"zero" biased polarized coupling caps

[Adam S]

I've have some cascaded op-amp stages, each providing +20dB, and worst
case DC offset is +-30mV at the output of each stage.

I want to AC couple each amplifier stage with polarized capacitors. I
cannot find any documentation recommending the use of polarized
capacitors for continuous reversed bias, even when the bias is very
small (< 50mV).

However there is the following PDF from kyocera,
http://www.kyocera.co.jp/prdct/electro/pdf/technical/revtant.pdf

basically saying tantalums should never be designed in circuits where a
continuous reverse bias exists and since they fail to present a
definition of revers biased I find it hard to believe a microvolt is
going to do harm them. So whats the truth about using tantalum's in such
applications ?

 

[Meindert Sprang]

I've have some cascaded op-amp stages, each providing +20dB, and worst
case DC offset is +-30mV at the output of each stage.

I want to AC couple each amplifier stage with polarized capacitors. I
cannot find any documentation recommending the use of polarized
capacitors for continuous reversed bias, even when the bias is very
small (< 50mV).

What size of capacitors do you need? I.e. do you really need polarized caps?

Meindert

 

[Tim Williams]

I've graphed the I-V characteristic of electrolytics and it looks pretty
flat up to 25% of the rated voltage, reverse. It improves with time when
voltage is applied.

I don't see anything wrong with it.

Tim

 

[Pooh Bear]

I've have some cascaded op-amp stages, each providing +20dB, and worst
case DC offset is +-30mV at the output of each stage.

I want to AC couple each amplifier stage with polarized capacitors. I
cannot find any documentation recommending the use of polarized
capacitors for continuous reversed bias, even when the bias is very
small (< 50mV).

What kind of signal ? Audio ?

Graham

 

[Spehro Pefhany]

I've have some cascaded op-amp stages, each providing +20dB, and worst
case DC offset is +-30mV at the output of each stage.

I want to AC couple each amplifier stage with polarized capacitors. I
cannot find any documentation recommending the use of polarized
capacitors for continuous reversed bias, even when the bias is very
small (< 50mV).

However there is the following PDF from kyocera,
http://www.kyocera.co.jp/prdct/electro/pdf/technical/revtant.pdf

Hmm. AVX (nee Aerovox) is owned by Kyocera (Kyoto Ceramics) eh?
Did not know that.

basically saying tantalums should never be designed in circuits where a
continuous reverse bias exists and since they fail to present a
definition of revers biased I find it hard to believe a microvolt is
going to do harm them. So whats the truth about using tantalum's in such
applications ?

I've used aluminum electrolytics in such applications with no
problems, in volume applications over many years. Tantalums- I have a
bad translation from Japanese that seems to say (not what it says, but
it doesn't make sense the way they wrote it) that for "short periods"
(not defined) 10% of rated voltage or 0.5V max is okay. Which fits
with AVX pdf.

I'd guess (which is worth what you paid for it) that -50 or -100mV
should be okay over temperature and time, and -30mV should be quite
safe.


Best regards,
Spehro Pefhany

 

[Adam S]

What size of capacitors do you need? I.e. do you really need polarized caps?

Meindert

They are 10uF into inverting amplifiers with Rin=2.2k, Rf=22k (gain -10)
with a 7Hz lower cutoff. Freq of interest is 100Hz~100kHz. Its for an
instrument to measure AC signal levlel at a single frequency. The
excessively low cutoff frequency to minimize phase error. DC coupling is
out of the question at +60dB.
Would you suggest X7R/X5R ceramics ?

 

[Spehro Pefhany]

They are 10uF into inverting amplifiers with Rin=2.2k, Rf=22k (gain -10)
with a 7Hz lower cutoff. Freq of interest is 100Hz~100kHz. Its for an
instrument to measure AC signal levlel at a single frequency. The
excessively low cutoff frequency to minimize phase error. DC coupling is
out of the question at +60dB.
Would you suggest X7R/X5R ceramics ?

Maybe. Microphonics might an issue. Can you increase the resistances
by 100:1 and use some nice 0.1uF film caps, or will bias currents or
something else screw things up?


Best regards,
Spehro Pefhany

 

[Meindert Sprang]

They are 10uF into inverting amplifiers with Rin=2.2k, Rf=22k (gain -10)
with a 7Hz lower cutoff. Freq of interest is 100Hz~100kHz. Its for an
instrument to measure AC signal levlel at a single frequency. The
excessively low cutoff frequency to minimize phase error. DC coupling is
out of the question at +60dB.
Would you suggest X7R/X5R ceramics ?

Indeed. Although Spehro's reply got me thinking. I didn't know ceramics
suffered from microphonics.

Meindert

 

[Ben Bradley]

Indeed. Although Spehro's reply got me thinking. I didn't know ceramics
suffered from microphonics.

They don't suffer from it, ceramics ENJOY being microphonic!

Okay, sorry. But I really must try making some "electronic drums"
using nothing but ceramic capacitors. Too bad they're so brittle.