buschfsu
Well-known member
why chose one over the other? i know what they are just not why you would need a polarized vs nonpolarized cap.
thanks
thanks
polarized vs nonpolar caps in circuit design
- Thread starter buschfsu
- Start date
Help Support GroupDIY Audio Forum:
NewYorkDave
Well-known member
One big factor is that, above a certain value and voltage rating, the only affordable, reasonably compact choice is a polarized cap.
jensenmann
Well-known member
Polarized caps don´t like to be used reversed. Being reversed they switch into a kind-of-diode-mode and cause distortion and will die in long term. But if the circuid provides enough polarization voltage nothing will speak against pol. caps.
[quote author="buttachunk"]polarized caps also cause audible distortion when not given the proper voltage, even if not reversed.
bipolar caps sound sweeter due to reduced distortion, generally, unless the designer specifies the proper voltages for the caps at all stages. this is not practical in many designs, as caps are generally there to bias other components (not the other way around).[/quote]
You may be making a semantic distinction. Polar caps are reasonably well behaved (well sort of) as long as + terminal is + wrt - terminal. If terminal voltage reverses due to DC and/or AC operating conditions, non-linearity happens... Are you talking about some non-linearity where terminals are not reversed wrt DC level?
JR
bipolar caps sound sweeter due to reduced distortion, generally, unless the designer specifies the proper voltages for the caps at all stages. this is not practical in many designs, as caps are generally there to bias other components (not the other way around).[/quote]
You may be making a semantic distinction. Polar caps are reasonably well behaved (well sort of) as long as + terminal is + wrt - terminal. If terminal voltage reverses due to DC and/or AC operating conditions, non-linearity happens... Are you talking about some non-linearity where terminals are not reversed wrt DC level?
JR
clintrubber
Well-known member
How should I imagine non-polarized electrolytic caps?
For some reason I've always imagined them as two polarized caps in anti-series and it seems to be correct, see for instance page 51 of
http://www.elfa.se/en/fakta.pdf:
Based on the above that I'd say (+) we can also use two physically separate polarized caps which is more convenient to find in various values and (+) has the benefit of being able to lift or pull the middle-node to a supply-rail by means of a high value resistor for correctly polarizing both caps. (-) It'll require somewhat more board space though.
Bye,
Peter
For some reason I've always imagined them as two polarized caps in anti-series and it seems to be correct, see for instance page 51 of
http://www.elfa.se/en/fakta.pdf:
Based on the above that I'd say (+) we can also use two physically separate polarized caps which is more convenient to find in various values and (+) has the benefit of being able to lift or pull the middle-node to a supply-rail by means of a high value resistor for correctly polarizing both caps. (-) It'll require somewhat more board space though.
Bye,
Peter
[quote author="buttachunk"][quote author="JohnRoberts"]Are you talking about some non-linearity where terminals are not reversed wrt DC level?[/quote]
John,
just noting that when given no positive DC on the positive terminal, or reverse voltage, polarized caps can cause distortion and/or other problems. I've seen designs utilizing polarized caps with no biasing voltage on them, which is known to cause distortion.[/quote]
I'm suggesting that distortions measured near the pole frequency of a polar blocking cap are in fact caused by that cap being reverse biased during a portion of the ac waveform. Proper application using arbitrarily low pole frequencies can be reasonably linear.
I would go as far as to suggest that polars should not be used to define any audible frequency pole but for DC blocking while tuned a few octaves lower should be neutral to the audio path.
JR
John,
just noting that when given no positive DC on the positive terminal, or reverse voltage, polarized caps can cause distortion and/or other problems. I've seen designs utilizing polarized caps with no biasing voltage on them, which is known to cause distortion.[/quote]
I'm suggesting that distortions measured near the pole frequency of a polar blocking cap are in fact caused by that cap being reverse biased during a portion of the ac waveform. Proper application using arbitrarily low pole frequencies can be reasonably linear.
I would go as far as to suggest that polars should not be used to define any audible frequency pole but for DC blocking while tuned a few octaves lower should be neutral to the audio path.
JR
walter
Well-known member
I have seen nonpolar caps used in a tape deck as a motor starting cap where the motor runs forward and reverse.
G
Guest
Guest
[quote author="walter"]I have seen nonpolar caps used in a tape deck as a motor starting cap where the motor runs forward and reverse.[/quote]
I have some NOS motor starting electrolytics.
I have some NOS motor starting electrolytics.
How did this get so complicated?
> why you would need a polarized vs nonpolarized cap.
If the voltage is always the same polarity, you may use a Polar Electro. (And you love to do so, because they are cheap.)
If the voltage ever reverses, you can NOT (safely) use a Polar Electro. When the voltage is backward, it decays to a pot of salt-water instead of a capacitor. Short-term, it is a lousy short. Long-term, it can explode (fun!) if the available power is large enough.
When you do discrete transistors, with +18V at a collector and +3V at the next base, a polar coupling-cap works "fine".
When you do +/- power supplies so "all your ins and outs are at zero V DC", and decide you need to block DC or subsonics, a Polar will work poorly, may fail. (And yet, many large polars tolerate years with a few tenths volt reversed across them.)
Non-polar Electrolytic is essentially two Polar back-to-back. They used to be literally two caps in one cardboard. I gather that now they are two foils in one can. Polars will work in reverse voltage and in alternating voltage applications, if you don't need a "perfect" cap. They are great for lo-price speaker crossovers. They can be used for motor-start, though it took decades to get them tough enough to displace film caps in HRRRRnnnngggg.... application.
Any way you slice them, electrolytics are great caps for the money but not particularly good caps. Electronics would be very different if we had to use paper-caps and chokes to make smooth DC power; OTOH indiscriminate use of cheap electros everywhere throughout the audo path is part of early "transistor sound".
> why you would need a polarized vs nonpolarized cap.
If the voltage is always the same polarity, you may use a Polar Electro. (And you love to do so, because they are cheap.)
If the voltage ever reverses, you can NOT (safely) use a Polar Electro. When the voltage is backward, it decays to a pot of salt-water instead of a capacitor. Short-term, it is a lousy short. Long-term, it can explode (fun!) if the available power is large enough.
When you do discrete transistors, with +18V at a collector and +3V at the next base, a polar coupling-cap works "fine".
When you do +/- power supplies so "all your ins and outs are at zero V DC", and decide you need to block DC or subsonics, a Polar will work poorly, may fail. (And yet, many large polars tolerate years with a few tenths volt reversed across them.)
Non-polar Electrolytic is essentially two Polar back-to-back. They used to be literally two caps in one cardboard. I gather that now they are two foils in one can. Polars will work in reverse voltage and in alternating voltage applications, if you don't need a "perfect" cap. They are great for lo-price speaker crossovers. They can be used for motor-start, though it took decades to get them tough enough to displace film caps in HRRRRnnnngggg.... application.
Any way you slice them, electrolytics are great caps for the money but not particularly good caps. Electronics would be very different if we had to use paper-caps and chokes to make smooth DC power; OTOH indiscriminate use of cheap electros everywhere throughout the audo path is part of early "transistor sound".
NewYorkDave
Well-known member
That brings me to the corollary of what I was saying earlier: we'd love to use film caps everywhere but sometimes it's just not practical. I mean, have you seen how huge/expensive a 22uF film cap can be?
PRR's example of a polarized coupling cap in an op-amp circuit with bipolar supplies does raise an interesting question. Should we worry as much as we tend to do about input offset voltages? A little DC offset (of the proper polarity) at the op-amp's output might not be so bad after all--at least as far as the health of a polarized electro coupling cap goes.
Good to see you, PRR.
PRR's example of a polarized coupling cap in an op-amp circuit with bipolar supplies does raise an interesting question. Should we worry as much as we tend to do about input offset voltages? A little DC offset (of the proper polarity) at the op-amp's output might not be so bad after all--at least as far as the health of a polarized electro coupling cap goes.
Good to see you, PRR.
Here is a filter cap.
A bipolar would be the same, only with back to back polarity.
Which reminds me, did we ever figure out if it matters which side is out?
Probably a foil orientation thing.
Funny. Look at the histogram for that org on red pic.
A little noise, then a big spike.
I wonder if that affects the file size?
A bipolar would be the same, only with back to back polarity.
Which reminds me, did we ever figure out if it matters which side is out?
Probably a foil orientation thing.
Funny. Look at the histogram for that org on red pic.
A little noise, then a big spike.
I wonder if that affects the file size?
...so this is ok?
(...plus diodes between the reg's IN/OUT, not shown)
Why you ask?...
........I have a big bag of them. :wink:
=FB=
(...plus diodes between the reg's IN/OUT, not shown)
Why you ask?...
........I have a big bag of them. :wink:
=FB=
buschfsu
Well-known member
thanks guys. glad i asked
> ...so this is ok?
Costs more than you have to spend. Will be bigger than the equivalent polar cap.
> ........I have a big bag of them.
Economics goes out the window. If you can't sell them, or keep sitting on them, use them everywhere. At least you don't have to figure out which end is which.
Costs more than you have to spend. Will be bigger than the equivalent polar cap.
> ........I have a big bag of them.
Economics goes out the window. If you can't sell them, or keep sitting on them, use them everywhere. At least you don't have to figure out which end is which.
clintrubber
Well-known member
Hello,
Let me re-phrase the somewhat hidden question in my previous message:
Assume situations that present changing polarities to the cap.
If board-space, component-count and eventual differences in cost are all of no concern,
why would we want to use a non-polar cap over two-polars-anti-series-with-a-biased-center-point (by-means-of-high-valued-R-to-supply-rail) ?
Thanks,
Peter
Let me re-phrase the somewhat hidden question in my previous message:
Assume situations that present changing polarities to the cap.
If board-space, component-count and eventual differences in cost are all of no concern,
why would we want to use a non-polar cap over two-polars-anti-series-with-a-biased-center-point (by-means-of-high-valued-R-to-supply-rail) ?
Thanks,
Peter
Harpo
Well-known member
[quote author="clintrubber"]over two-polars-anti-series-with-a-biased-center-point (by-means-of-high-valued-R-to-supply-rail) ?[/quote]
Hi Peter,
is there an advantage of your anti-series-polarised junction biased by a resistor to supply-rail adding another hpf section over leaving this junction floating, wich will require the double cap-value as these are wired in series. Maybe I've mixed up something.
Thanks,
Harpo
Hi Peter,
is there an advantage of your anti-series-polarised junction biased by a resistor to supply-rail adding another hpf section over leaving this junction floating, wich will require the double cap-value as these are wired in series. Maybe I've mixed up something.
Thanks,
Harpo
Most of you probably knows this allready but.....
The problem only arrises (IMHO) when we play with amplifiers which has both its input and output at half the supply voltage. At the same potential.
One way to avoid large expensive bipolarised caps is to tie the feedback string through a polarised cap (if youre using a single ended supply) to ground. If youre using a bipolar +- standard opamp supply you could tie it to DC+-.
This way we have half the supply voltage present at the cap terminal and you get away with cheap polarised electros.
0V ref. outputs driving heavy loads is best avoided. Better to use a SE supply and a big cheap electro.
(It has always been somewhat of a mystery to me, why people bother with bipolar supplys in audio circuits at all)? :?
No caps in the signal-path? Guess again Mr.
The signal path extends all the way back to the power trafo.
clintrubber
Well-known member
[quote author="Harpo"][quote author="clintrubber"]over two-polars-anti-series-with-a-biased-center-point (by-means-of-high-valued-R-to-supply-rail) ?[/quote]
Hi Peter,
is there an advantage of your anti-series-polarised junction biased by a resistor to supply-rail adding another hpf section over leaving this junction floating, wich will require the double cap-value as these are wired in series. Maybe I've mixed up something.
Thanks,
Harpo[/quote]
As I understood it (from the Walt Jung-articles IIRIC) it gives less distortion; both caps are always guaranteed to be properly biased.
The resistor-value can (& must) be high, so in practice the numbers are such that the added hpf you mentioned isn't a problem (at all).
You're right, you need double the values w.r.t. the non-polar. But that holds for the non-polar as well internally; it's just that it's already been done for you and the resulting value has been printed on the outside.
Bye,
Peter
Hi Peter,
is there an advantage of your anti-series-polarised junction biased by a resistor to supply-rail adding another hpf section over leaving this junction floating, wich will require the double cap-value as these are wired in series. Maybe I've mixed up something.
Thanks,
Harpo[/quote]
As I understood it (from the Walt Jung-articles IIRIC) it gives less distortion; both caps are always guaranteed to be properly biased.
The resistor-value can (& must) be high, so in practice the numbers are such that the added hpf you mentioned isn't a problem (at all).
You're right, you need double the values w.r.t. the non-polar. But that holds for the non-polar as well internally; it's just that it's already been done for you and the resulting value has been printed on the outside.
Bye,
Peter
clintrubber
Well-known member
[quote author="buttachunk"][quote author="clintrubber"]why would we want to use a non-polar cap over two-polars-anti-series-with-a-biased-center-point (by-means-of-high-valued-R-to-supply-rail) ? [/quote]
uh, if you mean 2 polar caps with positive terminals connected, and positive voltage tied to the center...
that would be impractical in many situations where there is positive voltage on either of the negative terminals. that can lead to not only more work and more expense, but a higher risk of component failure and distortion. bipolar caps are simple, don't distort, and are generally not as expensive as 2 polar caps.[/quote]
That high-value resistor for the center-tap is to be tied to the supply-rail, so except when TXs are near that give swings beyond the supply-rails, there can't usually be a voltage that's higher than the center-tap node. (We're assuming caps with decently low leakage)
[quote author="clintrubber"]If board-space, component-count and eventual differences in cost are all of no concern,
why would ...[/quote]
Bye,
Peter
uh, if you mean 2 polar caps with positive terminals connected, and positive voltage tied to the center...
that would be impractical in many situations where there is positive voltage on either of the negative terminals. that can lead to not only more work and more expense, but a higher risk of component failure and distortion. bipolar caps are simple, don't distort, and are generally not as expensive as 2 polar caps.[/quote]
That high-value resistor for the center-tap is to be tied to the supply-rail, so except when TXs are near that give swings beyond the supply-rails, there can't usually be a voltage that's higher than the center-tap node. (We're assuming caps with decently low leakage)
Sure, could be of course, but I was wondering which one would be best, given that those other concerns didn't matter, as in the original question:
[quote author="clintrubber"]If board-space, component-count and eventual differences in cost are all of no concern,
why would ...[/quote]
Bye,
Peter
mikep
Well-known member
one thing about using 2 back-to-back caps is you could specify 2 different parts to save a little money. for instance, the cap that connects to the outside world could be a 63V part (plus side out) so that a line output stage could withstand a bad patch into a mic input with phantom present. IMO, an important feature for broadcast audio equip.
mike p
mike p
Similar threads
