I need one bipolar 4.7uF 50v electrolytic capacitor. I have polarized 10uf 50v capacitors and bipolar 2.2uF 50v capacitors. I can place two of the 10uF in series (negative leads joined) or two of the 2.2uF in parallel to create one bipolar capacitor that is close to 4.7uF. Which is the better solution and why?
3
-
1\$\begingroup\$ What are the ESR and ESL of each? How does the series L and R in choice 1 / the parallel L and R in choice 2 affect your circuit? Choose whichever circuit has a self-resonant frequency that matches your use case. \$\endgroup\$– Matt SCommented Dec 25, 2022 at 4:13
-
1\$\begingroup\$ I was asking as a general question but I suppose there isn't a general answer. I don't have an ESR meter to check the BP capacitor that I'm replacing. It's a Nippon Chemi-Con KME series and the datasheets I found from 20 years ago when this would have been manufactured don't list an ESR value. I replaced it with two 10uF polarized Panasonic FR-A series capacitors and that works. The Panasonic datasheet doesn't specify an ESR value but "low ESR" is listed as a feature. The original BP capacitor was connected to a tiny board with a Murata IC (MMC-PCS03). I can't find a datasheet for it. \$\endgroup\$– tomasmCommented Dec 28, 2022 at 19:36
-
\$\begingroup\$ Why do you want to do this ? film would be better. \$\endgroup\$– AutisticCommented Apr 11 at 8:32
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
7
They appear to be roughly equivalent. But you have forgotten several things.
The diodes you will need with the series connection1. Each capacitor will need a diode across it, so that when it would otherwise be subject to a reverse voltage, its reverse is limited to less than a volt so that the capacitor will probably survive.
Ideally you would not simply allow the mid point to self bias via these diodes, but would provide some additional bias, so the capacitors come nowhere near to a negative voltage. This requires extra voltage headroom, so you ought to have a higher voltage rating in the series capacitors than your final target.
Tolerances. Electrolytics are famous (infamous?) for poor capacitance tolerance. If the two series capacitors are 9 uF and 11 uF, then the applied voltage will divide in the same ratio between them, and the mid point will rectify up to give you more than 50 V on the smaller capacitor, even with only 50 V peak applied.
1 - you could leave out the parallel diodes, but the capacitors would deal with the reverse voltage by destroying their dielectric, at best severely limiting their lifetime.
-
1\$\begingroup\$ There should be no need for parallel diodes in the case of the series connection of the two polarised capacitors (negatives connected or positives connected). It is a standard thing to do, to connect two polarised capacitors in series (without parallel diodes) to create what is effectively a non-polarized capacitor. What destroys a polarised capacitor is the large reverse current that flows when its reverse voltage is larger than about a volt. In the reverse series connection of two polarised capacitors this reverse current is limited to the leakage current of the forward biased capacitor. \$\endgroup\$– user173271Commented Dec 25, 2022 at 7:40
-
\$\begingroup\$ @James I'd like you to do a SPICE simulation of two series capacitors, starting from zero, applying an AC voltage. Plot the voltage waveform across each. Post it as another answer. This difference of our opinion deserves a proper exploration. \$\endgroup\$– Neil_UKCommented Dec 25, 2022 at 11:49
-
1\$\begingroup\$ I haven't got spice software and I've never used it, I've always preferred to learn from investigations using real components. My previous comment was based on real measurements from an investigation I undertook on two series, polarised electrolytics some time ago. I found that reverse biasing a single electrolytic by more than 1.0/1.5 V (or so) causes a large current to flow and it's the heating effect of that current that generates gas and pops the electrolytic! However with two series electrolytics (with opposite biases) the current flowing between the plates of the reverse .... \$\endgroup\$– user173271Commented Dec 25, 2022 at 13:02
-
1\$\begingroup\$ .... biased capacitor is limited to the leakage current of the forward biased capacitor (since the two capacitors are in series) even if the reverse biased capacitor charges to well over 1.5 V. I've just returned from having a Christmas pint in my local pub and so I hope that all makes sense! Generally I've learned its best not to post after returning from the pub! :-) \$\endgroup\$– user173271Commented Dec 25, 2022 at 13:06
-
\$\begingroup\$ @James never used SPICE? That's like fighting with one arm tied behind your back. The one you're presently using, the real world, is absolutely essential. But simulation is too useful a tool to ignore. Want to use ideal components? Measure voltage at THz or without loading the circuit at all? Iterate quickly between different values. I don't know any engineers who ignore simulation. \$\endgroup\$– Neil_UKCommented Dec 25, 2022 at 17:59