In the "Simplistic NJFET RIAA" thread, I suggested output coupling caps were not required if the preamp which you were feeding the signal to, has an input coupling cap.
One poster came back with what I thought was an interesting point - "If the caps are before the volume pot, maybe yes, but if they are not, then it's not so good to have DC running through the pot."
Now at the time, that seemed a very sensible statement but, after musing about this for a day, I have come up with a couple of questions that I would like answers to. The reason is, I have found out that my preamp does have its volume control before the input coupling caps and if I have to go with caps then:
A. 4 of them will cost me serious money (it's a double phono stage), and
B. I will have difficulty fitting them into the case.
So I decided to make this post; appreciate whatever knowledge people have to offer.
So:
1. Why is it a problem to have a little DC current continuously running through the volume control?
2. Does it make any difference if the volume control is a stepped attenuator rather than a pot?
Thanks,
Andy
One poster came back with what I thought was an interesting point - "If the caps are before the volume pot, maybe yes, but if they are not, then it's not so good to have DC running through the pot."
Now at the time, that seemed a very sensible statement but, after musing about this for a day, I have come up with a couple of questions that I would like answers to. The reason is, I have found out that my preamp does have its volume control before the input coupling caps and if I have to go with caps then:
A. 4 of them will cost me serious money (it's a double phono stage), and
B. I will have difficulty fitting them into the case.
So I decided to make this post; appreciate whatever knowledge people have to offer.
So:
1. Why is it a problem to have a little DC current continuously running through the volume control?
2. Does it make any difference if the volume control is a stepped attenuator rather than a pot?
Thanks,
Andy
A1. Because current through a resistance generates a voltage drop, and being a 3-terminal variable resistance exacerbates the effect.
A2. Only the obvious differences, eg. continuous vs. stepped variability.
I'm not familiar with the thread/project details. If there is insignificant DC into the preamp (ie, volume pot) this is all moot. I've seen other circuits with the volume pot sandwiched between two caps to block DC.
A2. Only the obvious differences, eg. continuous vs. stepped variability.
I'm not familiar with the thread/project details. If there is insignificant DC into the preamp (ie, volume pot) this is all moot. I've seen other circuits with the volume pot sandwiched between two caps to block DC.
I think the current going through the resistive element itself is "not good," as it will change the level on the wiper as the volume is changed, and if there is any "scratchiness" in the contact at all, the DC will make it sound worse as the volume is changed. This appears to be the "problem" you have.
But the worst problem, the one that such warnings really appear to be about, is DC current through the wiper (caused by no caps between the volume control and the following circuitry, and current from that circuitry going back into the control through the wiper). This causes bad things at the contact point, reduction in the life of the potentiometer, and I read on Facebook that it causes the spread of social disease. I'd like to see a "potenitometer expert" pipe in here, but I'm thinking the wiper, to do its best job, should be connected to a high impedance input, so there is as little current as possible, whether AC or DC, going through the contact point.
A stepped attentuator uses real switches, won't have the problem of reduced life and for that reason it won't matter if the current comes from the "input" side or the "wiper" side, but it will click when switching between volume levels. The loudness of the clicks depends on the amount of DC voltage, and will be most noticeable with quiet passages or no signal.
But the worst problem, the one that such warnings really appear to be about, is DC current through the wiper (caused by no caps between the volume control and the following circuitry, and current from that circuitry going back into the control through the wiper). This causes bad things at the contact point, reduction in the life of the potentiometer, and I read on Facebook that it causes the spread of social disease. I'd like to see a "potenitometer expert" pipe in here, but I'm thinking the wiper, to do its best job, should be connected to a high impedance input, so there is as little current as possible, whether AC or DC, going through the contact point.
A stepped attentuator uses real switches, won't have the problem of reduced life and for that reason it won't matter if the current comes from the "input" side or the "wiper" side, but it will click when switching between volume levels. The loudness of the clicks depends on the amount of DC voltage, and will be most noticeable with quiet passages or no signal.
Everything Ben has written is correct. The bottom line is you need the cap.
If you are having trouble with the size and cost of them, it would actually be fine to use good quality electrolytic or tantalum, as there is actually a DC bias present (looking at the schematic on the first page of the thread) so the caps will work well.
If you are having trouble with the size and cost of them, it would actually be fine to use good quality electrolytic or tantalum, as there is actually a DC bias present (looking at the schematic on the first page of the thread) so the caps will work well.
Last edited:
Jeez, I hate you guys! I post a reasonable question, hoping that someone will say "You'll be right, mate." ... and all you do is, unanimously, tell me I need to buy some caps! 😀
Haha - OK, thank you. 🙂 My idea that I could avoid paying A$1,000 for 4 output caps (my phono stage is a dual, so 4 modules = 4 caps) has been shown to be naïve.
So out comes the wallet! 🙂
Regards,
Andy
PS: No, sorry, richie00boy - good quality electrolytic or tantalum is not acceptable, IMO. I built one phono stage with AmpOhm PF-XAL-ALs ($30 each) and another with SoniCap Platinums (~$250 each) - the latter sounded much better. But someone who compared it against a phono stage that he had installed Duelunds in ... and suggested I needed Duelunds! ($500 each!) 🙁
Haha - OK, thank you. 🙂 My idea that I could avoid paying A$1,000 for 4 output caps (my phono stage is a dual, so 4 modules = 4 caps) has been shown to be naïve.
So out comes the wallet! 🙂
Regards,
Andy
PS: No, sorry, richie00boy - good quality electrolytic or tantalum is not acceptable, IMO. I built one phono stage with AmpOhm PF-XAL-ALs ($30 each) and another with SoniCap Platinums (~$250 each) - the latter sounded much better. But someone who compared it against a phono stage that he had installed Duelunds in ... and suggested I needed Duelunds! ($500 each!) 🙁
Last edited:
But was that phono stage a single rail design with bias on the caps? And was the phono stage the guy had with those 500 dollar caps in the same design and layout as yours?
Yes, richie00boy - my phono stage is single-rail design.
But I can't tell you how my mate's Vitus is arranged.
Are you suggesting that if it's got '+'/'-' DC rails, the output cap won't be biased? In which case, I can see how (while lesser caps certainly sound better when they are DC-biased), you need a much better cap to sound good when you have a '+'/'-' DC rail.
In which case, I'm confident I don't need to stump up the extra $ for the Duelunds. 🙂
Regards,
Andy
Yes, when a split supply is used there is no bias, so polarised caps like electrolytics and tantalums don't perform anywhere near what they are capable of.
As your mate had a completely different unit to yours, it's pointless to say that the 500 dollar caps in his have any bearing on the difference in sound.
My recommendation is that you fit a good quality tantalum cap to your unit and see how it fares, before spending money on expensive caps. You never know, the difference might be tiny or even undetectable.
As your mate had a completely different unit to yours, it's pointless to say that the 500 dollar caps in his have any bearing on the difference in sound.
My recommendation is that you fit a good quality tantalum cap to your unit and see how it fares, before spending money on expensive caps. You never know, the difference might be tiny or even undetectable.
Last edited:
Thank you very much, richie00boy, for:
a) explaining the implications of a '+'/'-' PS, and
b) pointing out that what caps are good in a completely different circuit may not deliver any benefit at all in my circuit (which is single-ended).
So I will not think any more about the Duelunds but, sorry, I have heard the result of using SoniCap Platinums in my phono stage and it sounded better than the one when I used AmpOhms. So that is what I will use. 🙂
Regards,
Andy
My understanding is that polarised caps need some DC bias voltage, otherwise they will deteriorate.
Non-polarised caps, such the various plastic/paper dielectrics, are likely to work better (i.e. produce less distortion) with no DC bias. Adding a DC bias means that any dielectric non-linearity (which in most cases will be very very small anyway - except for polyester where it is just very small) will be magnified but reduced in order. Hence a cap which produces a very tiny amount of 3rd without bias will instead produce a bigger amount of 2nd with bias. This assumes that the bias is larger than the peak signal - which will often be the case. Those who prefer their caps with bias may, just, be hearing this effect.
Non-polarised caps, such the various plastic/paper dielectrics, are likely to work better (i.e. produce less distortion) with no DC bias. Adding a DC bias means that any dielectric non-linearity (which in most cases will be very very small anyway - except for polyester where it is just very small) will be magnified but reduced in order. Hence a cap which produces a very tiny amount of 3rd without bias will instead produce a bigger amount of 2nd with bias. This assumes that the bias is larger than the peak signal - which will often be the case. Those who prefer their caps with bias may, just, be hearing this effect.
That is not the view of the amplifier designer whose amps I use. He says, putting quite a considerable bias on a film cap (like a coupling cap):
a) improves its sound, and
b) minimises the "sonic difference" between more expensive (and recognised "better sounding" caps) and cheaper caps (not so "better sounding").
Regards,
Andy
As I said, people who prefer DC bias on caps are probably hearing the tiny amount of 2nd this will produce and prefer this to the even tinier amount of 3rd without bias. If you look at a graph of dielectric nonlinearity it is almost always symmetric about zero - indeed it has to be unless the dielectric is poled. The dominant effect is 3rd. Add some DC and instead of distortion coming from (signal)^3 you get distortion from DCx(signal)^2 - this means 2nd instead of 3rd, and at a level which is boosted by (DC/signal).
Some people claim a 'zero voltage' effect in caps but as far as I know this has never been demonstrated. There is nothing in dielectrics to cause this. It would be seen in sensitive RF and measurement electronics long before it showed up in audio.
Some people claim a 'zero voltage' effect in caps but as far as I know this has never been demonstrated. There is nothing in dielectrics to cause this. It would be seen in sensitive RF and measurement electronics long before it showed up in audio.
DF96, are you talking solely about the effect of bias on non-polarised caps? Adding a bias to polarised caps always improves their performance.
In the application here, there is a bias present - it's intrinsic to the design and the function of the cap is to block that DC. So let's not get caught up with a red herring 🙂 Although what you are saying could be construed as the polarised cap performing better than the non-polarised in this application?
In the application here, there is a bias present - it's intrinsic to the design and the function of the cap is to block that DC. So let's not get caught up with a red herring 🙂 Although what you are saying could be construed as the polarised cap performing better than the non-polarised in this application?
Andy
the major problem with DC through a pot is wiper noise on movement. This is due to discontinuity of contact as the wiper is forced to lift off the track when it encounters dust or dislodged particles of track material. This was a big problem with carbon composition pots as the carbon is readily dislodged, and the problem got worse as the pots aged since there was more material to dodge.
It is far less of a problem with modern cermet pots and almost non-existent with conductive plastic pots (though this last doesn't stop them sounding like (English batsmen)* to my ears).
Overall advice: suck it and see. It will probably be fine.
Regarding stepped attenuators: if you use a break before make switch, the presence of DC can only make the noise problem inherent to this design worse as it willl promote arcing as the contacts approach.
I can't see that it's a problem with make before break switches, but there may be something I'm missing.
* the software wouldn't allow me to write Shyte so I chose a simile.
the major problem with DC through a pot is wiper noise on movement. This is due to discontinuity of contact as the wiper is forced to lift off the track when it encounters dust or dislodged particles of track material. This was a big problem with carbon composition pots as the carbon is readily dislodged, and the problem got worse as the pots aged since there was more material to dodge.
It is far less of a problem with modern cermet pots and almost non-existent with conductive plastic pots (though this last doesn't stop them sounding like (English batsmen)* to my ears).
Overall advice: suck it and see. It will probably be fine.
Regarding stepped attenuators: if you use a break before make switch, the presence of DC can only make the noise problem inherent to this design worse as it willl promote arcing as the contacts approach.
I can't see that it's a problem with make before break switches, but there may be something I'm missing.
* the software wouldn't allow me to write Shyte so I chose a simile.
Last edited:
No. Generally, non-polarised caps are better.richi00boy said:
Of course most applications of caps have a DC voltage necessarily present. Most dielectrics used in caps are sufficiently linear that this creates no problem. I was simply offering a possible explanation as to why some people prefer a DC bias even when the circuit does not require it. If the dielectric is linear and the cap is mechanically solid then a DC bias does nothing; nothing will be heard. Any lack in these areas could create distortion which a significant DC bias will worsen (in magnitude) but reduce (in order); it could sound better while actually adding more distortion. It will be a small effect in any case, unless a wholly inappropriate dielectric has been used.
All I am saying is that if you genuinely hear a change when adding a DC bias to a non-polar (e.g. plastic film) then it probably comes from adding some 2nd. If this makes it sound more like an expensive cap then I draw a conclusion which some may find unpalatable!
In an ideal world, an amp design with split supply rails would have no DC on the output. In the real world, you get a small DC offset. Often (such as when using opamps) you can't predict whether it will be positive or negative.
Thank you, Mark - I was hoping you would chip in. First thing I will do before making any decision is measure the DC offset. I may be worrying over nothing! 😉
I used a stereo TKD stepped attenuator in my preamp - I have no idea whether it is 'break before make' or 'make before break'?
Haha - good one! 😀
Regards,
Andy
Last edited:
If I understand their construction correctly they are neither but they mimic a make before break switch.
Thanks, Mark. I measured the DCv at the Emitter of the BC550 output buffer and, yes, I can see there is obviously a large "DC reading". 😱 The Base is at about half supply - so 12v - and the Emitter is about 0.6v lower, right?
I thought I might be dealing with a few 10s of mV but for what it is, I will use a cap.
Thanks for all the input here.
Andy
- Status
- Not open for further replies.