are a little more critcal. Bare in mind though, electrolytics have a very wide tolerance on value anyway.

[GeoffK]

Valve radios are, with a few exceptions, high voltage high impendence voltage operated circuits and the capacitive reactance of the capacitors usually have smaller capacitive values along with higher voltage ratings that determine the physical size of the capacitor. Transistor circuitry is lower impendence current operated with lower capacitive reactances and having larger capacitive values for the lower reactance but having lower voltages can still be made smaller sizes for larger capacitive values. Better manufacturing methods has made capacitors smaller for a given size and voltage.
The electrolytic capacitors in older equipment can be very reliable, it is lack of use that eventually leads to them going faulty by being unable to reform. The capacitors in the power supply are the reservoir, then the resistor and then the smoothing that smooths out the ripple voltage even further, older radios sometimes have an inductor instead of the resistor to further smooth the DC voltage by inductive reactance. In a small transistor radio the reservoir could be as high as 200uF and the smoothing capacitor 100uF, compared to much smaller capacitive values in a valve radio.
Any leaking capacitor should be removed completely, not soldered on the old tags, though old capacitor cases can be re-stuffed with new capacitors.
May have written something similar to above post.

[ms660]

"I find myself in disagreement with some of the forgoing.
Capacitor values are arrived at by calculation, using circuit requirements as factors. EG, a given capacitor may be required to pass a certain frequency range and attenuate other frequencies, at the same time blocking a certain operating voltage. Relevant calculations will determine the value required."

Equals

"Designers arrive at the values that will do the job and be cost effective."

Lawrence.

[DigitalNoMore]

Being a retired telecommunications equipment designer and used to designing for reliability as well as function and cost I sometimes despair at the inadequate voltage ratings of some of the electrolytics designed in.
This equally applies to more modern equipment where the designer can't use lack of availability as an excuse. Speaker coupling capacitors seem to be a case in point. I've replaced a number of these in relatively modern transistor radios and in each case the voltage rating has been too low (okay for normal operation but not high enough to cope with the voltage surge on switch-on).

[threeseven]

Occasionally that was an issue with some Philips stuff back in the 70's. Bench service tech's would ignore company policy and use a higher rated components where a common fault became apparent.

[DigitalNoMore]

As well as value out of spec, open-circuit or leaky, electrolytics can also fail with a dead short.

Another good reason for not connecting replacement new capacitors across the tags of old ones unless you've removed the guts first.

[Biggles]

Correct me if I am wrong, but to hold the same amount of charge, if the voltage is lower, the capacitor has to be of greater value. The amount of charge required to be held is the important thing here. Hence the greater values normally found in transistor based lower voltage equipment, particularly on supply rails. There are of course other limiting factors to be considered, frequency response, maximum ratings of rectifiers, ripple current, availability and cost.
Alan.

[G8HQP Dave]

'Charge stored' (=CV) is not usually the important thing. 'Energy stored' (= 0.5CV^2)may be important for a reservoir capacitor. For most other caps it is impedance which matters.

[GeoffK]

The charge does depend on the voltage applied to a capacitor Q = C*V
Q = Coulombs, the charge stored on the capacitor plates. Transistor circuits have lower impendence circuits, the frequencies at the lower impendence need a larger value capacitor for the reactance to be large enough to accept the wanted frequencies. High impendence valve circuits need smaller value capacitors with a less reactance values to accept the same frequencies.

[ms660]

Originally Posted by Hetrodyne

the frequencies at the lower impendence need a larger value capacitor for the reactance to be large enough to accept the wanted frequencies. High impendence valve circuits need smaller value capacitors with a less reactance values to accept the same frequencies.

As read, I don't quite follow that.

Lawrence.

[GeoffK]

A circuit with a higher impendence, an input to a valve control grid for example because it is a voltage on the grid to influence the flow of electrons in the valve and not a current being passed, needs a smaller capacitive value to pass a particular frequency via the grid coupling capacitor due to the high impedance. The same frequency at a lower impedance such as the base input to a transistor that is current driven via the base current needs a larger capacitor value to pass the same frequency. Of course the capacitor doesn't actual pass the frequency it flows round the external circuit to charge and discharge the capacitor. At least that is my understanding of why different capacitor values are chosen so the reaction of the capacitor is large enough for the required frequency to 'pass'. In general capacitors in transistor circuits have larger values than in valve circuits because of the reactances at lower circuit impendences. If I have it wrong then I would like to know as I thought this was the main reason.

[ms660]

Originally Posted by Hetrodyne

the frequencies at the lower impendence need a larger value capacitor for the reactance to be large enough to accept the wanted frequencies. High impendence valve circuits need smaller value capacitors with a less reactance values to accept the same frequencies.

To me, what you imply in the above is that a larger value of capacitance will result in a larger value of reactance and that for a smaller value of capacitance the reactance value will be less, that's opposite to what I have always understood.

Lawrence.

[Station X]

As you say Lawrence the reactance of a capacitor is given by 1/2*PI*F*C, so the higher the capacitance value the less the reactance and vice versa.

Perhaps Hetrodyne is considering the circuit as a whole and not just the capacitor?

[G4_Pete]

On the subject of ripple current I have noticed that some maunfactures quote 120 Hz as the normalised value and then apply frequency modifiers and some quote 100 Khz as the normalised value. This means if the catalogue short listing does not specify frequency you have to check the data sheet to get the true picture.
So in choosing a capacitor ripple loss and related disipation losses have to be considered which is related to the physical size and how much heat can be got rid of if the thing is to last just beyond the guarentee!!

So in choosing modern replacements for old radios this is a point to note.

Sample derating factors from a random data sheet.
Frequency
Coefficient
50Hz 0.40
120Hz 0.50
1kHz 0.80
10kHz 0.90
100kHz or more 1.00

[GeoffK]

May not have got it quite right about the reactance for larger or smaller value capacitors at different frequencies and impendences. The reactance of a capacitor becomes larger as the frequency becomes lower, the reaction should be low enough, not large enough, at particular wanted frequencies. Transistor circuits tend to be lower impendence circuits than valve circuits so capacitor values for doing the same task can be different values, usually larger, in transistor circuits compared to valve circuits.

[Herald1360]

Can we please talk about impedance not impeNdence?

It's difficult to take seriously any discussion where even the basic terminology is confused.

Reactance is the capacitive or inductive component of an impedance.

Impedance is the combined effect of the resistance and reactance of a circuit element.

Strictly speaking any real world component has an impedance, though depending on the frequency considered the reactive or resistive parts may be trivial in its effect on circuit function.

Going back to the original question- it does seem to have been answered (if in a piecemeal fashion).

[GeoffK]

Apologies for spelling, was in a hurry, but no excuse for poor spelling, of course I meant impedance.

[Biggles]

The relative impedance of the capacitor at particular frequencies to the rest of the circuit can be looked upon as a potential divider circuit. A capacitor of a certain impedance when presented with a relatively low impedance circuit such as those typically found in transistor circuits will suffer more of a loss due to the potential divider effect. A high impedance capacitor presented with a high impedance load will suffer less of a "pull down" than if it was presented with a low impedance load.
Alan.