A question for the electronics guys

[Amphiprion]

Hi,

I am in need of a capacitor technology. I am looking for an AC coupling capacitor for a signal between gain stages in a filter circuit I am designing. It has to be 1uF or larger (pref. MUCH larger for maintaining linear phase in the frequency range of interest). It has to be as low or lower distortion as an aluminum electrolytic DC blocking cap in this application. Unfortunately, I cannot use an aluminum electrolytic. Don't ask why, I just can't. As I see it, my options are:

1) Large (HUGE) C0G ceramic. Still have trouble even getting to 1uF in SMT packages. (THD ~-100 to 130db depending on cap selection)

2) Tantalum. Can't use because of the distortion plots I've seen for them indicate they would be very bad (THD approx. -70dB)

3) Film. Only downside is large size and still can't get very big values. But currently what I am leaning towards. (THD ~-100 to 130dB depending on dielectric)

4) Polymer aluminum electrolytic. A700 series from Kemet. I have no idea what the distortion specs on these are like. I am hoping they are more like standard aluminum electrolytic in distortion (~-80 to 100dB as DC blocks) than tantalums. I am really interested if anyone has info on these.

5)? Recommend me something.

Other constraints are 5mm total height (although I can bend a radial leaded part over and RTV it to the board or something similar). SMT preferred.

If anyone out there can help, it would be much appreciated. Polarized caps are obviously OK. Thanks for any help!

[mazurek]

I generally use panasonic ECQ-P(Z) series polypropylene capacitors. I believe they are available up to 0.47uF or so from Digikey, and I have paralleled those to get higher. Make sure to get the 50V rated ones if you can as they are the smallest. I know these measure quite well in filter circuits.

I've used higher valued film caps of different types, but I've got no experience in xover circuits though.

The Jensen Transformers app notes on input and output stages recommend panasonic bipolar electrolytic capacitors as being pretty good, but I've not used these myself.

Some will say bypass a bad capacitor with a good one, but I'm sure you are experienced enough to know that without measuring, it can be an unexpected combination of R-C-L's that don't necessarily improve things.

Another thing, I'm sure you already know, but can be helpful to think about. If you are active, you have leeway in sizing. You could scale R's proportionally (and see whether thermal noise becomes important), or decompose a filter stage with giant values into two filter stages with more reasonable values. I have a bass adjust option in my crossover that moves a f3=20Hz Q=0.5 LR2 -> f3=40Hz Q=0.5. I used two first order filters for the job rather than one Linkwitz transform circuit.

Sorry, I have no real experience with SMT so my thoughts may not have been helpful.

[Face]

http://www.wima.de/EN/products.htm

[benchtester]

parallel as many 0.22 uF as you like:

http://www.niccomp.com/Catalog/nsmx.pdf