Capacitor Voltage Question

[tdstotler]

Quick Newbie question for the wisemen...


When building a EQ circuit, does the voltage of the Cap effect the audio? For instance if I use a 50V cap versus a 400 V Cap,
will there be any noticable differences in sound?
I just prototyped a new eq with 400V caps that I had available and it sounds amazing, just curious though if the higher voltage rating would
make a difference, and should I go lower.

Thank you!

Todd

 

[JohnRoberts]

Capacitors have a voltage coefficient that quantifies how much their capacitance changes with terminal voltage. When used in a filter this changing capacitance causes distortion and is the difference between using cheap ceramic and better film caps in audio filters.

For blocking caps, there generally isn't much changing voltage across the capacitor so voltage coefficient is less important, but other metrics can interact with current being passed.

In general the significance of breakdown voltage is when you don't have enough.

JR

 

[gridcurrent]

High voltage electrolytic capacitors have MUCH higher leakage current than an equivalent value in a lower voltage rating.  This is one reason why many capacitor "upgrades" aren't, as users often install higher voltage parts.  Do a search for voltage verses leakage current.

Having said that, it is my opinion that some circuitry benefit from a measure of increased leakage.  If you are happy with the sonics of your eq, be happy.

 

[JohnRoberts]

I am not a fan of "more" leakage, I even had to black ball one manufacturer's new series when the leakage made audible noise in the phantom blocking caps in the input of a mixer mic preamp.

This one cap series was the only time I ever experienced audible noise from leakage.

JR

 

[Rochey]

I'm about to finish an app note here about customers using Conductive Polymer Electrolytic caps.

Conductive Polymer is the new boy in town, offering lower ESR (Effective Series Resistance) than it's regular Aluminum Cap brothers, and better frequency response, however, from the looks of things, it's leakage current is about 20x higher!

please please please, don't use them on things like audio converter VCOM/VREF pins.

Sorry - the whole topic of leaky caps is a tough one right now.

/R

 

[gridcurrent]

This one cap series was the only time I ever experienced audible noise from leakage.

JR

Noise from leakage.  Tell us more, John.

 

[JohnRoberts]

This one cap series was the only time I ever experienced audible noise from leakage.

JR

Noise from leakage.  Tell us more, John.

I don't recall the specific details, I was managing a mixer engineering group at the time and as I recall the QA folks out in the factory identified a noise problem with a single production run of mixer channel strips in a mature product. One of my engineers traced the problem back to a new series of capacitors that had recently been approved.

Cap makers are constantly working on size and cost reductions. Prior to this experience engineering didn't bother to test for noise. AFAIK this is still a pretty rare example of leakage current being audible, but be warned it has happened at least once.

I don't recall the brand of capacitor, but IIRC is was one of the major cap vendors. We used them all. We gonged just that one new series and continued to use their other capacitors with no problems.

------------

I find it interesting that Rochey is dealing with leakage in another new series...  We always need to be alert when cap makers improve one characteristic dramatically... I recall having problems with the then new low ESR caps shorting out... While perhaps not as spectacularly as tantalum, the result was the same.. circuit no workee.

Don't ASSume that because some new cap is very good at one thing, it will be good at all things.

JR

 

[gridcurrent]

There's probably a reason why the circuit called for 400v capacitors. Increasing the voltage to be safe typically has no ill effects other than has been mentioned, of course lower than what potential will be present across the terminals will seriously ruin the capacitor and your day. The capacitance value and the dielectric has more to play with the sound, and even then isn't as drastic as one may expect to a certain extent.

tdstotler did not specify that the circuit called for 400V capacitors.  On the issue of increased leakage in higher voltage capacitors:  I repaired a Pultec MEQ-5 in which that some smart tech installed a high voltage coupling capacitor; the leakage destroyed the Triad HS-50. I have repaired a few mics including a C24 in which the leakage in the "upgrade" coupling capacitor destroyed the output transformer.  "no ill effects"?
Techs should spend more time with their Electronics 101 textbooks than the Digikey catalog capacitor pages.

 

[JohnRoberts]

tdstotler did not specify that the circuit called for 400V capacitors.  On the issue of increased leakage in higher voltage capacitors:  I repaired a Pultec MEQ-5 in which that some smart tech installed a high voltage coupling capacitor; the leakage destroyed the Triad HS-50. I have repaired a few mics including a C24 in which the leakage in the "upgrade" coupling capacitor destroyed the output transformer.  "no ill effects"?
Techs should spend more time with their Electronics 101 textbooks than the Digikey catalog capacitor pages.

I am not familiar with the mechanism you suggest that higher voltage capacitors would exhibit larger leakage than a lower rated voltage part in a given circuit application. Could you point me toward an Electronics 101 textbook that explains this?

Capacitor data sheets often describe leakage in terms of a constant X the amount of capacitance X the rated voltage..

This is logical. More capacitance means less resistance between the plates (more dielectric in parallel), and more voltage cranked into ohms law means more current leaking through that resistance.

A high voltage cap only leaks it's max leakage current at max rated voltage, so at lower voltage it's a fraction of that max.

You have me curious now, what is it about higher voltage caps that makes them more leaky?

JR



edit- For a given capacitance to have higher breakdown voltage for a given insulation, I guess the insulation needs to be thicker spacing the plates further apart. To keep the capacitance the same the plates need to have more area, so series resistance goes up, but there's more of it in parallel, seems like a wash..   

 

[gridcurrent]

"Could you point me toward an Electronics 101 textbook that explains this? "

Boylestad, 10th edition.
Further lab research confirmed the pattern.
As to the actual mechanism, you are the expert, tell us.

 

[JohnRoberts]

"Could you point me toward an Electronics 101 textbook that explains this? "

Boylestad, 10th edition.
Further lab research confirmed the pattern.
As to the actual mechanism, you are the expert, tell us.

If I knew I wouldn't be asking. 

I expect there to be differences between types of capacitor, and leakage modulated by actual working voltage but within a type of capacitor it seems resistive. On one cap data sheet I looked at yesterday they  used the same multiplier constant for the C x rated voltage for less than 25V and 35v to 50V, with a slight fixed minimum leakage in the higher voltage series (like 1 uA vs. .7 uA), suggesting a linear mechanism with some minor housekeeping for size of package.

Remarkably I found a copy of that textbook on the internet, that could be searched and found 40 references to leakage. I still didn't find the answer to my query.

They did describe leakage generally as a resistance (on the order of 100 megOhm) in parallel with the capacitance, which is consistent with my understanding, i.e. that more applied voltage results in more leakage due to simple resistance and ohms law.

I don't doubt that different dielectric types exhibit different leakage behavior.

Can you point me more specifically to the chapter or page in that book, I didn't find it in a quick scan and the book is over 1k pages.

I will resist the temptation to throw some caps on the bench since I can't easily parse fraction of uA, or hundreds of megohms resistance.  Sounds like a project for Bob Pease.

JR

 

 

[PRR]

> I can't easily parse fraction of uA

And OTOH, the smallest wire used in transformers can handle 10mA.

So I, an experienced thinker and smoker with a collection of EE 101 texts, am also mystified how "leakage destroyed the Triad HS-50 ... leakage in the "upgrade" coupling capacitor destroyed the output transformer."

Unless the caps or installation were just bad.

 

[johnR]

I too am a little mystified by the assertion that higher voltage caps leak more. Manufacturers normally specify leakage in terms of CV (ie. capacitance x applied voltage). Quoted figures are typically the same over a wide spread of voltage ratings for any given capacitor series. If a capacitor doesn't meet the leakage spec it's faulty.

 

[Gus]

100Hp AC drives I repaired in the past used some 1.2uf 1200VDC film caps.  I would leakage check them with a Sencore Lc102 at 999VDC(the highest voltage it went too)nice tool you could adjust the leakage voltage in 1 VDC steps.  The leakage was no better or worse than 1uf 50VDC film caps and the ESR measured about the same.

 

[tdstotler]

The circuit I was working on did not call for any specific voltages, I was using the 400V Poly Pro caps in the filter section. To clarify my question does
the voltage rating have any effect on the sound if the cap is only being used in the filter circuit.

 

[JohnRoberts]

The voltage rating affects the internal geometry of the capacitor while results in larger spacing and more plate area, so all things equal a much larger physical capacitor for a given dielectric and capacitance.

In principle the voltage should not make a difference, if the cap is 2" long instead of .2" that physical size and perhaps lead length could make a difference in practice. 

JR

 

[JTraverse]

The circuit I was working on did not call for any specific voltages, I was using the 400V Poly Pro caps in the filter section. To clarify my question does
the voltage rating have any effect on the sound if the cap is only being used in the filter circuit.

It will not.