Bipolar Electrolytics in Phantom power lines?

Searched the forum but couldn't seem to find suficcient info on this one...

For Phantom power decoupling of mic lines we usually use Electrolytics (provided there's no Xformer to do the job etc.).
Unipolar electrolytics - at least good ones - seem to work well as long as they're actually polarized. With phantom power off, though (e.g. when connecting ribbon mic's), they would be run unpolarized and thus introduce significant distortion to the signal.
Bipolar electrolytics don't seem to have this disadvantage IIRC. Hence the question:

Should one rather use bi-polar 'lytics for Phantom power decoupling?
Do bi-polar types even have drawbacks in comparison to uni-polar types regarding signal quality?


Thanks for any input,
cheers,
Volker

i don'tunderstand.

[quote author="CJ"]i don'tunderstand.[/quote]
Er... which part is it that you don't understand?

"Aluminum electrolytic capacitors are polar capacitors. Where necessary, voltages of opposite polarity should be prevented by connecting a diode. The diode?s conducting-state voltage of approximately 0,8 V is permissible. Reverse voltages ≤ 1,5 V are tolerable for a duration of less than 1 second, but not in continuous or repetitive operation."

IMO, an electrolytic capacitor works just fine (without distortions) with a small reverse audio voltage (up to line levels).

Regards,
Milan

You have to be aware that a 'lytic in your application shows a frequency dependent distortion mechanism. Distortion starts to get measurable about at 1/10 of the -3 dB point of the high-pass filter. So what you can do is to simply chose the cap/resistor such that the -3 dB point is below 2 Hz (say 1 Hz for safety). Another solution is to take a bipolar, as suggested. A third solution is the back-to-back trick, preferably with some bias in between.

If you want excellent phase linearity, you need to chose the -3 dB point below 1 Hz anyway.

Samuel

Samuel,
I do know that distortion rises at the edge of the passband, i.e. when the capacitor begins to "work" harder, storing/releasing large amounts of charge. The logical consequence of which would be to move the cutoff freq well outside the desired freq range.
What I ask myself is if
a) unipolar lytics maybe even distort within the passband if operated in biased state? I remember sources that state that the requirement to operate unipolar ones in biased state isn't only due to possible damage by temporary reverse bias, but also due to sonic alterations...
b) bipolar lytics have any inferior properties in comparison to unipolar ones operated in polarized, or any other drawbacks such as higher ESR etc.?

Thanks again,
Volker

According to capacitor experts like Cyril Bateman, who has written about caps extensively in Electronics World and worked in the industry iirc, and supported by data from Jensen Transformers, (specific) bipolar 'lytics work better than either biased electrolytics, or back-to-back polar 'lytics with their midpoint connection being biased appropriately. I know this runs counter to intuition and also what many people swear by, but that's life. Greg Timbers at JBL even made speakers that had a battery inside to bias the back-to-back 'lytics, and he is definitely of the aurically aural persuasion, but Jensen et al. 's data would indicate that this strategy is not optimal.

I don't time to dig up the references but they have been at least partially noted in other Prodigy threads.

It's still a good idea as mentioned above to have the important cutoff frequencies associated with the caps well below the lowest signal frequency of interest.

Thanks all! :thumb:

I've got a prize for the first person to tell me what the problem is.

Click to expand...

Now you're in for it. Keith AKA SSL Tech is in da hooouse...

:wink: :sam:

[quote author="mediatechnology"]

See www.tenmilecreek.net/images/ssl_82e149.pdf

There is a problem with this circuit I found commissioning a console not related to the caps. The problem is to the left of the mic pad. Hint: This console was noisy (bubbly/rocky) with phantom on but very quiet when it was off. Several of the console's 56 inputs had the problem and the 48V supply was clean. I've got a prize for the first person to tell me what the problem is.[/quote]

With the the phantom supplied to pin 2 there's now flowing a DC current through those 100k resistors (R_8 & R_punchhole). If these resistors are fiddly that could cause noise.

Bye,

Peter

Thanks Wayne ! :thumb:

The fun of joining your quiz was already enjoyable enough, but those 1510's will maximize the experience even further ! Much appreciated.

And indeed quite surprising those 100k gave you this trouble.

I'm not sure the reasoning for putting them there - the only thing I can think of is to keep a little bit of bias across the caps in a phantom on unpatched input condition.

Click to expand...

I guess it wouldn't have cost them much to provide say a regulated 54V supply for that center-cap-point (as now connected to the unswitched phantom,node '2') and then all benefits would be there without risk of noisy crackling.

Uh,and there's the (quite remote) possibility that a (very) leaky cap generates DC-current through the other set of 100k resistors, possibly resulting in noise as well when these 100k are from that same bad bad batch :wink:

Best regards,

Peter

the only thing I can think of is to keep a little bit of bias across the caps in a phantom on unpatched input condition.

Click to expand...

If I'm not mistaken they could have had all goodies even without that additional regulated supply that I proposed.
And they could have saved 2 resistors per channel !

Simply use the -18V supply to connect those center-cap-100k resistors to and adjust cap-polarity accordingly.

Whether phantom is on or off, all four caps are now always seeing a healthy dose of DC-voltage.

I saw these are 100V caps already, so also no VDC-upgrading required.

Bye,

Peter