Capacitors...debunking the shame of vintage crossovers

[JohnVF]

Here is a simple exercise, think about the recording studio and how many electrolytic capacitors the recordings we love have passed through. It's probably hundreds with all the recording equipment. Does it really make a lot of sense that the last one in the speaker is going to have such a drastic effect on the signal? If it did then it would mean that the hundreds that signal passed through before you listened to it would have already rendered the musical instruments so "smeared" and "distorted" it couldn't hardly be recognized as the original sound clip. Of course we know that's ridiculous, just as ridiculous as claiming such drastic effects from a measly capacitor inside a crossover network. Don't get me wrong I have measured some very minor differences in some applications but it's far below audible. I find it hilarious that they want to bring up the "transient response" of a capacitor yet they ignore the transducer. The latter, which takes electrical signal and converts it to mechanical work, I.e. a motor to push sound pressure waves. These transducers have mass, they are going to have a magnitude higher amount of distortion in regard to instant impulse response. It's like ignoring the elephant in the room while trying to say Bigfoot or a unicorn is somewhere there instead and to blame for all the elephant like effects we are perceiving at the time.

You would probably be shocked at just how much purposeful voicing and subjective manipulation happens in recording studios, from the choices of transformers in microphone preamps, to use old/outdated limiters and other devices to 'smear' or otherwise distort the sound in pleasing ways determined only by the subjective choices of the producers and engineers involved, on down through the guitarist who voices his tone with some of the things you say simply aren't audible or that they don't even exist.

Chill with the condescension. The biggest sign of somebody's advice to ignore is somebody who knows just enough to think they know everything, without having learned enough to know they don't.

 

[S0und Dragon]

It saddens me when folks take the open and hospitable attitude of a place like this and seem to go out of their way to make it like other places. There are definitely places that are way more receptive to aggressive numbers only thinking. This place is very much not that. While we embrace the objective side of things, we do not discount the subjective either. Both camps have a safe haven here. It is regrettable that there are folks who can’t seem to understand the vibe of this place and thrive in it. Some folks would rather watch Rome burn. Oh well… there are no more warnings to give.

 

[EarlK]

I think the "Numbers Only Gamers" should spend their energies on proving that "Imaging > both Depth and Width" doesn't exist , since after all, it's immeasurable.

 

[Tables_and_Tubes]

At first, I was very interested in this thread. My wife inherited a pair of mid seventies Klipshhorns and I need to clean them up, and get them setup in the living room. I am very concerned about cap choices for the crossovers, as I don't want the upper mids and treble to get any more "aggressive".

Like many other sites, the thread went off the rails when the " techs" and EEs came in and told us all what we can't hear.....sigh.

Dan

 

[JohnVF]

I would also like to point out that almost every successful and well-regarded engineer in this industry embraces both the objective and subjective side of thinking and experiencing things. It is only on the more hobbyist-level and in a very small set of manufacturers who have found a niche catering specific measurements of their devices, like SINAD, to please the rankings on certain forums to induce sales, that the objective-only, nothing else matters, mindset exists. Which is FINE... because there are people who purchase things based on a limited scope of measurements or even a single measurement.

If anybody wants to see a great balance of objective and subjective analysis in speaker review, @ErinsAudioCorner on Youtube is like a fantastic course in bridging the two worlds, with a wealth of well-done objective measuring of speakers mixed with subjective correlation (or not) when listening. It really helps explain the type of measurements that are done more on the manufacturing end of things, while bridging that with the consumer end of things. Plus he's just a cool non-know-it-all who is curious about why things are the way they are. I can't recommend him enough, to both sides of the objective/subjective divide (hello from the middle).

 

[JohnVF]

At first, I was very interested in this thread. My wife inherited a pair of mid seventies Klipshhorns and I need to clean them up, and get them setup in the living room. I am very concerned about cap choices for the crossovers, as I don't want the upper mids and treble to get any more "aggressive".

Like many other sites, the thread went off the rails when the " techs" and EEs came in and told us all what we can't hear.....sigh.

Dan

There are plenty of experienced EEs who are just as die-hard in their opinion about changing out old caps for newer ones. I would love to get the last two EEs who have restored tube gear for me in the same room and have them fight each other over the choice of caps each put into two different old Fisher tube amps of mine...as each had a completely opposite opinion to the other when it comes to using something more period-correct and leaving well-enough alone, vs saying period-correct is idiotic as time and reliability has moved on, but each was just as obstinate in why they were right and everybody else was wrong. Sometimes its less about objectivity and more about having a personality type that is quite rigid to the notion that there are other opinions than your own. "Objectivity" is just the smoke-screen employed for being extremely opinionated.

 

[captainbeefheart]

I'm not all about the numbers, in fact the opposite which is why I stated that beyond a certain point low enough things are just not audible. We have background noise, device noise, distortion from transducers, distortion from devices etc.. that are magnitudes larger in their effect of on the signal to which things down -140db are going to be lost and not important.

I'm trying to put perspective into the discussion is all. I also never stated that some capacitors might have slight frequency response effect via enhanced brightness. But distortions and impulse response aren't going to be of any concern if they are even measured since when they are it's typically extremely small in these applications.

I think the "Numbers Only Gamers" should spend their energies on proving that "Imaging > both Depth and Width" doesn't exist , since after all, it's immeasurable.

Except it is without question measurable and well understood. For the most part this is discussed in speaker design or speaker placement and room treatment side of things. It's time based measurements from delays/reflections in the listening space.

This of course can be done at the signal level like adding reverb to make a speech sound like it was recorded in a large room. Imaging/depth is from natural live recordings where the mic is picking up the in room reflections and delays. Again it's easily measurable and explainable.

 

[captainbeefheart]

There are plenty of experienced EEs who are just as die-hard in their opinion about changing out old caps for newer ones. I would love to get the last two EEs who have restored tube gear for me in the same room and have them fight each other over the choice of caps each put into two different old Fisher tube amps of mine...as each had a completely opposite opinion to the other when it comes to using something more period-correct of saying period-correct is idiotic as time and reliability has moved on, but each was just as obstinate in why they were right and everybody else was wrong. Sometimes its less about objectivity and more about having a personality type that is quite rigid to the notion that there are other opinions than your own. "Objectivity" is just the smoke-screen employed for being extremely opinionated.

Was it the same piece of equipment at both techs? I ask because at a certain point caps were made very reliable. For example there might be a piece from the 60's with Polyester caps throughout it and from experience I know these are extremely reliable. Of course I'll make sure there is no DC leaking by to the next stage. If the output of the amplifier is meeting or exceeding performance specifications then there is really no point to replace the capacitors in my opinion. On the other hand if I open a piece of gear that I know have paper caps that are extremely prone to failure or leakage I'll advise replacing them especially since most likely they will be leaking some DC by throwing off the operating point of the next stage. It really depends on the exact type of capacitor used and not generalization.

Speaking of generalization. A coupling capacitor is supposed to be of low impedance and pass intended frequencies, not attenuate them. By the nature of the application there is very little signal ACROSS the capacitor so they cannot impart their dielectric properties strongly onto the signal. I.e. type of coupling cap matters very little in regard to sound quality, it's not even measurable if they are sized appropriately. Some designs will purposely size them small to attenuate low frequencies, typically for stability and in these applications where at say 35Hz you may see 8% of the signal across the cap. Here is where the dielectric properties might start to give their nonlinear influence of the matter and you'll measure a rise in distortion. In this case it depends how low the distortion is, most likely not audible at all but it would make sense here to use a more linear capacitor so no ceramics and no electrolytics, most films and paper are quite linear.

 

[EarlK]

Except it is without question measurable and well understood. For the most part this is discussed in speaker design or speaker placement and room treatment side of things. It's time based measurements from delays/reflections in the listening space.

This of course can be done at the signal level like adding reverb to make a speech sound like it was recorded in a large room. Imaging/depth is from natural live recordings where the mic is picking up the in room reflections and delays. Again it's easily measurable and explainable.

Imaging is certainly explainable ( since most people attest to hear the phenomenon ) and many have spent copious energies on promoting explanations that most of us can abide by. I believe it comes down to how many doctorate students have done a paper on the subject.

Measurable?? Please enlighten me here on how-to.
I'd love to be able to stick a mic ( or two ) in front of a pair a speakers, capture the data, process it into visual presentation form and then have everyone ( if not most ) view that graph to either agree that my speakers are imaging champs ( or not ).

 

[JohnVF]

Was it the same piece of equipment at both techs? I ask because at a certain point caps were made very reliable. For example there might be a piece from the 60's with Polyester caps throughout it and from experience I know these are extremely reliable. Of course I'll make sure there is no DC leaking by to the next stage. If the output of the amplifier is meeting or exceeding performance specifications then there is really no point to replace the capacitors in my opinion. On the other hand if I open a piece of gear that I know have paper caps that are extremely prone to failure or leakage I'll advise replacing them especially since most likely they will be leaking some DC by throwing off the operating point of the next stage. It really depends on the exact type of capacitor used and not generalization.

Speaking of generalization. A coupling capacitor is supposed to be of low impedance and pass intended frequencies, not attenuate them. By the nature of the application there is very little signal ACROSS the capacitor so they cannot impart their dielectric properties strongly onto the signal. I.e. type of coupling cap matters very little in regard to sound quality, it's not even measurable if they are sized appropriately. Some designs will purposely size them small to attenuate low frequencies, typically for stability and in these applications where at say 35Hz you may see 8% of the signal across the cap. Here is where the dielectric properties might start to give their nonlinear influence of the matter and you'll measure a rise in distortion. In this case it depends how low the distortion is, most likely not audible at all but it would make sense here to use a more linear capacitor so no ceramics and no electrolytics, most films and paper are quite linear.

One tech filled my Fisher 400c with NOS orange-drops that dated back decades as well as other period-correct replacement parts, and another looked at it recently after it developed a balance issue and just shook his head and almost got angry with what the other guy had done "all these guys think there's some magic in this old crap and there's not" "he was adamant that it should sound like it did when new" "well it doesn't, because these aren't new and now it just sounds old". The rest of us are just left in the middle wanting our gear to be repaired without having to choose one dogma over another. Its really frustrating, to be honest.

 

[captainbeefheart]

Measurable?? Please enlighten me here on how-to.
I'd love to be able to stick a mic ( or two ) in front of a pair a speakers, capture the data, process it into visual presentation form and then have everyone ( if not most ) view that graph to either agree that my speakers are imaging champs ( or not ).

You can't stick the mic just anywhere, it would have to be exactly where your ears are at your listening position. You can download REW or other software to help yourself plot your room out and work on treatment and speaker placement. This is why DSP and active crossovers are becoming so popular because it can process the signal to more easily and quickly achieve these results without having to physically alter the room reflections manually.

 

[JohnVF]

Imaging is certainly explainable ( since most people attest to hear the phenomenon ) and many have spent copious energies on promoting explanations that most of us can abide by. I believe it comes down to how many doctorate students have done a paper on the subject.

Measurable?? Please enlighten me here on how-to.
I'd love to be able to stick a mic ( or two ) in front of a pair a speakers, capture the data, process it into visual presentation form and then have everyone ( if not most ) view that graph to either agree that my speakers are imaging champs ( or not ).

You can tell a bit by looking at vertical and horizontal radiation/response graphs, and also you can tell a bit from those if the speaker will be more or less room dependent. But its easier to just listen..... that said, its going to vary from room to room if they're not well treated, and that variation will also depend on that radiation as will what the speaker sounds like on and off axis.

And yes they do stick, essentially, a mic in front of the speaker to do these measurements. But its more complicated than that.

*I'm not a speaker measurer and have only a limited understanding of it, I'm just a nerd who enjoys watching other people do it.

 

[captainbeefheart]

One tech filled my Fisher 400c with NOS orange-drops that dated back decades as well as other period-correct replacement parts, and another looked at it recently after it developed a balance issue and just shook his head and almost got angry with what the other guy had done "all these guys think there's some magic in this old crap and there's not" "he was adamant that it should sound like it did when new" "well it doesn't, because these aren't new and now it just sounds old". The rest of us are just left in the middle wanting our gear to be repaired without having to choose one dogma over another. Its really frustrating, to be honest.

Ya that's ridiculous. Those old orange drops are well built and very reliable, they were manufactured well after capacitor development had many of the bugs worked out. They most certainly are not adding anything positive or negative to the sound so long as they are healthy, which most likely they are since I rarely find a failed one in anything. Did either of them test the output of the amplifier to see if it's meeting spec and also check stability? If the output is good then it is good. I have found it's rarely anything to do with the type of capacitor used, it's more to do with the entire amplifier operating as it should be, that includes stability criteria.

In this situation I would completely ignore the second tech, he sounds like a curmudgeon that doesn't quite understand things well. Yes there are plenty of EE's that don't actually understand things properly, that paper/degree is meaningless many times. The first guy probably understood even the old orange drops are well made capacitors and reliable, and many people prefer their vintage equipment to look period correct as it will bring in more money at sale time. It's a balance of period correct restoration and performance. If all you had was a balance control issue the second tech should have kept his mouth shut and just fixed the problem it was there for.

 

[JohnVF]

Ya that's ridiculous. Those old orange drops are well built and very reliable, they were manufactured well after capacitor development had many of the bugs worked out. They most certainly are not adding anything positive or negative to the sound so long as they are healthy, which most likely they are since I rarely find a failed one in anything. Did either of them test the output of the amplifier to see if it's meeting spec and also check stability? If the output is good then it is good. I have found it's rarely anything to do with the type of capacitor used, it's more to do with the entire amplifier operating as it should be, that includes stability criteria.

In this situation I would completely ignore the second tech, he sounds like a curmudgeon that doesn't quite understand things well. Yes there are plenty of EE's that don't actually understand things properly, that paper/degree is meaningless many times. The first guy probably understood even the old orange drops are well made capacitors and reliable, and many people prefer their vintage equipment to look period correct as it will bring in more money at sale time. It's a balance of period correct restoration and performance. If all you had was a balance control issue the second tech should have kept his mouth shut and just fixed the problem it was there for.

The 2nd tech put it on a scope and everything was fine except for a hint of noise that we both guessed was it just needing a good cleaning of switches and connectors. The balance issue was in the selector knobs being gunked up after 60 years, I need to take the thing apart a bit to get at the switches.

 

[captainbeefheart]

And yes they do stick, essentially, a mic in front of the speaker to do these measurements. But its more complicated than that.

That's wrong if you want your listening spot to be your intended holy grail position. I only see speaker designers stick the mics in front of the speaker because for speaker design you don't want room reflections to be part of the equation. To get the end result, which is in room response you have to place the mic where you are sitting in order to get a handle on all the different room reflections and their delay times. There are direct reflections that come from just one reflected point and there are also second and third point reflections that will reach the listening position later. All of these need to be taken into account for your personal optimal listening enjoyment.

I agree, I don't use DSP, I just do it by ear until I can get the coveted phantom source where the speakers disappear. It's time consuming and sometimes frustrating but you can get darn close to excellent with some effort.

 

[captainbeefheart]

The 2nd tech put it on a scope and everything was fine except for a hint of noise that we both guessed was it just needing a good cleaning of switches and connectors. The balance issue was in the selector knobs being gunked up after 60 years, I need to take the thing apart a bit to get at the switches.

Get a new tech.

A scope isn't good enough for analysis. You can have 1% distortion that's barely even noticeable on a scope when viewing a test tone sine wave. A distortion analyzer will tell them much more what's happening even with noise, you'll see a bump in the FFT plot at either 120Hz or 60Hz. You can see on the scope what noise it is if you adjust the time domain to see if it's 60Hz or 120Hz, but since they are octaves of each other it's best to view in FFT plot where you can actually see clearly the amount of each one.

I like to also test for IMD and beat note distortion of amplifiers while they are on the bench along with stability and impulse response tests.

 

[JohnVF]

That's wrong if you want your listening spot to be your intended holy grail position. I only see speaker designers stick the mics in front of the speaker because for speaker design you don't want room reflections to be part of the equation. To get the end result, which is in room response you have to place the mic where you are sitting in order to get a handle on all the different room reflections and their delay times. There are direct reflections that come from just one reflected point and there are also second and third point reflections that will reach the listening position later. All of these need to be taken into account for your personal optimal listening enjoyment.

I agree, I don't use DSP, I just do it by ear until I can get the coveted phantom source where the speakers disappear. It's time consuming and sometimes frustrating but you can get darn close to excellent with some effort.

I'm talking about measuring the speaker itself not what it will sound like in a specific room. Something like these from the previously mentioned Erin's Audio Corner. I'm not great at evaluating these things but it would seem to have a balanced directivity without a super wide dispersion (even less so vertically so get the height right). Each room will be different but this might help determine -how- different.

 

[captainbeefheart]

I'm talking about measuring the speaker itself not what it will sound like in a specific room. Something like these from the previously mentioned Erin's Audio Corner. I'm not great at evaluating these things but it would seem to have a balanced directivity without a super wide dispersion (even less so vertically so get the height right). Each room will be different but this might help determine -how- different.

Ahh gotcha.

Yes you could extrapolate that type of data and use it for what reflections you'll be dealing with in room. If I can find it I have a nice image showing different speaker types, and in varying positions in different rooms and where all the various 1st, 2nd, 3rd reflection points are going to be. I believe it was from one of my Floyd Toole books. Essentially you could overlay the speaker polar pattern onto these images and draw out each angle throughout the room. Do some backhand calculations on what sort of time delays you might be dealing with.

I implore anyone that spending time with these type of modifications will have the largest effect on your listening experience, far more than changing capacitor dielectrics. But to each their own and I'm not trying to take peoples joy away. I only want to give some perspective on what's the bigger fish to fry.

I'll have a look for those parts of the books and upload them. They were quite useful when I want to work on speaker placement and room treatment.

 

[JohnVF]

Ahh gotcha.

Yes you could extrapolate that type of data and use it for what reflections you'll be dealing with in room. If I can find it I have a nice image showing different speaker types, and in varying positions in different rooms and where all the various 1st, 2nd, 3rd reflection points are going to be. I believe it was from one of my Floyd Toole books. Essentially you could overlay the speaker polar pattern onto these images and draw out each angle throughout the room. Do some backhand calculations on what sort of time delays you might be dealing with.

I implore anyone that spending time with these type of modifications will have the largest effect on your listening experience, far more than changing capacitor dielectrics. But to each their own and I'm not trying to take peoples joy away. I only want to give some perspective on what's the bigger fish to fry.

I'll have a look for those parts of the books and upload them. They were quite useful when I want to work on speaker placement and room treatment.

On the pair of speakers I'm listening to now, Dynaudio Special 40s, I looked up the vertical and horizontal dispersion and realized that they were far less forgiving than I'd assumed for vertical placement. I happened to have two different stands around here and put them on the pair that lowered them about 4 inches, and they really sounded a lot better. Yes its something I could have arrived at by just messing around with the different stands, but the lower pair was in the basement and I didn't dig them out until seeing the measurements. They sounded good before, they sound better now. So a case of objective measurements helping out my subjective listening experience. I didn't do any measurements, just listened to the results of an actual inspired by looking at measurements.

 

[MrEd]

The earth was believed to be flat at one time until we had that ability to see and realize differently.
There many things we can not measure, as advanced as we may think we are.