Crossover Capacitor Degradation over time

[LoudnClear]

Bob/Dean/Erik,

Can you detail a bit the degradation modes in the caps as they age and drift out of tolerance?

Is there a change in capacitance? Increase or decrease? Magnitude? Rate of change?
DaddyDee mentioned a resistive change, What is the phenomenon? Magnitude? Rate of change?
Does the degradation occur more on certain types of caps?
Do SoniCaps/etc have features that slow or defeat these degradations?
Do SoniCaps/etc have features that make them more suited/uniform in audio applications?

Thanks,
Kimball

[BEC]

OK,  I will take a shot at this.

Capacitors used in electronic circuits are generally specified as having a given value of capacitance and a given working voltage.  The circuit always assumes a "perfect" capacitor, but no perfect capacitor exists.  That "perfect" capacitor would add nothing at all to the circuit except capacitance.  Anything else added to the circuit other than capacitance can be represented as a "loss" or something that makes the capacitor look like a capacitor plus a resistor.

Now, with the right meter, we can measure these losses and express them as an equivalent value of resistance that is in series with the capacitor.  This measurement, ESR (Equivalent Series Resistance) should always be as low as achievable, since it is, by definition, a bad thing.  That is because it is something the capacitor is doing other than adding capacitance.

When I said that no "perfect" capacitor exists, I can also say that good polypropylene capacitors come darned close to being "perfect".  They typically measure at around a few hundredths of an ohm for ESR.

The older the capacitor, the higher this measured ESR.  I guess that this increase is from perhaps a combination of corrosion on the internal connections and perhaps also, the purity of materials originally used in the manufacturing process.  

Anyway, I think I can hear the difference in the sound of a crossover when the ESR of the caps has reached about 0.5 ohms.  That is a rather typical value I measure on caps in a 20 year old crossover.  The more caps in series in the crossover, the more the effect is audible. I think the highest ESR I have measured in a cap from a Klipsch crossover is something over 300 ohms.  

I hope that starting out fresh with a capacitor with a very low value of ESR will mean that it takes longer for ESR to reach a value that is audible.  We need to revisit this in 20 years or so to see if the new caps with very low values of ESR hold up better.  At about 5 years of use on some caps I have, I still measure ESR to be almost exactly the same as they started out with.

By the way, capacitance may still measure good or in spec after ESR has risen to a huge value.  If fact it is normal for caps I have measured to still be good as far as capacitance is concerned and be total non-functional because of very high ESR. 

Bob Crites

[tigerwoodKhorns]

Really good post Bob.

[LoudnClear]

Bob,

Thanks for the great commentary, you managed to explain it in a way that I understood every last word you said. Perfect!

So it sounds like every self respecting [vintage] crossover guy needs an ESR meter. (vintage modifies crossover, not guy)

The Atlas ESR60 meter looks like a nice accurate little fellow. I'm going to ask around at work and see if anyone out in the labs uses an ESR meter.

I'm planning on ordering caps for my cornwalls and heresys, clipping one lead of the existing caps to open them from the circuit, while soldering the new cap into the circuit. I plan on initially running a bit of extra wire out the port on the cornwall so I can have my wife switch old cap/ new cap while I'm back listening for difference. Of course the extra run of wire would add some effective ESR-like resistance, but it would be common to both caps, so should wash out somewhat.

Anyway, thanks for taking the time to bring me up to speed on what ESR is, I really appreciate it.

Kimball

[BEC]

For anyone getting serious about measuring crossover components, I would recommend the B&K Precision Model 885. This meter does it all. Capacitance, Inductance, Resistance and ESR all at good accuracy. An added feature is that the meter directly reads Impedance. Of course if just doing ones own crossovers, sort of hard to justify the cost of around $550.00 with the Kelvin Clip 4 wire leads and gold contacts.

Bob Crites

[WopOnTour]

Great post Bob!

A quick related question? (You've probably answered this one a few times I spose)

Have you observed oil-filled motor run capacitors to have higher ESR than poly-props?

I'm using Solens right now, but happended to come across some new 2 and 12.5uFd motor runs at an industrial electrical shop and couldnt resist grabbing them ($5-9 each)

They are quite large and look VERY similar to the ones I removed from the AA networks when I overhauled them a few months back. I was contemplating just patching them in with a switch so I could do some A-B comparisons.

I know you USED to be a proponenet of the motor runs...

Do you expect I will notice a difference in the sound? (ie would YOU?) and if so can this difference be attributed to a combination of the different (higher?) ESR and the electrolytic properties of the materials used?

Many Thanks

***

[BEC]

The oil filled caps I used were GE and they were polypropylene.  They had an excellent ESR.  Since GE stopped making them, I haven't found any more of the oil filled caps that are as good.  So how good yours are will depend on who made them and what materials were used and how old they are.

Bob Crites

[two dot]

Bob, if you are reading this please shoot me a pm and tell me where to install my "midrange gaskets"???

Thanks

You have a pm

[BEC]

And your also have a PM reply.

Bob