I’m about to buy parts for my first tube amp.
I’m told the coupling cap is the most important cap. I read two ‘top of the pops’ comments here for V-caps, by Trombone and serengetiplains (one of whom has tried most types).
It was suggested to me to get MutiCap film & foil RTX polystyrenes “should be 95% as good” and about $80 US the pair less.
Would correct extrapolations of the above suggestion be:
A *completed amp* *with RTX polystyrene coupling caps would be about 95% as good as one with V-caps
And therefore that a *completed amp* with V-caps coupling caps would sound about *5% better than one with RTX polystyrene?
Thanks
I’m told the coupling cap is the most important cap. I read two ‘top of the pops’ comments here for V-caps, by Trombone and serengetiplains (one of whom has tried most types).
It was suggested to me to get MutiCap film & foil RTX polystyrenes “should be 95% as good” and about $80 US the pair less.
Would correct extrapolations of the above suggestion be:
A *completed amp* *with RTX polystyrene coupling caps would be about 95% as good as one with V-caps
And therefore that a *completed amp* with V-caps coupling caps would sound about *5% better than one with RTX polystyrene?
Thanks
I feel building my first with premium parts was a mistake. Now I have version 4 of a different circuit and a pile of $4 resistors with short leads. My recommendation for a first cut is to not go crazy on parts quality. If you're comfortable with Ebay maybe some Russian PIOs for coupling plus non-magnetic generic carbon comp resistors, $10 ASC poly in oil (or Blackgates, still nowhere near $80) for cathode bypasses. Leave the leads long to start. Put the real money into the iron (my second mistake) and save the final voicing for when you're absolutely, positively certain it's the amp you want.
As I mentioned already privately, what's best for someone else may not be the best for you and your system. In fact, I'd say it's likely you wont like the same thing as someone else - no matter how many caps they tried.
I suggested the RTX polystyrenes as good candidates because they aren't particularly expensive in the value you require and lots of people seem to like them even if they're rarely judged "the best" - whatever that means. I don't really care which you choose, just don't spend a lot on caps right now. That's my advice.
And, man, you take everything far too literally... What does "sound about 5% better" mean exactly? I used 95% to talk you out of spending the money because I KNOW you are much more likely to get a sound YOU LIKE by trying several different caps than by trying only one cap no matter how much you spend on it.
When I said "most important" I didn't mean "spend heaps of money". I meant "listen to a lot of different ones and pick the one you like best." You're not building this amp for serengetiplains to listen to are you?
I suggested the RTX polystyrenes as good candidates because they aren't particularly expensive in the value you require and lots of people seem to like them even if they're rarely judged "the best" - whatever that means. I don't really care which you choose, just don't spend a lot on caps right now. That's my advice.
And, man, you take everything far too literally... What does "sound about 5% better" mean exactly? I used 95% to talk you out of spending the money because I KNOW you are much more likely to get a sound YOU LIKE by trying several different caps than by trying only one cap no matter how much you spend on it.
When I said "most important" I didn't mean "spend heaps of money". I meant "listen to a lot of different ones and pick the one you like best." You're not building this amp for serengetiplains to listen to are you?
One place where you might not think of spending money on a cap -- especially if an SE -- is the last cap in the power supply. This is often ignored and an elco is used, but since it is in the current loop it is in the signal path.
I do recommend starting with an elso and then swapping in a Solen (or somesuch)... we did that a couple times, and don't bother with anything but poly there.
dave
I do recommend starting with an elso and then swapping in a Solen (or somesuch)... we did that a couple times, and don't bother with anything but poly there.
dave
The "sound" of a cap has a lot to do with the circuit that it is in. If the cap is working into a high impedance load then the differences between caps is minimal. If the cap sees a low impedance, or time varying load (the usual case) then there could be an audible difference between the various capacitors.
As stated above, the output transformer will have the greatest effect of the sound quality by far. You are not likely to change it once the amplifier is built, so get a good one from the start. Caps can be changed easily, so you have two choices. Put a cheap cap in the amp to get everything working, and plan on changing it. Or (my preference) put in a good, but not outrageous cap (I use Auricaps) and change it later if you feel the need. Leave the leads long enough to use it over on the next amp.
As also stated above, in a SE amp the last power supply cap is very important. Leave room in your design for a upgrade later. Polypropelene motor RUN (not start) caps are relatively cheap possibilities. An electrolytic bypassed with a Solen also works.
As stated above, the output transformer will have the greatest effect of the sound quality by far. You are not likely to change it once the amplifier is built, so get a good one from the start. Caps can be changed easily, so you have two choices. Put a cheap cap in the amp to get everything working, and plan on changing it. Or (my preference) put in a good, but not outrageous cap (I use Auricaps) and change it later if you feel the need. Leave the leads long enough to use it over on the next amp.
As also stated above, in a SE amp the last power supply cap is very important. Leave room in your design for a upgrade later. Polypropelene motor RUN (not start) caps are relatively cheap possibilities. An electrolytic bypassed with a Solen also works.
Thanks tubelab, excellent tips.
How do I calculate of the schematic the impedance load it is working into (it’s the Bugle 45), or a time varying load?
The last cap in the 330 v power supply is marked in Gordon’s schematic marked “SCR”. I was thinking of BG VK (350 v), but it seems poly Solen may be better.
Is SCR = Solen?
Thanks guys!
How do I calculate of the schematic the impedance load it is working into (it’s the Bugle 45), or a time varying load?
The last cap in the 330 v power supply is marked in Gordon’s schematic marked “SCR”. I was thinking of BG VK (350 v), but it seems poly Solen may be better.
Is SCR = Solen?
Thanks guys!
Under small signal conditions the cap sees the 470K resistor in parallel with the input capacitance of the 45. Both of these are constant and predictable. The cap sees two unpredictable, time variant impedances. The first is the miller capacitance of the 45. In the ideal world the signal that the cap sees is a replica of the input signal that is 180 degrees out of phase with a high pass characteristic. In the real world the output transformer causes a phase shift at the frequency extremes especially with a speaker attached. The second effect comes in to play as the signal level is increased, especially with non-sinusoidal (real music) signals. As the instantaneous grid voltage approaches zero the grid impedance begins to drop. At a slightly negative voltage (-2or-3 volts) it drops rapidly, and is a few K ohms at zero volts. Under strong transient conditions this effect will upset the charge on the coupling cap causing blocking distortion. The recovery from this is determined by the size (value) of the cap, and its dissipation factor. This happens more often than you think on highly dynamic music.
There are probably more effects that I can't identify. The effects of these two can be seen by scoping the amp input with one trace, and the grid of the 45 with the other trace. Then drive the amp with a pulse wave of a small duty cycle. You can see the grid do funny things as the amplitude, frequency and pulse width are varied. The effects will be different between a cheap mylar cap and a good cap, although it is hard to see the difference between good caps.
How would you calculate these effects? I have no idea. Even if you could, how would you apply this information to the selection of a cap.
I have never used BG caps because of the cost, but there are probably many different opinions on caps. I use the Panasonic electrolytics paralleled by a PIO or a Solen.
I have a friend who built the Bugle 45 with custom Electra-Print transformers, and an external Heathkit adjustable power supply. It sounded so good that I ordered a pair of these transformers for my 45 amp.
There are probably more effects that I can't identify. The effects of these two can be seen by scoping the amp input with one trace, and the grid of the 45 with the other trace. Then drive the amp with a pulse wave of a small duty cycle. You can see the grid do funny things as the amplitude, frequency and pulse width are varied. The effects will be different between a cheap mylar cap and a good cap, although it is hard to see the difference between good caps.
How would you calculate these effects? I have no idea. Even if you could, how would you apply this information to the selection of a cap.
I have never used BG caps because of the cost, but there are probably many different opinions on caps. I use the Panasonic electrolytics paralleled by a PIO or a Solen.
I have a friend who built the Bugle 45 with custom Electra-Print transformers, and an external Heathkit adjustable power supply. It sounded so good that I ordered a pair of these transformers for my 45 amp.
Tubelab
Thanks again for some more understanding on the impedances and caps front.
What is the advantage in paralleling electrolytics with a PIO or a Solen?
You friend’s Bugle 45 – the Electra-Prints were OPTs? And why have an *adjustable power supply?
(BTW, my amp after that will be a low - medium volt 845, for the bass, I’ll have to check out your boards etc then) plenty of room...
dave *plenty of room...
Cheers
Thanks again for some more understanding on the impedances and caps front.
What is the advantage in paralleling electrolytics with a PIO or a Solen?
You friend’s Bugle 45 – the Electra-Prints were OPTs? And why have an *adjustable power supply?
(BTW, my amp after that will be a low - medium volt 845, for the bass, I’ll have to check out your boards etc then) plenty of room...
dave *plenty of room...
Cheers
Most electrolytics (and most other capacitors)are made by winding the aluminum/ electrolyte/spacer sandwich up into a roll and stuffing it into a can. This gives the capacitor some internal inductance. There is already some internal resistance, and a few other imperfections in all capacitors. In an electrolytic cap these imperfections tend to manifest themselves in the upper audio frequency ranges.
The first function of a power supply capacitor is to supply the peak current that the amplifier needs and to provide a low impedance path to ground for the line frequency ripple current. Electrolytics do this well, although some are better than others.
The second function of a power supply capacitor is to provide a low (and uniform) power supply output impedance for all audio frequencies. This is rarticularly important in a SE amp. Many electrolytic caps exhibit a relatively high internal impedance at high audio frequencies, so there has been a general practice of paralleling the electrolytic with a capacitor that has good characteristics in this range such as a 1uF PIO or polypropelyne.
More recent findings claim that this cap can resonate with the internal inductance of the electrolytic to cause problems elsewhere in the audio band. The proponents of this theory say that you should put a 10 ohm resistor in series with the 1uF cap. These are generalizations which probably have merit for some specific cases. I am sure that each particular combination of capacitors will have their own unique characteristics, and only careful testing can find a good combination that exhibits low impedance across the entire audio range.
I have seen some of this testing explained in an article in Audio Xpress magazine and on the web. I plan to do my own testing when I get the time.
My friends 45 amp uses a variable power supply because he used a Heathkit power supply for testing of the initial design, and never got around to building the final power supply. I belong to the theory that you should eliminate as much risk as possible when building something new. When I am designing an amp I will use a variable lab type power supply (I have several) to get the amp working. Then I can find the voltage and current settings that the amp is happy with. Then I will design and build the proper power supply. The power supply is not trivial in an SE amp since it is directly in the signal path. I have noticed that these old vacuum tube regulated power supplies have a low output impedance from DC to tens of KHz and they sound really good too.
The Electra-Prints were the output transformers and they were custom built for his amp. My friend told Jack that he was building a Bugle 45 and only needed an 8 ohm output. Jack stated that single impedance transformers were better and cheaper than those with multiple speaker windings. They worked well, so I got a pair for my TubelabSE 45 amp. Jack knows these as the Albert Pagan design. They were $100 each a couple of years ago. I am sure that the price has gone up.
The first function of a power supply capacitor is to supply the peak current that the amplifier needs and to provide a low impedance path to ground for the line frequency ripple current. Electrolytics do this well, although some are better than others.
The second function of a power supply capacitor is to provide a low (and uniform) power supply output impedance for all audio frequencies. This is rarticularly important in a SE amp. Many electrolytic caps exhibit a relatively high internal impedance at high audio frequencies, so there has been a general practice of paralleling the electrolytic with a capacitor that has good characteristics in this range such as a 1uF PIO or polypropelyne.
More recent findings claim that this cap can resonate with the internal inductance of the electrolytic to cause problems elsewhere in the audio band. The proponents of this theory say that you should put a 10 ohm resistor in series with the 1uF cap. These are generalizations which probably have merit for some specific cases. I am sure that each particular combination of capacitors will have their own unique characteristics, and only careful testing can find a good combination that exhibits low impedance across the entire audio range.
I have seen some of this testing explained in an article in Audio Xpress magazine and on the web. I plan to do my own testing when I get the time.
My friends 45 amp uses a variable power supply because he used a Heathkit power supply for testing of the initial design, and never got around to building the final power supply. I belong to the theory that you should eliminate as much risk as possible when building something new. When I am designing an amp I will use a variable lab type power supply (I have several) to get the amp working. Then I can find the voltage and current settings that the amp is happy with. Then I will design and build the proper power supply. The power supply is not trivial in an SE amp since it is directly in the signal path. I have noticed that these old vacuum tube regulated power supplies have a low output impedance from DC to tens of KHz and they sound really good too.
The Electra-Prints were the output transformers and they were custom built for his amp. My friend told Jack that he was building a Bugle 45 and only needed an 8 ohm output. Jack stated that single impedance transformers were better and cheaper than those with multiple speaker windings. They worked well, so I got a pair for my TubelabSE 45 amp. Jack knows these as the Albert Pagan design. They were $100 each a couple of years ago. I am sure that the price has gone up.
Just about all said when I arrived at this thread, so also just to reassure:
There was a series of articles by Cyril Bateman in Electronics World about 4 years ago. He did just about all the tests possible on capacitors from a few pF to large electrolytics; apparently with over 180 spectrum analyses. Obviously not all of these were published, but quite a few myths were exposed. I "mislayed" my copies and so cannot give a direct reference (I will have to go back to the library and recopy one of these days.)
For coupling capacitors he found none of the exotic (and pricey) stuff to really do practically better than a polyester. For electrolytics I recall that the non-polarised types were superior where available, also with due respect, the Black Gates were marginally better than good other types but not worth the extra expenditure according to him. I judged this a really good series and am satisfied that just about all that is needed was mentioned there. As soon as I can get hold of the title and issue ......
Then also just to support, it is strange that coupling caps are so often revered as "directly in the signal path". So are power supply capacitors, if one cares to draw a.c. equivalent diagrams! Advice by Tubelab suffices - this to my mind is often an overrated subject. Effects that are 80 - 100 dB down, when the impedances as coupling capacitors are a % or so only of the full circuit impedance to begin with......
I have also read about the possibility of resonance with the electrolytic's inductance but could never find anything except very small indications of that, well above the audio range (as in several 100 KHz). This would act in series with the capacitance, and if that is as lossy as purported at those frequencies, then such resonance effect would be damped anyway.
I did read that modern electrolytics had the two "electrodes" area-soldered at either end, thus shorting the "windings" and any inductance. I have not recently taken one apart; can anybody comment on this? (Tubelab?)
Regards.
There was a series of articles by Cyril Bateman in Electronics World about 4 years ago. He did just about all the tests possible on capacitors from a few pF to large electrolytics; apparently with over 180 spectrum analyses. Obviously not all of these were published, but quite a few myths were exposed. I "mislayed" my copies and so cannot give a direct reference
(I will have to go back to the library and recopy one of these days.)For coupling capacitors he found none of the exotic (and pricey) stuff to really do practically better than a polyester. For electrolytics I recall that the non-polarised types were superior where available, also with due respect, the Black Gates were marginally better than good other types but not worth the extra expenditure according to him. I judged this a really good series and am satisfied that just about all that is needed was mentioned there. As soon as I can get hold of the title and issue ......
Then also just to support, it is strange that coupling caps are so often revered as "directly in the signal path". So are power supply capacitors, if one cares to draw a.c. equivalent diagrams! Advice by Tubelab suffices - this to my mind is often an overrated subject. Effects that are 80 - 100 dB down, when the impedances as coupling capacitors are a % or so only of the full circuit impedance to begin with......
I have also read about the possibility of resonance with the electrolytic's inductance but could never find anything except very small indications of that, well above the audio range (as in several 100 KHz). This would act in series with the capacitance, and if that is as lossy as purported at those frequencies, then such resonance effect would be damped anyway.
I did read that modern electrolytics had the two "electrodes" area-soldered at either end, thus shorting the "windings" and any inductance. I have not recently taken one apart; can anybody comment on this? (Tubelab?)
Regards.
I have not taken apart any of the newer good quality electrolytics. I have a problem taking apart perfectly good parts, and I haven't blown any of them up yet.
I can hear the difference between different power supplies on the same SE amp and I don't claim to have good hearing. The sound that flows from a good SE amp when powered by a good tube regulated supply is hard to beat. I use my Knight or the Fluke 407D to develop an amp, then I try to come up with a power supply that sounds as good.
I have tried several types of coupling caps in some of my amps. In an amp with a fixed (predominantly resistive) load like the TubelabSE, I can only notice the sound of really cheap caps (10 for a dollar kind of cheap). The rest all sound very similar. In an amp that has the cap connected directly to the grid of output tube, the differences are more noticable. I believe that I like the sound of the Auricaps better, but I am not totally convinced. I have not tried anything more expensive than those.
When both channels in a stereo SE amp are powered from the same power supply, there can be coupling between the channels through the power supply. Put a load resistor on one channel and a speaker on the other channel. Drive the channel with the load to full power with a real music signal. Short the input of the channel with a speaker. Ideally there should be no sound from the speaker. Sound from the output transformer is normal. If you hear anything from the speaker, there is some coupling from one channel to the other.
You can try that polyprop or oil cap in parallel with the last power supply cap. I find that it helps on most SE amps. One lead should be connected to the B+ ends of the output transformer wires. The other end should be connected to the point where the output tubes cathodes are returned to ground.
If your amp uses cathode bias, some will use two such caps connecting the negative ends directly to the output tubes cathodes. This is often called an "ultrapath" connection.
I can hear the difference between different power supplies on the same SE amp and I don't claim to have good hearing. The sound that flows from a good SE amp when powered by a good tube regulated supply is hard to beat. I use my Knight or the Fluke 407D to develop an amp, then I try to come up with a power supply that sounds as good.
I have tried several types of coupling caps in some of my amps. In an amp with a fixed (predominantly resistive) load like the TubelabSE, I can only notice the sound of really cheap caps (10 for a dollar kind of cheap). The rest all sound very similar. In an amp that has the cap connected directly to the grid of output tube, the differences are more noticable. I believe that I like the sound of the Auricaps better, but I am not totally convinced. I have not tried anything more expensive than those.
When both channels in a stereo SE amp are powered from the same power supply, there can be coupling between the channels through the power supply. Put a load resistor on one channel and a speaker on the other channel. Drive the channel with the load to full power with a real music signal. Short the input of the channel with a speaker. Ideally there should be no sound from the speaker. Sound from the output transformer is normal. If you hear anything from the speaker, there is some coupling from one channel to the other.
You can try that polyprop or oil cap in parallel with the last power supply cap. I find that it helps on most SE amps. One lead should be connected to the B+ ends of the output transformer wires. The other end should be connected to the point where the output tubes cathodes are returned to ground.
If your amp uses cathode bias, some will use two such caps connecting the negative ends directly to the output tubes cathodes. This is often called an "ultrapath" connection.
Thanks Johan . .
Rod Elliott’s analysis: http://sound.westhost.com/articles/capacitors.htm includes a ‘synopsis’ of Bateman’s “Capacitor Sound” - Jul, Sep, Oct- Dec 2002 & Jan 03 Electronics World.
Tubelab
Thanks again (my interpretation is that the Bugle at least, has the cap connected directly to the grid of output tube. CB suggests polyester, but I am leaning to MIT RTX film & foil polystyrene.
Cheers
Rod Elliott’s analysis: http://sound.westhost.com/articles/capacitors.htm includes a ‘synopsis’ of Bateman’s “Capacitor Sound” - Jul, Sep, Oct- Dec 2002 & Jan 03 Electronics World.
Tubelab
Thanks again (my interpretation is that the Bugle at least, has the cap connected directly to the grid of output tube. CB suggests polyester, but I am leaning to MIT RTX film & foil polystyrene.
Cheers
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.