Hi,
Sometimes I see the use of polarized capacitors exposed to the outside world, such as the 10uF output capacitor in this pedal: http://www.4qdtec.com/gifs/phaser.gif
This surprises me a bit. Suppose the next pedal in the chain has an input coupling cap that is not orders larger than 10uF, and also DC couples the input to a virtual ground at -4.5V. (I have never seen such a circuit like that, but I have seen circuits with only a negative rail, at least in the context of germanium transistors. And so it seems conceivable to me that an audio circuit might use a virtual ground significantly below 0.) In this case, wouldn't the output cap see a DC reverse voltage of up to 4.5V, depending on the size of the input cap of the next pedal, and then possibly suffer reverse breakdown? Is this really OK?
Sometimes I see the use of polarized capacitors exposed to the outside world, such as the 10uF output capacitor in this pedal: http://www.4qdtec.com/gifs/phaser.gif
This surprises me a bit. Suppose the next pedal in the chain has an input coupling cap that is not orders larger than 10uF, and also DC couples the input to a virtual ground at -4.5V. (I have never seen such a circuit like that, but I have seen circuits with only a negative rail, at least in the context of germanium transistors. And so it seems conceivable to me that an audio circuit might use a virtual ground significantly below 0.) In this case, wouldn't the output cap see a DC reverse voltage of up to 4.5V, depending on the size of the input cap of the next pedal, and then possibly suffer reverse breakdown? Is this really OK?
It's never ok to reverse-bias a polarized capacitor. But, you can use a bipolar type,
which is essentially two polarized capacitors in anti-series. Then any DC voltage
(and polarity) within the rating is fine.
http://www.mouser.com/ProductDetail...=sGAEpiMZZMtZ1n0r9vR22ZGaUoI0JcRf3FdaTwYcMPQ=
TI use low ESR polarized electrolytic capacitors to decouple the output stage of some of their class D amp reference designs and you can apply 10% of reverse votage on a polarized electrolytic capacitors witout any damage on the long term.According to measurements some high quality polarized electrolytic capacitors are able of extremely low levels of THD
Last edited:
Negative ground has been most popular at least since 1960. So designers that use a polarized electrolytic cap for input with the minus pointed out have a pretty good chance of getting it right.
Polyprophylene has 10x the cost and 10x the size of a polarized electroltyic, so in a market where low price makes the sale, you see a lot of them used for input caps.
Polyprophylene has 10x the cost and 10x the size of a polarized electroltyic, so in a market where low price makes the sale, you see a lot of them used for input caps.
Hi Chaerin,
I don't think the case you describe would actually do any damage, because as long as at least one of the series-connected electrolytic capacitors is biased with the correct polarity, it will limit the DC current to very low values (only leakage). It is unlikely that that current will be big enough to overheat the other polarized capacitor. Nonetheless, I would prefer to have a large resistor to ground at the output, so you know that the DC will settle to something near zero.
In general, polarized aluminium electrolytic capacitors can handle about 2 V of reverse voltage. Apply more reverse voltage and they will start to conduct DC. Apply a lot more than 2 V reverse voltage and they will start to conduct a lot of direct current, dissipate a lot of power, become very hot and explode.
I don't think the case you describe would actually do any damage, because as long as at least one of the series-connected electrolytic capacitors is biased with the correct polarity, it will limit the DC current to very low values (only leakage). It is unlikely that that current will be big enough to overheat the other polarized capacitor. Nonetheless, I would prefer to have a large resistor to ground at the output, so you know that the DC will settle to something near zero.
In general, polarized aluminium electrolytic capacitors can handle about 2 V of reverse voltage. Apply more reverse voltage and they will start to conduct DC. Apply a lot more than 2 V reverse voltage and they will start to conduct a lot of direct current, dissipate a lot of power, become very hot and explode.
Post1 sch shows a high Meg leakage grounding resistor across the input terminals.
A similar high Meg leakage grounding resistor should be added across the output terminals.
C20 & C21 in the post10 sch should also have the high Meg resistors.
470uF is exceptionally high.
This may be a remnant from the old fashioned requirement to have 600ohms driving capability.
Rane show 220uF bi-polar Panasonic for this duty.
But no one specifies these output for 600ohms loading. Interference attenuation gets better as the load impedance is increased.
One can usefully reduce the bi-polars to <100uF and might get away with non-polar 1uF if you know the loading will only be 100k.
A similar high Meg leakage grounding resistor should be added across the output terminals.
C20 & C21 in the post10 sch should also have the high Meg resistors.
470uF is exceptionally high.
This may be a remnant from the old fashioned requirement to have 600ohms driving capability.
Rane show 220uF bi-polar Panasonic for this duty.
But no one specifies these output for 600ohms loading. Interference attenuation gets better as the load impedance is increased.
One can usefully reduce the bi-polars to <100uF and might get away with non-polar 1uF if you know the loading will only be 100k.
Last edited:
Thanks to everyone for the responses.
Interesting... so in the case that C1 enters reverse breakdown and wants to act like a short circuit, C2 will charge completely and the voltage across C1 will tend toward zero, regardless of the nominal ratio between C1 and C2, or, oh, I guess precisely a bipolar capacitor has been formed.
So the only requirement for capacitors not to explode is that everyone follows the convention of putting polarized capacitors in a polarity which assumes external ports are at DC 0. Then even if it is not DC 0, only one capacitor will be backwards. And it's reasonable to assume everyone will follow this convention, otherwise things would break under extremely common conditions.
So, I think I can see why it works and probably won't ever explode, and I think I can also see why it's probably a bad idea anyway? (e.g. what if the next pedal has a wiring fault and the input is shorted to -9V for no reason)
This is designed to be connected to a passive load, though, right? So they would always see DC 0 anyway.
Interesting... so in the case that C1 enters reverse breakdown and wants to act like a short circuit, C2 will charge completely and the voltage across C1 will tend toward zero, regardless of the nominal ratio between C1 and C2, or, oh, I guess precisely a bipolar capacitor has been formed.
So the only requirement for capacitors not to explode is that everyone follows the convention of putting polarized capacitors in a polarity which assumes external ports are at DC 0. Then even if it is not DC 0, only one capacitor will be backwards. And it's reasonable to assume everyone will follow this convention, otherwise things would break under extremely common conditions.
So, I think I can see why it works and probably won't ever explode, and I think I can also see why it's probably a bad idea anyway? (e.g. what if the next pedal has a wiring fault and the input is shorted to -9V for no reason)
This is designed to be connected to a passive load, though, right? So they would always see DC 0 anyway.
Last edited:
Sorry DF96, but I don't believe that's correct. For the reasons that chaerin and I explained, the rectifying properties of the electrolytic capacitors will always make the output DC voltage settle to some safe value, whatever it may be. The same happens when you connect two normal electrolytics in anti-series and use the series connection as a bipolar electrolytic capacitor.
Of course it is good practice to add ground leak resistors, for example to prevent loud thumps when you switch from one signal source to another. But that has nothing to do with explosions, even though it may sound like one when the volume is turned up high.
Of course it is good practice to add ground leak resistors, for example to prevent loud thumps when you switch from one signal source to another. But that has nothing to do with explosions, even though it may sound like one when the volume is turned up high.
Hello,
As far as the size of the ecetro caps goes - according to tests published by Cycil Bateman (spelling?) on using caps in the signal path -
Electrolytic caps:
1. The larger the (microfarads) size, the smaller the distortion from the cap.
2. The larger the voltage rating of the cap, the smaller the distortion as well.
3. Using bi-polar cap instead of polar electrolytic gives quite a bit less distortion.
Combining all three (larger that really needed, higher voltage rating and making it bipolar) caused small signal distortion to get way down close to measurement limits.
Bottom line, if use electrolytic's in the signal path, it is possible to really minimize the distortion if you keep the previous facts in mind.
As far as the size of the ecetro caps goes - according to tests published by Cycil Bateman (spelling?) on using caps in the signal path -
Electrolytic caps:
1. The larger the (microfarads) size, the smaller the distortion from the cap.
2. The larger the voltage rating of the cap, the smaller the distortion as well.
3. Using bi-polar cap instead of polar electrolytic gives quite a bit less distortion.
Combining all three (larger that really needed, higher voltage rating and making it bipolar) caused small signal distortion to get way down close to measurement limits.
Bottom line, if use electrolytic's in the signal path, it is possible to really minimize the distortion if you keep the previous facts in mind.
In a recent modification I am doing, a circuit called for a 2.2uF capacitor. So I did the comparison:
Electrolytic--2.2uF 50v 5mm dia. x 11mm H; ESR=1.8Ω at 1K Hz $0.36; Panasonic EEU-FC1H2R2H
Polypropylene--2.2uF 450v 17x11mm, 7.5mm LS, 14mm H; ESR =0.08Ω at 1 K Hz; $0.93. Panasonic ECW-FD2W225K
So the PP is ~3x the cost of the electrolytic, 22x less ESR, and, with the 7.5mm lead spacing, will fit on the board. Yes, it is bigger than the electrolytic, but MUCH MUCH better performance at a reasonable cost increase.
I have found, however than the 2.2uF is about the limit that you can reasonably fit a polypropylene cap as a replacement, Anything that needs more capacitance just gets too big for polypro.
But I DID find a great new 10uF, 50 v polyester cap that will fit most places---a WIMA MKS2B051001N00JO00--only 7x11mm, and 8 x less ESR than an a 10uF electrolytic.
Film caps rule!!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.