[Roger A. Modjeski]

Does the audio signal pass through a coupling cap? 

Why do I hear changes in sound quality between different caps?

The first is a good question, the second I cannot answer.  "Does the audio signal pass through a coupling cap?"  Pass through to me means there is a current. In a crossover there is certainly a substantial current that can be several amperes. Indeed things like ESR make a big difference. Dielectric absorption (DA) also makes a difference as the low frequencies are absolutely pumping the dielectric material up and down with many volts. The louder you play the speaker, the lower the impedance, the larger the effects. Yes a big bass transient leave behind a "ghost voltage"in the tweeter series cap due to DA. Lets not have that. I advise people to change caps in their crossoves without reservation. However, I go further to encourage them to save that money and get and electronic crossover (where the caps pass no current) and go direct to their drivers. Crossovers overboard!  .

Of course, changing the caps is easier. How many are willing to take that big step and make a big difference in their systems. The point about multi-amping is that you can choose specific amps for their specific job. I just hooked the OTL-1 to my horribly inefficient (less than 80 dB/watt) ESL speaker. By design that amp only sees 100 Hz up and a 500 watt transistor amp of no distinction plays 100 Hz down. Even with that inefficient ESL I can play plenty loud. With a conventional speaker of that low efficiency playing full range I would be running out of power in the bass region. 

I recall the whole boutique cap thing started in crossovers and with that I totally agree. Damn the 10 cent electrolytics with their high ESR and loss of capacitance over time. They are usually 10-20% accurate in value so the crossover points and slopes can vary widely. Bypass them but that won't fix the DA problem because the bypass is to small, but it will help the ESR.

Having said all that in my honest opinion as one who does understand how electrons get about in a system, there is virtually no current (a few microamps) at best and no voltage (the cap has little or no reactance at audio frequencies) that the audio signal does not pass through the anymore than it passes through a wire. (I am not inviting a discussion of hook up wire in an amp )

Are you saying that coupling caps in an amp make as big a difference in your hearing as crossover caps?

The best analogy I can think of right now is that high octane gas vs low octane gas will make a big difference in a high compression engine. It will make less difference in a low compression engine and will make no difference, in my experience, in starting a fire. 

[BobM]

How about adding a smaller, quicker bypass cap to those large value electrolytics in the power section of an amp or preamp or CD player or whatever? IYO is there any benefit to that?

[corndog71]

Let me give an example.

I have a Bottlehead Sex SET headphone amplifier.  http://bottlehead.com/s-e-x-2-1-specs/

There are two cap positions in particular where changing caps made audible differences.  The tubes are 6DN7.  There's an interstage cap (0.1uF 630V) and output coupling cap (1uF 630V).  In the original kit an inexpensive poly cap is used for interstage and a Solen for output cap.



At some point I replaced those caps with ERSE brand caps.  I really didn't hear much if any difference.  For the record I have AKG 702 headphones.



At some point I added the Constant Current Source upgrade that Bottlehead sells.  It's a little circuit board; one for each channel.  That made a significant improvement to the sound.

Later, when I had some money to burn I picked up a pair of Clarity Cap ESA and MR series for the output caps.  I also had a pair of 0.47uF V-Caps from another project.  I tried various combinations of interstage and coupling caps and each time the sound changed.  Sometimes for the better but more often for the worse.  With each combination I played and listened for hundreds of hours.





With the V-caps in the interstage position and the Clarity ESA in the coupling cap position.  I got a nice, detailed yet warm on the top sound.  More detail and clarity over the original and ERSE caps.  Swapping the ESA cap for the MR cap was a revelation!  Everything became more detailed.  Separation of instruments and voices became more defined.  Details emerged that had been muted or obscured.  But over time I found it too revealing!  My ears would get fatigued after about 30-45 minutes of listening.  I loved the resolution I was getting but hated the fatigue.  I was disappointed because in my tube preamp the MR caps remained highly resolving yet never fatiguing.

I ended up switching back to the ESA output caps and the warmth returned and I could listen once again with no fatigue.

Recently I acquired a bunch of Russian military caps in various values.  Most of them are paper in oil types sealed in glass with metal outer bodies.  A couple are large teflon caps. (0.1uF 630V)

I hooked up the teflon FT-3 caps as interstage caps and a pair of PIO K42y-2 output caps.  After running them for a while I gave them a listen and now feel they have better resolution than the ESA caps especially in the treble region yet retain the overall warmth that keeps my ears from getting fatigued.

[Roger A. Modjeski]

It is interesting to hear that some boutique caps made it worse some made it better. All in all I guess you ended up spending as much or more on the caps than the original Bottlehead amp. I know the V- caps are very expensive, especially in the large sizes. I bought some for customer at his request. As I said, I don't mind doing it, I don't feel its the best use of money when other more important things can be done in a system.

As to what you heard. The 1 uF output coupling cap is conducting current so that is the one that should make the most difference. The difference in current is more than ten thousand to one vs the interstage cap. The output coupling cap is driving a load of a few thousand ohms, the interstage cap is driving something like 470,000 ohms and at lower 15 times lower signal level (15 is the mu of tube).

I also think the output coupling cap should be a lot bigger. If you are driving a 3000 ohm transformer the -3 dB down point is 53 Hz with 1 uF. This should be a problem, however 32 ohm phones move the -3 point down by 4 to 1 is the inductance of the output transformer is sufficient. Has anyone measured that? I clicked on specs but couldn't find any.

I never use less than 33 uF in my rather similar 6EM7 parafeed amp.  Why is that cap so small? These are the things that I think need attention.
Here is a nice calculator for that sort of thing. http://www.sengpielaudio.com/calculator-XLC.htm

I am not surprised to hear the current source mod made a big difference. That is a major circuit change. In my mind it has much more influence over the sound than a cap change. A current source unloads the input triode reduces distortion and raises gain. Did you notice the gain change?

If you had a variety of output transformers and chokes to try you would hear bigger differences and it's not about the materials. Magnequest promoted some unusual expensive iron some time ago, where is it now? Seems everyone making parts has become a self appointed materials expert.

However who wants to fool around with changing transformers, though yours are pretty easy to get to.  Do you know the primary and choke inductance values. They had better be large. Why dont people fool with transformers more than capacitors. The have a bigger chance of making a difference. I will propose that a tranformerm is too difficult and varied to specify. You have to express the impedance ratios, the input inductance, the leakage inductance, the copper loss, the iron loss, the Low frequency saturation point. A cap is easy to specify, capacity and voltage, done.  The rest is going on and on about exotic materials, winding techniques, leads, yada, yada , yada.

How about changing the 6DN7 to a 6EM7 which has the same pinout but  lower output impedance. You would have to change the cathode resistors but little else I imagine. The output impedance is 750 vs 2000 ohms, almost 3 times lower. That will make a big difference.

I am simply stating that there are bigger things to do that fooling with caps, however, for some reason fooling with caps has become very popular perhaps only one step below tube rolling. I think it is obvious in any hobby that people change the things that are easiest to change, while leaving the more difficult yet more important things alone.

I could make you some iron that would blow the socks off the Bottlehead stuff. Look at the pricing of the kit. Break it down by the parts you know the cost of and you will see that the two output chokes and two output transformers can only cost about $25- $50 each. If you have $250 to spend on a Vcap how about the same for a transformer or audio choke?

How about replacing that $2 volume control with a Noble or Alps.

I really don't understand how more resolution caused listener fatigue. I find the more detail I get the more I can turn down the volume and that certainly reduces fatigue. Of course the good recording sound better and the bad recordings sound worse. So go play the bad ones in the car. 

[corndog71]

Thanks, Roger.  I have homework now as there are many points in your post to respond to. I'll have to do some research.  I also rode a bike 8.5 miles after work!

[neobop]

The first is a good question, the second I cannot answer.  "Does the audio signal pass through a coupling cap?"  Pass through to me means there is a current. In a crossover there is certainly a substantial current that can be several amperes. Indeed things like ESR make a big difference. Dielectric absorption (DA) also makes a difference as the low frequencies are absolutely pumping the dielectric material up and down with many volts. The louder you play the speaker, the lower the impedance, the larger the effects. Yes a big bass transient leave behind a "ghost voltage"in the tweeter series cap due to DA. Lets not have that. I advise people to change caps in their crossoves without reservation. However, I go further to encourage them to save that money and get and electronic crossover (where the caps pass no current) and go direct to their drivers. Crossovers overboard!  .

Of course, changing the caps is easier. How many are willing to take that big step and make a big difference in their systems. The point about multi-amping is that you can choose specific amps for their specific job. I just hooked the OTL-1 to my horribly inefficient (less than 80 dB/watt) ESL speaker. By design that amp only sees 100 Hz up and a 500 watt transistor amp of no distinction plays 100 Hz down. Even with that inefficient ESL I can play plenty loud. With a conventional speaker of that low efficiency playing full range I would be running out of power in the bass region. 

I recall the whole boutique cap thing started in crossovers and with that I totally agree. Damn the 10 cent electrolytics with their high ESR and loss of capacitance over time. They are usually 10-20% accurate in value so the crossover points and slopes can vary widely. Bypass them but that won't fix the DA problem because the bypass is to small, but it will help the ESR.

Having said all that in my honest opinion as one who does understand how electrons get about in a system, there is virtually no current (a few microamps) at best and no voltage (the cap has little or no reactance at audio frequencies) that the audio signal does not pass through the anymore than it passes through a wire. (I am not inviting a discussion of hook up wire in an amp )

Are you saying that coupling caps in an amp make as big a difference in your hearing as crossover caps?

The best analogy I can think of right now is that high octane gas vs low octane gas will make a big difference in a high compression engine. It will make less difference in a low compression engine and will make no difference, in my experience, in starting a fire. 

Looking at different amps/preamps and gestalt on the net, it seems like you're a minority of 1.  I'm not an EE or a circuit designer and certainly don't understand every possible movement of electrons, but it seems virtually everyone else hears significant differences using technically superior coupling caps.

What's the function of a coupling cap, to block DC ?   Then it's very much like a speaker crossover cap only not in the middle of the audio band.  It might not matter if very little current goes through it, it's still in the signal path.   I was told by an EE that every part in a circuit is in the signal path, whether it's in series or a shunt capacitor.  True?  Why do high end preamps have no tone controls, even ones that can be switched out?  I was told that even having the wires connected whether the circuit can be switched out or not, would degrade the sound.  Do your preamps have tone controls?  If not maybe it's for a different reason?

A speaker crossover might do more than divide frequencies, and slopes on different drivers aren't necessarily the same.  There might be impedance compensation and/or amplitude response correction.  I don't how common zobels or series notch filters are, but they certainly exist and their exclusion might degrade performance of an individual driver without them.  Crossover design is aided by computer these days and such things as phase performance between drivers is complex.  In a general way it might be beneficial to lose the crossover and bi-amp or tri-amp, but that depends on the speaker.

At the top of this thread you asked for suggestions for those looking for better quality coupling caps:
http://tubelab.com/designs/tubelab-se/manual/capacitors/

http://diy.ecpaudio.com/p/some-notes-on-coupling-capacitors.html

http://www.diyaudio.com/forums/tubes-valves/82502-best-coupling-cap-under-20-a.html

http://www.high-endaudio.com/Modifications.html

http://www.gearslutz.com/board/geekslutz-forum/434325-what-best-musical-sounding-capacitors-use-recap-pro-audio-gear-2.html 

[DaveC113]

Looking at different amps/preamps and gestalt on the net, it seems like you're a minority of 1.  I'm not an EE or a circuit designer and certainly don't understand every possible movement of electrons, but it seems virtually everyone else hears significant differences using technically superior coupling caps.

Eh, check the diyaudio tube amp forum. LOTS of very talented tube designers seem to have hearing impediments. 

Sorry, I couldn't resist, just a joke... 

[corndog71]

It is interesting to hear that some boutique caps made it worse some made it better. All in all I guess you ended up spending as much or more on the caps than the original Bottlehead amp. I know the V- caps are very expensive, especially in the large sizes. I bought some for customer at his request. As I said, I don't mind doing it, I don't feel its the best use of money when other more important things can be done in a system.

As to what you heard. The 1 uF output coupling cap is conducting current so that is the one that should make the most difference. The difference in current is more than ten thousand to one vs the interstage cap. The output coupling cap is driving a load of a few thousand ohms, the interstage cap is driving something like 470,000 ohms and at lower 15 times lower signal level (15 is the mu of tube).

I also think the output coupling cap should be a lot bigger. If you are driving a 3000 ohm transformer the -3 dB down point is 53 Hz with 1 uF. This should be a problem, however 32 ohm phones move the -3 point down by 4 to 1 is the inductance of the output transformer is sufficient. Has anyone measured that? I clicked on specs but couldn't find any.

I never use less than 33 uF in my rather similar 6EM7 parafeed amp.  Why is that cap so small? These are the things that I think need attention.
Here is a nice calculator for that sort of thing. http://www.sengpielaudio.com/calculator-XLC.htm

I am not surprised to hear the current source mod made a big difference. That is a major circuit change. In my mind it has much more influence over the sound than a cap change. A current source unloads the input triode reduces distortion and raises gain. Did you notice the gain change?

If you had a variety of output transformers and chokes to try you would hear bigger differences and it's not about the materials. Magnequest promoted some unusual expensive iron some time ago, where is it now? Seems everyone making parts has become a self appointed materials expert.

However who wants to fool around with changing transformers, though yours are pretty easy to get to.  Do you know the primary and choke inductance values. They had better be large. Why dont people fool with transformers more than capacitors. The have a bigger chance of making a difference. I will propose that a tranformerm is too difficult and varied to specify. You have to express the impedance ratios, the input inductance, the leakage inductance, the copper loss, the iron loss, the Low frequency saturation point. A cap is easy to specify, capacity and voltage, done.  The rest is going on and on about exotic materials, winding techniques, leads, yada, yada , yada.

How about changing the 6DN7 to a 6EM7 which has the same pinout but  lower output impedance. You would have to change the cathode resistors but little else I imagine. The output impedance is 750 vs 2000 ohms, almost 3 times lower. That will make a big difference.

I am simply stating that there are bigger things to do that fooling with caps, however, for some reason fooling with caps has become very popular perhaps only one step below tube rolling. I think it is obvious in any hobby that people change the things that are easiest to change, while leaving the more difficult yet more important things alone.

I could make you some iron that would blow the socks off the Bottlehead stuff. Look at the pricing of the kit. Break it down by the parts you know the cost of and you will see that the two output chokes and two output transformers can only cost about $25- $50 each. If you have $250 to spend on a Vcap how about the same for a transformer or audio choke?

How about replacing that $2 volume control with a Noble or Alps.

I really don't understand how more resolution caused listener fatigue. I find the more detail I get the more I can turn down the volume and that certainly reduces fatigue. Of course the good recording sound better and the bad recordings sound worse. So go play the bad ones in the car. 

I bought the V-caps out of curiosity.  I probably wouldn't spend that much on them again although I do think they sound excellent.

I don't have the current version of the the Sex amp which is version 2.1.  Mine is the previous 2.0 version.  Mine has off the shelf chokes and ouput transformers that have since become discontinued.  I believe the choke is 30mH and the OT is 8K.  The newer version has custom-made iron.  Bottlehead does offer their version 2.1 iron as an upgrade.  They don't post the values so I can't comment on that.  Magnequest still offers upgraded sets as well and uses Nickel as the special metal to increase the price.  According to the bottlehead guys the more nickel the better the sound.  I think they wrote some more about it but I can't seem to find it.
I was planning on getting the upgraded bottlehead iron as it's only $200.  If you can make better iron then I would certainly be interested.   
They supplied 1uF for the outputs but also recommended 2uF as an option; especially with the upgraded iron.

Switching to 6EM7 sounds like fun too.  The bottlehead guys did tinker a bit with the 6EM7 tubes and posted the needed changes although I haven't heard of anyone actually doing it.  The whole point of the kit was that it was flexible enough to try various circuits/ tubes.

The current source did improve the amp considerably.  I do recall a bump in gain.

I do think you're right about why people mess with caps.  They are easy to figure out the values needed and swap out.  As a seemingly never-ending novice to this stuff I can personally confirm this.    The bottlehead gear is designed for guys who want to put the puzzle together but don't necessarily want to figure out the details behind each piece. 

I've considered upgrading the volume knob but usually I just max it out as I have a separate preamp connected upstream.  The sex amp only has one input whereas my preamp has two. 

Yeah, the MR caps being too revealing really threw me.  It did sound fantastic but at the same time irritating after a while.  I have the MR caps in my preamp and don't get the same effect.  It's weird.  Maybe I need a Shakti stone or ultrasonic filter.   

[rpf]

Okay, I'll jump into this morass. And hopefully won't get attacked too badly.   

Yes, for those of us with no technical knowledge, or even soldering skills, having caps changed is an easy, relatively inexpensive, path to a potentially significant upgrade.

I've had caps changed on a bunch of equipment as follows and usually found an easily noticeable difference. And, no, I'm not prone to a prolonged placebo effect.

Changed the output coupling caps in a pair of Dodd Mono 50s from the stock Sonicap Is to Sonicap Platinum teflons. The latter had more detail but were lacking in warmth and had an overly smooth, "slippery" quality to them that I hated: preferred the less expensive originals. Changed them out again for Rel AudioCap PCUs (copper foil) and was very happy: more resolution, warmth and tonal accuracy across the board.

Changed the output coupling caps on a MHDT Havana DAC from the stock proprietary ones to Jupiter HTs. Heard a very large increase in resolution, dynamics, soundstaging dimensions and tonal accuracy. Changed the input caps and an interstage coupling cap from the stock ones to Jensen copper foil caps and think I heard a slight improvement in tonality. The latter were not worth the difference.

Changed the output coupling caps on an Eastern Electric Minimax DAC from stock to Jupiter HTs. Large increase, as above in the Havana. Changed some other (interstage coupling?) caps to Rel PCUs and heard no difference.

Changed the output coupling caps on an Eastern Electric DAC Plus from stock to VH Audio OIMP and heard a more relaxed, less digital presentation in the highs, and a bit more warmth to the sound.

Changed the output coupling caps on a Rogue Cronus Magnum from the stock un-labelled low cost Mundorfs to Mundorf Evo Silver/Gold/Oil and heard a slight increase in detail and a larger increase in refinement.

My takeaway is that the type/brand of output coupling caps used have a significant effect on the sound: other caps not so much.

All of the before and after values were identical except for the output caps on the Havana which were 10% higher in value after the change. 

[Speedskater]

Boutique capacitor manufacturing:



Mainline capacitor manufacturing:

[Roger A. Modjeski]

It takes big expensive machines to make capacitors of the highest quality construction. "boutique" capacitors (unless custom manufactured by a major manufacturer) are seldom of the highest quality construction.

I agree about the unreliability of Boutique caps. Thanks for the pictures of the machines. The machine in the bottom picture is quite something and cost big $$$. Do you have a link to the company that made it?

In any art there is the combination of the machine and person using it. There is a lovely book "In Marconi's Footsteps" where the author reproduced the early commercial radio equipment made by the Marconi company circa Titanic. In the book he has lovely color pics of the wood and brass radio sets (before tubes) that he made with a hand drill and a hand jig saw. I was impressed to see such beautiful work done with the most basic, non precision tools.

In the early days I made some very good transformers on a very crude winder I made my self. Futterman made his power transformers on a lathe. Bass Lim at RelCaps made capacitors on equipment he made himself. His company's claim to fame was that he could make 1% capacitors consistently. While 1% caps are nice to make RIAA EQ stages accurate, they make no difference in coupling where the value is larger than the lowest frequency. I bought all my 1% RIAA caps for the RM-5 from him. His main customers were telecommunications companies making very accurate, repeatable filters where 1% caps are necessary. He further told me that the big makers (Sprague, ITT, Mallory, etc) would also sell me 1% caps but they would do it by winding a big lot, measuring each one and selling me the ones that were within 1%, the rest they could sell in their regular 10 % line. Out of perhaps 7,000 rel-caps I bought from him over the years only 1 has failed open. None have ever shorted.

Bass also taught me about self-healing capacitors. Metalized film caps will heal from a short all by themselves. This is not true of film and foil capacitors. Silver, tinfoil, oil and paper and copper caps are not self-healing. The caps in the Jadis were not self healing. Once shorted they stay shorted.

There is a lot to know about capacitors. This is just the tip of the iceberg.

[BobM]

Roger - I think this question may have inadvertantly gotten lost in the recent large posts, so I will ask again.

How about adding a smaller, quicker bypass cap to those large value electrolytics in the power section of an amp or preamp or CD player or whatever? IYO is there any benefit to that?

[Roger A. Modjeski]

I bought the V-caps out of curiosity.  I probably wouldn't spend that much on them again although I do think they sound excellent.

I don't have the current version of the the Sex amp which is version 2.1.  Mine is the previous 2.0 version.  Mine has off the shelf chokes and ouput transformers that have since become discontinued.  I believe the choke is 30mH and the OT is 8K.  The newer version has custom-made iron.  Bottlehead does offer their version 2.1 iron as an upgrade.  They don't post the values so I can't comment on that.  Magnequest still offers upgraded sets as well and uses Nickel as the special metal to increase the price.  According to the bottlehead guys the more nickel the better the sound.  I think they wrote some more about it but I can't seem to find it.
I was planning on getting the upgraded bottlehead iron as it's only $200.  If you can make better iron then I would certainly be interested.   
They supplied 1uF for the outputs but also recommended 2uF as an option; especially with the upgraded iron.

Switching to 6EM7 sounds like fun too.  The bottlehead guys did tinker a bit with the 6EM7 tubes and posted the needed changes although I haven't heard of anyone actually doing it.  The whole point of the kit was that it was flexible enough to try various circuits/ tubes.

The current source did improve the amp considerably.  I do recall a bump in gain.

I do think you're right about why people mess with caps.  They are easy to figure out the values needed and swap out.  As a seemingly never-ending novice to this stuff I can personally confirm this.    The bottlehead gear is designed for guys who want to put the puzzle together but don't necessarily want to figure out the details behind each piece. 

I've considered upgrading the volume knob but usually I just max it out as I have a separate preamp connected upstream.  The sex amp only has one input whereas my preamp has two. 

Yeah, the MR caps being too revealing really threw me.  It did sound fantastic but at the same time irritating after a while.  I have the MR caps in my preamp and don't get the same effect.  It's weird.  Maybe I need a Shakti stone or ultrasonic filter.   

I read about your Bottlehead amp last night. I have been watching Bottlehead for years and noted their line has expanded significantly. I noticed the $199 upgrade for the transformers and chokes. It would be nice for them to give some comparative specs. I note that Doc is more of a story teller than a numbers guy. I hardly see any numbers for the equipment.

On the transformers: They clearly admit using a line matching transformer (available everywhere for $5 to $9) as the output transformer. I have measured these and they start to fall off around 5 KHz and are down several dB at 20KHz  They are indeed attractive as they are a common off the shelf item used in every distributed sound system. When the principal came over the PA system in your school you were listening to him through these transformers. There is one on each speaker and the taps on it allowed the system to play each speaker at a different volume as required. The taps are literally labeled in watts. The input is 25 or 70.0 volts constant level. The installer chooses a power tap to get the appropriate volume for that room, hallway, auditorium or cafeteria speaker. Since they are intended for PA the bandwidth is limited. The windings are not interleaved and they are made in the simplest way. They are just fine for what they are intended. From a price point of view he is clever to use them. I was on to them at about the same time. Their bandwidth is load dependent and they do work much better at 32 ohms than 8.  At 32 ohms at headphone levels they are actually pretty good if properly applied.

A 2 uF coupling cap would help the bottom end for an 8 ohm load but the 1 uF is probably acceptable but not ideal at 32 ohms. The only way to know is to make a few simple measurements or just make the cap bigger and listen to the bass.

I know this suggestion is beyond what some may want to do but for a few hundred  $$ one can buy a scope, generator and AC voltmeter used on eBay and make many of the important measurements such as gain and frequency response.  A distortion analyzer can be had for $200 these days. That includes the generator and voltmeter. Anyone wanting to set up a bench should let me know here and I will make some recommendations. There is so much really good used test equipment available these days as America has given up manufacturing. You can set up a bench for less than the cost of a couple of Boutique caps.

In my classes in SF and Berkeley we will be using test equipment so we can combine measurements and listening tests. Measurements can go a long way into explaining and modifying the sound of an amplifier. However there are some things that appear to defy measurement and have to be determined by listening. I do both, however I want to know that when I hear great bass its not because the response is bumped up. Myself and my fellow designers generally want to make sure something measures reasonably before bothering to haul it into the listening room.

[Roger A. Modjeski]

Roger - I think this question may have inadvertantly gotten lost in the recent large posts, so I will ask again.

How about adding a smaller, quicker bypass cap to those large value electrolytics in the power section of an amp or preamp or CD player or whatever? IYO is there any benefit to that?

What is a "quicker bypass cap"? Caps don't have speed, though I know that language is prevalent in the audio community. However I can translate audiospeak into techspeak. On its face the idea makes sense. The big cap does the low frequencies but is not so good at the high so we put in parallel a little film cap that is good at the high end. Unfortunately, and no one talks about this, it makes things worse. Perhaps they should do the experiment I will describe below.

In 1978 I made a simple experiment to test this. I took a 22 uF electrolytic and a variety of smaller caps (0.01 to 1 uF) and put them in parallel as is suggested everywhere. Then I put them across the output of a square wave generator. The combination rang like hell at the top of the audio band. In the power section, as you inquire, this could easily bump up the response at high frequencies or cause other problems because we want the cap to be a short. Without the bypass there was no ringing and the capacitor did its job properly. Given this discovery I stopped using bypass caps very early in my career.

In most electrolytics the point where the cap starts to give up its ability to be a short is typically above the audio band at 50-100 KHz. Adding the bypass cap actually pushes this point down into the audio band where we don't want it.

[steve f]

Roger, that's one of the reasons I enjoy the Music Reference forum. I learn something almost every time I visit.  Quite a few years ago, I picked up a PS Audio II amp. It was known for being among the first amps to bypass a bunch of caps with various film types.  Being curious, young, foolish, and probably voiding the warranty, I removed some of the bypass caps to try them in a couple of other circuits. I did notice that the other circuits did not sound better, but after removal, the PS did. I didn't understand why. Now I do. Thank you.

[GMuffley]

I have been DIYing tube equipment since 1976 and the availability of better passive components, resistors and capacitors, has been a boon to the hobbyist. Changing coupling capacitors can yield significant and audible changes to the reproduced sound--good and bad.  I have a large assortment of boutique caps that I have tried over the years and my "go to" cap for years has been the Rel-Cap polystyrene; it seems to work in many different amps and preamps and it is a high performance component.  It has temperature limitations compared to polypropylene, but I simply keep it away from high heat sources.  I haven't tried the most expensive V-Caps, but I have not hopped on the teflon band wagon.  Tried the paper and oils as well, but htey didn't give me the results I was looking for.

[DaveC113]

How about adding a smaller, quicker bypass cap to those large value electrolytics in the power section of an amp or preamp or CD player or whatever? IYO is there any benefit to that?

On this one, I have settled with unbypassed Clarity TC caps rather than use any electrolytics at all. They are expensive and large, but the results have been worth it imo.

[corndog71]

On this one, I have settled with unbypassed Clarity TC caps rather than use any electrolytics at all. They are expensive and large, but the results have been worth it imo.

I agree that they are worth it.  And Clarity does say these are self-healing.

[paul79]

IME, a good polypropylene film & foil cap works well bypassing power supply electrolytics.

[Freo-1]

Mundorf Silver/Gold are my favorites.  Great coupling caps.