I'm looking for good quality output coupling caps to modify my OPPO player. There are endless dicussions and listening tests everywhere, but one thing is still unclear to me. Is there a difference in the type of cap you need for crossover applications and (signal) coupling? Most online tests, like the famous 'Humble homemade hifi' pages, use the caps in crossovers. Would the results transfer to coupling applications, or are the two different?
In general, what would be the requirements for the perfect coupling cap? And how would they be different to the requirements for crossover caps?
In general, what would be the requirements for the perfect coupling cap? And how would they be different to the requirements for crossover caps?
There is a significant difference between coupling and any 'tone control' caps - which includes crossovers, equalisation etc. The difference is the signal voltage. Coupling caps have little signal voltage across them, apart from the cap which sets the LF rolloff. Tone caps have significant signal voltage across them, so you need linear dielectrics and very rigid construction. Almost any cap can do coupling, with a few exceptions.
My fault, I should have been clearer about what I wanted. Thanks for your answer.
Do I understand then that coupling is an easier job for a cap, so the quality of the cap should matter less then when used in a crossover?
Do I understand then that coupling is an easier job for a cap, so the quality of the cap should matter less then when used in a crossover?
Yes, but I rarely see this mentioned outside DIYaudio.
And only recently has this become a regular sight.
D.Self introduced the idea to me nearly 20years ago, but I cannot recall the distortion idea being extended to all coupling caps. As far as I was aware it sort of stayed with the NFB cap for a long time and did not get extended to other uses/duties.
The closest I saw was Rane (or Jensen) talking about a pair of 220uF bipolar electrolytic being used for a balanced line level output.
And only recently has this become a regular sight.
D.Self introduced the idea to me nearly 20years ago, but I cannot recall the distortion idea being extended to all coupling caps. As far as I was aware it sort of stayed with the NFB cap for a long time and did not get extended to other uses/duties.
The closest I saw was Rane (or Jensen) talking about a pair of 220uF bipolar electrolytic being used for a balanced line level output.
Last edited:
For coupling, the idea is to have zero audio voltage across the cap. If there's no audio voltage across the cap, despite any protestations to the contrary, it can't affect the sound.
Naturally it depends on your definition of "zero". There will always be some small voltage across the cap, due to its reactance and dissipation factor. You can reduce these by making the cap value larger and choosing caps with lower dissipation factor, like switching to a giant polypropylene!
The voltage across the cap is an easy thing to measure and I don't know why there's so much disagreement and discussion and so little data. OK, I do, but that becomes a conversation on human nature!
For output coupling caps, technically, the best choices are probably polystyrene, polypropylene and Teflon. They're linear and have very low DF (translates to low esr). The downside is that people often choose too low a value because the needed value won't physically fit or costs an arm and a leg. IMO, this is where a lot of the beliefs about different sound quality of different caps comes from. But I've been wrong before.
IMO again, the best solution for output coupling caps, if you have to use an electrolytic, is to choose the value for near zero audio voltage across the cap, then choose a relatively high voltage rating, say 50 VDC. Even though you may not need the higher voltage, if you look at the dissipation factor tables in the data sheets, you'll see that it goes down with increasing voltage rating.
This can be confirmed on the test bench very easily by driving the cap with a square wave through a couple kohm resistor and looking at the voltage on the cap at a very high gain. The cap will integrate the square wave into a triangle. At each peak there will be a voltage shift that depends on esr. Higher voltage caps will show almost no shift at all, whereas low voltage caps, particularly under 16 VDC, will show a definite shift or offset. They'll also show distortion of the triangle, particularly tantalums. You're really answering the question, "how good an integrator does this cap make", which is a very good test of how perfect a cap is. Note that this test is only useful for electrolytics, as film caps are too good to observe any effect. They need to be tested by other means like sensitive bridge measurements.
A cap used to control filter break points will always have audio voltage across it it order to do its job, be it a low level active filter or a high level speaker crossover. In these cases you need to choose a part with as close to ideal properties as possible.
An electrolytic would typically be your last choice, but for the fact that speaker crossovers sometimes require 50 or more uF at a reasonably high voltage. Film caps just aren't a cost effective solution compared to nonpolar electrolytics. Fortunately, used properly, they sound very good.
Naturally it depends on your definition of "zero". There will always be some small voltage across the cap, due to its reactance and dissipation factor. You can reduce these by making the cap value larger and choosing caps with lower dissipation factor, like switching to a giant polypropylene!
The voltage across the cap is an easy thing to measure and I don't know why there's so much disagreement and discussion and so little data. OK, I do, but that becomes a conversation on human nature!
For output coupling caps, technically, the best choices are probably polystyrene, polypropylene and Teflon. They're linear and have very low DF (translates to low esr). The downside is that people often choose too low a value because the needed value won't physically fit or costs an arm and a leg. IMO, this is where a lot of the beliefs about different sound quality of different caps comes from. But I've been wrong before.
IMO again, the best solution for output coupling caps, if you have to use an electrolytic, is to choose the value for near zero audio voltage across the cap, then choose a relatively high voltage rating, say 50 VDC. Even though you may not need the higher voltage, if you look at the dissipation factor tables in the data sheets, you'll see that it goes down with increasing voltage rating.
This can be confirmed on the test bench very easily by driving the cap with a square wave through a couple kohm resistor and looking at the voltage on the cap at a very high gain. The cap will integrate the square wave into a triangle. At each peak there will be a voltage shift that depends on esr. Higher voltage caps will show almost no shift at all, whereas low voltage caps, particularly under 16 VDC, will show a definite shift or offset. They'll also show distortion of the triangle, particularly tantalums. You're really answering the question, "how good an integrator does this cap make", which is a very good test of how perfect a cap is. Note that this test is only useful for electrolytics, as film caps are too good to observe any effect. They need to be tested by other means like sensitive bridge measurements.
A cap used to control filter break points will always have audio voltage across it it order to do its job, be it a low level active filter or a high level speaker crossover. In these cases you need to choose a part with as close to ideal properties as possible.
An electrolytic would typically be your last choice, but for the fact that speaker crossovers sometimes require 50 or more uF at a reasonably high voltage. Film caps just aren't a cost effective solution compared to nonpolar electrolytics. Fortunately, used properly, they sound very good.
Don't forget the details
Generally speaking Polypropylene is a good choice. But the electrode type/material and the diameter to length ratio is very important as well
Generally speaking Polypropylene is a good choice. But the electrode type/material and the diameter to length ratio is very important as well
Hi,
The cheapest way to upgrade coupling caps is to inspect
all outputs and inputs and remove / bypass output or
input if they both have coupling caps. Generally I'd
remove output caps and optimise input caps.
rgds, sreten.
The cheapest way to upgrade coupling caps is to inspect
all outputs and inputs and remove / bypass output or
input if they both have coupling caps. Generally I'd
remove output caps and optimise input caps.
rgds, sreten.
Yes, Yes, Yes !
Optimise the input cap of the Receiver for best quality of signal passing to suit the INPUT parameters that load the DC blocking capacitor.
Once this is done, just about any Source can drive the Receiver without risk of damage.
The only equipment where further thought should be given would be a Valve/Tube output stage of Source equipment that sends a very high DC voltage out. But this must be rare, very rare, maybe even illegal.
- Status
- Not open for further replies.