Even capacitor manufacturers sometimes get it wrong... - UK Vintage Radio Repair and Restoration Discussion Forum

daviddeakin · 14th Apr 2021, 4:07 pm

I recently discovered this app note by Cornell Dubilier, which had an interesting claim I had not seen before:

"If two, same-value, aluminum electrolytic capacitorsare connected in series, back-to-back with the positive terminals or the negative terminals connected, the resulting single capacitor is a non-polar capacitor equal in capacitance to either of the original pair. The two capacitors rectify the applied voltage and act as if they had been bypassed by diodes. When voltage is applied, the correct-polarity capacitor gets the full voltage. However, on a capacitor meter with no bias voltage the two capacitors measure half capacitance as you expect from capacitors in series."

Well after a number of frequency response tests I've satisfied myself that they are not correct. The effective capacitance is always half, as you would expect, whatever the bias voltage or signal amplitude is used. The 'happy' capacitor always prevents the 'unhappy' one from behaving like a short. Just goes to show even the experts can get it wrong!

duncanlowe · 14th Apr 2021, 4:14 pm

Interesting. I mentioned this ability to create a non polarised this way once, in an interview. I was told I was talking rubbish. The same was said about a particular DC motor application where I said the intitial current (at stall) was around 400A and again I was told, 'no maybe 40A'

I didn't get a place on that course, kind of glad really. For a while I wondered if the 'experts were actually right, as they must be clever as those very same 'experts' would have been the ones teaching me.

ortek_service · 14th Apr 2021, 4:59 pm

I had heard that if you do this with Tantalums, then one would soon breakdown with reverse-voltage and effectively be a s/c. But wasn't entirely convinced, have seen Tantalums fitted backwards across supplies that didn't always go bang straight-away.

But 'Bipolar' / Non-Polarised (NP) Electrolytics are commonly used in speaker cross-overs, so there must be some info about these (which I believed to be just 2 back-to-back Polarised ones)

G6Tanuki · 14th Apr 2021, 6:48 pm

When I've had to use a pair of back-to-back electrolytics, I've always added a pair of diodes in parallel with the capacitors - that way the capacitors are only ever subjected to a 'reverse' voltage equivalent to the forward-voltage-drop-of-a-diode (0,6v for silicon).

"Audio" types might fret about the risk of the diodes introducing non-linearities, but I'd respond by pointing out that the diode's non-linearity is bypassed by a rather large capacitor!

Radio Wrangler · 14th Apr 2021, 7:14 pm

Ah but the capacitor isn't infinite, so there will always be a little bit of distortion leaking past... and through such cracks in our universe, the dungeon dimensions leak in. People with infinite hearing capability will hear it.

You can't argue with people who don't use logic or a sense of proportion.

It works OK for us mere mortals

David

Diabolical Artificer · 15th Apr 2021, 6:13 am

How do back to back electrolytic caps work in a speaker, there is no polarising voltage, no DC only AC? How is a non polarised electrolytic made? It's a puzzler this one, need some coffee.

Andy.

kalee20 · 15th Apr 2021, 10:01 am

My understanding is bi-polar electrolytics are made with two foils which have both been pre-formed with oxide, so either will stand the working voltage.

It means the resulting component is larger than a 'normal' electrolytic, and leakage current is a bit higher because the component can't be reformed after rolling up and stuffing into a can, which inevitably is going to cause cracks in the oxide.

But as for connecting two in series - I have heard of the practice, though it sounds flaky to me. And I bet the resulting capacitance is signal-level dependent, at very low levels, where neither capacitor depolarises, the capacitance is equivalent to the two in series, but at high voltage levels, one of them DOES behave like a diode so the capacitance goes up, and probably non-linearly, and not very reformably either!

14th Apr 2021, 4:07 pm	#1
daviddeakin Hexode Join Date: Apr 2007 Location: W Yorks, UK. Posts: 407	Even capacitor manufacturers sometimes get it wrong... I recently discovered this app note by Cornell Dubilier, which had an interesting claim I had not seen before: "If two, same-value, aluminum electrolytic capacitorsare connected in series, back-to-back with the positive terminals or the negative terminals connected, the resulting single capacitor is a non-polar capacitor equal in capacitance to either of the original pair. The two capacitors rectify the applied voltage and act as if they had been bypassed by diodes. When voltage is applied, the correct-polarity capacitor gets the full voltage. However, on a capacitor meter with no bias voltage the two capacitors measure half capacitance as you expect from capacitors in series." Well after a number of frequency response tests I've satisfied myself that they are not correct. The effective capacitance is always half, as you would expect, whatever the bias voltage or signal amplitude is used. The 'happy' capacitor always prevents the 'unhappy' one from behaving like a short. Just goes to show even the experts can get it wrong!

14th Apr 2021, 4:14 pm	#2
duncanlowe Nonode Join Date: Jan 2016 Location: Stafford, Staffs. UK. Posts: 2,532	Re: Even capacitor manufacturers sometimes get it wrong... Interesting. I mentioned this ability to create a non polarised this way once, in an interview. I was told I was talking rubbish. The same was said about a particular DC motor application where I said the intitial current (at stall) was around 400A and again I was told, 'no maybe 40A' I didn't get a place on that course, kind of glad really. For a while I wondered if the 'experts were actually right, as they must be clever as those very same 'experts' would have been the ones teaching me.

14th Apr 2021, 4:59 pm	#3
ortek_service Octode Join Date: May 2018 Location: Northampton, Northamptonshire, UK. Posts: 1,440	Re: Even capacitor manufacturers sometimes get it wrong... I had heard that if you do this with Tantalums, then one would soon breakdown with reverse-voltage and effectively be a s/c. But wasn't entirely convinced, have seen Tantalums fitted backwards across supplies that didn't always go bang straight-away. But 'Bipolar' / Non-Polarised (NP) Electrolytics are commonly used in speaker cross-overs, so there must be some info about these (which I believed to be just 2 back-to-back Polarised ones)

14th Apr 2021, 6:48 pm	#4
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 13,998	Re: Even capacitor manufacturers sometimes get it wrong... When I've had to use a pair of back-to-back electrolytics, I've always added a pair of diodes in parallel with the capacitors - that way the capacitors are only ever subjected to a 'reverse' voltage equivalent to the forward-voltage-drop-of-a-diode (0,6v for silicon). "Audio" types might fret about the risk of the diodes introducing non-linearities, but I'd respond by pointing out that the diode's non-linearity is bypassed by a rather large capacitor!

14th Apr 2021, 7:14 pm	#5
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,874	Re: Even capacitor manufacturers sometimes get it wrong... Ah but the capacitor isn't infinite, so there will always be a little bit of distortion leaking past... and through such cracks in our universe, the dungeon dimensions leak in. People with infinite hearing capability will hear it. You can't argue with people who don't use logic or a sense of proportion. It works OK for us mere mortals David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

15th Apr 2021, 6:13 am	#6
Diabolical Artificer Dekatron Join Date: Jul 2009 Location: Sleaford, Lincs. UK. Posts: 7,661	Re: Even capacitor manufacturers sometimes get it wrong... How do back to back electrolytic caps work in a speaker, there is no polarising voltage, no DC only AC? How is a non polarised electrolytic made? It's a puzzler this one, need some coffee. Andy. __________________ Curiosity hasn't killed this cat...so far.

15th Apr 2021, 10:01 am	#7
kalee20 Dekatron Join Date: Feb 2007 Location: Lynton, N. Devon, UK. Posts: 7,081	Re: Even capacitor manufacturers sometimes get it wrong... My understanding is bi-polar electrolytics are made with two foils which have both been pre-formed with oxide, so either will stand the working voltage. It means the resulting component is larger than a 'normal' electrolytic, and leakage current is a bit higher because the component can't be reformed after rolling up and stuffing into a can, which inevitably is going to cause cracks in the oxide. But as for connecting two in series - I have heard of the practice, though it sounds flaky to me. And I bet the resulting capacitance is signal-level dependent, at very low levels, where neither capacitor depolarises, the capacitance is equivalent to the two in series, but at high voltage levels, one of them DOES behave like a diode so the capacitance goes up, and probably non-linearly, and not very reformably either!