Capacitors in Series - phase changes?

[GeorgeBoles]

Hello folks,

I am aware that putting capacitors in series produces an overall capacitance of 1/Ct = 1/Ca + 1/Cb + ... etc., and had never thought about it any more.

But what happens to the phase of a signal which passes through, say, three capacitors in series - is it changed by only 90 degrees (as it would be for a single capacitor, as if they were all working together as one), or is it retarded by 90 degrees three times over - (once for each capacitor)?

Your advice would be helpful if multi-amping my soon-to-be-borne Linkwitz Orions with different sorts of power amplifier, or if I fiddle with more or fewer capacitors in the signal chain, say in my Nelson Pass "NS10" type pre-amp... depending on what is in my spares box or what is cheap at my supplier.

All advice gratefully and gracefully received.

Regards,
George.

[GeorgeBoles]

Also, while I am at it, I read somewhere on this forum that it was unsafe to simply put two capacitors in series and expect them to handle twice their rated voltage, e.g. using two 35V rated capacitors in series in a power supply to cope with 50V rails, because the voltage may NOT be shared equally between them,

In this discussion, there was a diagram showing how to ensure that the voltage was split between the capacitors, I think by placing resistors in parallel with each of the capacitors and letting the resistors split the voltage.

Can anyone either point me to the thread, or point me in the correct direction.

Regards again,
George.

[Babowana]

Originally posted by GeorgeBoles


I am aware that putting capacitors in series produces an overall capacitance of 1/Ct = 1/Ca + 1/Cb + ... etc., and had never thought about it any more.

But what happens to the phase of a signal which passes through, say, three capacitors in series - is it changed by only 90 degrees (as it would be for a single capacitor, as if they were all working together as one), or is it retarded by 90 degrees three times over - (once for each capacitor)?

The phase shift is always a relative value.
For example, when we have a low pass RC filter (lag-network) of an amplifier, we could consider the phase shift.

Let's say, we are using three caps having the same value (Ca) in series, the total capacitance is calulated as Ca/3.
And, if we compare the phase shift between R-CaCaCa (three caps in series) filter with R-C (one cap of Ca/3) filter, the phase shift is maintained the same.
But, if we compare the phase shift between R-CaCaCa (three caps in series) filter and R-C (one cap of Ca) filter, the phase shift is different. In this case, the phase shift is increased in R-CaCaCa filter according to the number of caps in series. Each additional phase shift is less than 90deg and the increase rate is non-linear.

The same principle is applied to the high pass filter, but the phase shift is decreased when there is the phase shift.

Sorry I'm bad in writing . . . somewhat drunk . . .

[Babowana]

Originally posted by Babowana


The phase shift is always a relative value.
For example, when we have a low pass RC filter (lag-network) of an amplifier, we could consider the phase shift.

Let's say, we are using three caps having the same value (Ca) in series, the total capacitance is calulated as Ca/3.
And, if we compare the phase shift between R-CaCaCa (three caps in series) filter with R-C (one cap of Ca/3) filter, the phase shift is maintained the same.
But, if we compare the phase shift between R-CaCaCa (three caps in series) filter and R-C (one cap of Ca) filter, the phase shift is different. In this case, the phase shift is increased in R-CaCaCa filter according to the number of caps in series. Each additional phase shift is less than 90deg and the increase rate is non-linear.

The same principle is applied to the high pass filter, but the phase shift is decreased when there is the phase shift.

Sorry I'm bad in writing . . . somewhat drunk . . .

I was wrong. It should have been:

The phase shift is always a relative value.
For example, when we have a low pass RC filter (lag-network) of an amplifier, we could consider the phase shift.

Let's say, we are using three caps having the same value (Ca) in series, the total capacitance is calculated as Ca/3.
And, if we compare the phase shift between R-CaCaCa (three caps in series) filter with R-C (one cap of Ca/3) filter, the phase shift is maintained the same.
But, if we compare the phase shift between R-CaCaCa (three caps in series) filter and R-C (one cap of Ca) filter, the phase shift is different. In this case, the phase shift is decreased in R-CaCaCa filter according to the number of caps in series. Each phase shift is less than 90deg and the decreasing rate is non-linear.

The same principle is applied to the high pass filter, but the phase shift is increased when there is the phase shift.

[cliffforrest]

Consider a black box containing two 2uF capacitors in series.

The outside world can only see a single 1uF capacitor.

What parameters of capacitance could produce a difference in phase shift between the black box and a single 1uF cap?

None.

Yes, the DC voltage across each one will depend on the effective leakage resistances of the caps. Swamp this with a parallel resistor to define the voltage share.

[GeorgeBoles]

Many thanks to you both, Babowana and Cliff.

You have solved my problem.

Regards,
George.