Hello,
This is my first post on this forum, so first of all I would like to greet everyone. Here's what brings me here this time 😉
My speakers (ReVox BX-350) will soon be 50 years old, it's time to check the crossovers 😉 Since I already have experience with the renovation of ReVox Emporium B speakers and I don't want to "modernize" unnecessarily what is original and works, I have a few questions for you:
1) Has anyone of you unsoldered the old foil capacitor with plastic housing from the 70's crossover (in the ReVox BX-350 crossover it is Wima MKS 4 6.8 uF 100 V) and measured its capacitance? I ask because these foil capacitors in crossovers are considered "eternal", but during a recent repair of a home cinema projector I discovered that the foil capacitors installed there had lost 50 percent of their capacitance after 15 years of operation. If such a loss occurred in the speaker crossover, it would mean a very significant change in parameters and characteristics. To measure a foil capacitor in BX-350 speakers, you first need to unglue and desolder it, which can destroy the capacitor, because the glue is very strong. Please write if any of you have removed such a foil capacitor from any old speaker crossover and measured its capacitance.
2) Do you know of any chemical that dissolves the glue that is used to glue capacitors to the board in speaker crossovers? Maybe toluene would work - it is used to unglue speaker membranes from foam and rubber suspension?
And I will also share my experience with replacing electrolytic capacitors in other Revox crossovers. When my other speakers - Revox Emporium B - started giving strange results in the characteristic measurements in the midrange speaker range, I decided to start by replacing the capacitors in the crossover. Unfortunately, to measure their parameters, you have to unsolder and unglue them, and there was a lot of glue and it was so strong that ungluing required destroying the capacitors (bending the metal casing). Fortunately, the housings remained sealed and it was possible to measure the capacitance. And now I will share the measurement results 😉
Emporium B are about 35 years old, 3-way, closed cabinet speakers. The capacitors removed from them had identical or better parameters than the new Mundorf capacitors with smooth foil. Capacitance, ESR, Vloss - everything was fine! The only thing I noticed was that both the new and old capacitors had a much higher capacitance than the one given on the housing, by about 15 percent. So replacing these capacitors - and they were very difficult to obtain and very expensive - turned out to be completely pointless. The problem turned out to be not the crossover, but the hardened rubber suspension of one of the midrange speakers, as a result of which the resonance of this speaker shifted into the range of its operation in the speaker column.
Maybe someday I will share the Emporium B renovation process in detail, I have a lot of photos and it was a very interesting adventure - I can say right away that the greatest effect was achieved by thoroughly sealing the housing and the cable passages inside it (multi-chamber cabinet). For now, I'll leave the questions and invite you to the discussion 😉
And actually, this is another interesting topic related to crossovers in vintage speakers, designed without computers. Since almost every bipolar electrolytic capacitor, whether new or old, has a larger capacity than the one written on its housing... Wasn't it by any chance that the designers of these speakers selected electrolytic capacitors according to their actual capacity, and not the one written on the housing? Because if so, then considering an old capacitor, which instead of 15 uF has 18 uF, as "worn out" makes no sense. I think this question is quite reasonable, especially since new electrolytic capacitors also have higher capacities than the manufacturers' data. Perhaps the designers assumed this higher, actual capacity for calculations.
Best regards,
Mike
This is my first post on this forum, so first of all I would like to greet everyone. Here's what brings me here this time 😉
My speakers (ReVox BX-350) will soon be 50 years old, it's time to check the crossovers 😉 Since I already have experience with the renovation of ReVox Emporium B speakers and I don't want to "modernize" unnecessarily what is original and works, I have a few questions for you:
1) Has anyone of you unsoldered the old foil capacitor with plastic housing from the 70's crossover (in the ReVox BX-350 crossover it is Wima MKS 4 6.8 uF 100 V) and measured its capacitance? I ask because these foil capacitors in crossovers are considered "eternal", but during a recent repair of a home cinema projector I discovered that the foil capacitors installed there had lost 50 percent of their capacitance after 15 years of operation. If such a loss occurred in the speaker crossover, it would mean a very significant change in parameters and characteristics. To measure a foil capacitor in BX-350 speakers, you first need to unglue and desolder it, which can destroy the capacitor, because the glue is very strong. Please write if any of you have removed such a foil capacitor from any old speaker crossover and measured its capacitance.
2) Do you know of any chemical that dissolves the glue that is used to glue capacitors to the board in speaker crossovers? Maybe toluene would work - it is used to unglue speaker membranes from foam and rubber suspension?
And I will also share my experience with replacing electrolytic capacitors in other Revox crossovers. When my other speakers - Revox Emporium B - started giving strange results in the characteristic measurements in the midrange speaker range, I decided to start by replacing the capacitors in the crossover. Unfortunately, to measure their parameters, you have to unsolder and unglue them, and there was a lot of glue and it was so strong that ungluing required destroying the capacitors (bending the metal casing). Fortunately, the housings remained sealed and it was possible to measure the capacitance. And now I will share the measurement results 😉
Emporium B are about 35 years old, 3-way, closed cabinet speakers. The capacitors removed from them had identical or better parameters than the new Mundorf capacitors with smooth foil. Capacitance, ESR, Vloss - everything was fine! The only thing I noticed was that both the new and old capacitors had a much higher capacitance than the one given on the housing, by about 15 percent. So replacing these capacitors - and they were very difficult to obtain and very expensive - turned out to be completely pointless. The problem turned out to be not the crossover, but the hardened rubber suspension of one of the midrange speakers, as a result of which the resonance of this speaker shifted into the range of its operation in the speaker column.
Maybe someday I will share the Emporium B renovation process in detail, I have a lot of photos and it was a very interesting adventure - I can say right away that the greatest effect was achieved by thoroughly sealing the housing and the cable passages inside it (multi-chamber cabinet). For now, I'll leave the questions and invite you to the discussion 😉
And actually, this is another interesting topic related to crossovers in vintage speakers, designed without computers. Since almost every bipolar electrolytic capacitor, whether new or old, has a larger capacity than the one written on its housing... Wasn't it by any chance that the designers of these speakers selected electrolytic capacitors according to their actual capacity, and not the one written on the housing? Because if so, then considering an old capacitor, which instead of 15 uF has 18 uF, as "worn out" makes no sense. I think this question is quite reasonable, especially since new electrolytic capacitors also have higher capacities than the manufacturers' data. Perhaps the designers assumed this higher, actual capacity for calculations.
Best regards,
Mike
Hi Mike,
Film capacitors commonly lose capacitance over time as the "foil" burns away. Foil capacitors shouldn't lose capacitance as they can carry the current. They will be larger than the film types.
Electrolytic capacitors used to have the marked capacitance, normally a little higher. That was the way back then. Today's capacitors often are lower than the marked value (to save money). You would replace like types with like types (electrolytic for electrolytic, film or foil replaced with foil in a crossover).
See if you can get a razor blade between the capacitor body and PCB. The PCB is the valuable part, so destroy the capacitor if you have to. Sometimes a heat gun will soften the glue depending on what they used. I do get crossovers in for service now and then, and I used to be authorized warranty for Revox. Never had the need to warranty one of their speakers. I have rebuilt a couple internal actively crossed over models. Generally very good speakers.
Film capacitors commonly lose capacitance over time as the "foil" burns away. Foil capacitors shouldn't lose capacitance as they can carry the current. They will be larger than the film types.
Electrolytic capacitors used to have the marked capacitance, normally a little higher. That was the way back then. Today's capacitors often are lower than the marked value (to save money). You would replace like types with like types (electrolytic for electrolytic, film or foil replaced with foil in a crossover).
See if you can get a razor blade between the capacitor body and PCB. The PCB is the valuable part, so destroy the capacitor if you have to. Sometimes a heat gun will soften the glue depending on what they used. I do get crossovers in for service now and then, and I used to be authorized warranty for Revox. Never had the need to warranty one of their speakers. I have rebuilt a couple internal actively crossed over models. Generally very good speakers.
You are probably speaking about x-capacitors. They are subjected to voltage spikes from mains, which cause local dielectric breakdown and (safe) degradation of both dielectric and metallisation layers. After thousands of such events capacitance will decrease significantly.
https://www.se.com/eg/en/faqs/FA406791/
Obviously, there is no such conditions in loudspeaker crossovers, so film capacitors there have indefinitely long lifespan, as you have seen in Emporium B 🙂
Xylene or MEK would likely work, but don't bother.
Decent designs employing electrolytics shoould be tolerant to their tolerance 🙂
As you noted yourself, there can be more serious problems than some cap being 10% out of spec; unless drivers themselves are known to be really stable during decades of service and overall speaker design have enough merits to be considered adequate even 40 or 50 years later, measurements in electrical domain are of limited value. Axial frequency response measurements should be taken at the very least, and only then you need to pay attention to the crossover.
After all, a loudspeaker is an acoustic device, so any attempt to diagnose or improve it should begin from the acoustic domain.
Another thing to consider is the oxidation of the metal film inside the foil capacitor.
All plastics is hygroscopic and is also open to oxygen, at the same time there is ion contaminations in the plastics as well.
If you then include a bias voltage then you have the perfect mix for agening.
Seen a lot of X and Y caps that has degraded a lot.
All plastics is hygroscopic and is also open to oxygen, at the same time there is ion contaminations in the plastics as well.
If you then include a bias voltage then you have the perfect mix for agening.
Seen a lot of X and Y caps that has degraded a lot.
MKC and polystyrene caps are prone to that. Old stuff had a long lifetime but even with care and cautiousness things still age till a year where they are officially EOL.
Nothing is eternal thankfully. In a distant past I repaired KEF stuff and I have seen enough worn out crossover filters that I just refurbished them in time. Also sometimes left and right not being equal in parts. I recall black caps with cracks in seals.
Yeah they are “original”, original worn out parts.
Nothing is eternal thankfully. In a distant past I repaired KEF stuff and I have seen enough worn out crossover filters that I just refurbished them in time. Also sometimes left and right not being equal in parts. I recall black caps with cracks in seals.
Yeah they are “original”, original worn out parts.
Last edited:
Are we talking about film or foil capacitors. Let's be clear about this. Also keep the application in mind.
Foil capacitors are extremely robust and generally well sealed unless you get cheap ones. Film capacitors are also typically well sealed (again, unless you get cheap ones). The main difference is that the film types cannot handle current levels in crossovers. Foil capacitors are larger in size for the same values, and more expensive. You may well find film capacitors in speaker crossovers when they are trying to save money.
In signal applications I sometimes find Polystyrene capacitors shorted. They don't take well to wave soldering since they are low temperature dielectric types. I have yet to find any decent polystyrene capacitors that have dropped in value. They are typically open (lead pulled probably) or completely shorted. These are parts in excess of 45 years old.
If the electronics have served in humid environments, you're going to see everything from chassis rust to all manner of component problems. You may well see capacitor issues along with all kinds of other issues. Never mind switches. Repairing those is a never ending story.
AC line capacitors see severe service, and I think manufacturers used the cheapest caps they could find that would pass certification. I'm going to bet those have the very highest failure rate of any application. I see Mylar most often in older equipment.
Foil capacitors are extremely robust and generally well sealed unless you get cheap ones. Film capacitors are also typically well sealed (again, unless you get cheap ones). The main difference is that the film types cannot handle current levels in crossovers. Foil capacitors are larger in size for the same values, and more expensive. You may well find film capacitors in speaker crossovers when they are trying to save money.
In signal applications I sometimes find Polystyrene capacitors shorted. They don't take well to wave soldering since they are low temperature dielectric types. I have yet to find any decent polystyrene capacitors that have dropped in value. They are typically open (lead pulled probably) or completely shorted. These are parts in excess of 45 years old.
If the electronics have served in humid environments, you're going to see everything from chassis rust to all manner of component problems. You may well see capacitor issues along with all kinds of other issues. Never mind switches. Repairing those is a never ending story.
AC line capacitors see severe service, and I think manufacturers used the cheapest caps they could find that would pass certification. I'm going to bet those have the very highest failure rate of any application. I see Mylar most often in older equipment.
This sentence should be in bold in all new threads!
I hate to mention this inconvenient fact, but there is relationship between the voltage across the driver terminals and what disturbs the air around your ears.
You will find a transfer function between the acoustic pickup from a microphone and what you apply to the outer speaker terminals. The behavior of components are very predictable, so you can infer a lot of information from the electrical behaviour.
For example. When receiving re-coned drivers (or new drivers) I check linearity - electrically. You can see non-linearity through testing this way before you can hear it. This was commonly done since the 1970's. One plus of doing some checking electrically is that the signal to noise ratio is higher electrically. Other acoustic noises can really interfere with what you are trying to hear. Of course we listen to the thing as well, but do not discount getting extra information from means other than a microphone or your ears.
You will find a transfer function between the acoustic pickup from a microphone and what you apply to the outer speaker terminals. The behavior of components are very predictable, so you can infer a lot of information from the electrical behaviour.
For example. When receiving re-coned drivers (or new drivers) I check linearity - electrically. You can see non-linearity through testing this way before you can hear it. This was commonly done since the 1970's. One plus of doing some checking electrically is that the signal to noise ratio is higher electrically. Other acoustic noises can really interfere with what you are trying to hear. Of course we listen to the thing as well, but do not discount getting extra information from means other than a microphone or your ears.
Couple of thoughts. An impedance measurement tool such as the REW + homemade jig or Dayton DATS is a real accelerator to a lot of this.
In the 1990s and before my experience is that electrolytic caps also had higher ESR specs than their modern bi-polar counterparts. Obviously, if you measure a cap that's 50% off the capacitance spec it deserves to be replaced, but be aware that replacing caps you may significantly alter the series resistance of the crossover.
For this reason I feel it's always important to:
E
In the 1990s and before my experience is that electrolytic caps also had higher ESR specs than their modern bi-polar counterparts. Obviously, if you measure a cap that's 50% off the capacitance spec it deserves to be replaced, but be aware that replacing caps you may significantly alter the series resistance of the crossover.
For this reason I feel it's always important to:
- Measure the entire speaker impedance (both speakers)
- Measure the driver impedance (all drivers, this tests for failure)
- Do a complete crossover simulation with XSim or VituixCAD
- Evaluate ESR changes via the transfer function graphs
- Retest when everything is back in place
E
Generally speaking, as long as you replace like caps with the same family of capacitor you'll be fine. Driver variances are greater than the minute changes in ESR between old and new electrolytic caps, and I have tested old ones with lower ESR than new ones.
You're probably further ahead checking each driver for linearity before doing anything. You'll need a dual trace scope, variable generator with high output or a small amplifier and oscillator, and a 1 K resistor (approximately). You can then accurately get the resonant frequency for each driver. The entire box as well. Put the scope in X-Y mode, adjust for a usable display and look a the pattern. It should look like a propeller on a 45° angle. Deviations from that shape mean the driver has a problem.
You're probably further ahead checking each driver for linearity before doing anything. You'll need a dual trace scope, variable generator with high output or a small amplifier and oscillator, and a 1 K resistor (approximately). You can then accurately get the resonant frequency for each driver. The entire box as well. Put the scope in X-Y mode, adjust for a usable display and look a the pattern. It should look like a propeller on a 45° angle. Deviations from that shape mean the driver has a problem.
I repaired many speaker crossovers and it's mostly due to dried out or leaky Elco caps that they fail. For foil or film caps it's rare that they go far off spec, except with polysterene caps. Inductors are almost never a problem, very old (mostly carbon) resistors tend also to deviate over time, but less frequent that elco's.
If you repair a crossover, you should always measure the drivers in the box to do, and use the old crossover as inspiration to get close. But as drivers (especially cone drivers) age over decades, the t/s parameters and frequency response changes and you need often adapt the crossover to that. Just replacing an old elco with a modern film cap without measuring and simulating is mostly a receipe for bad sound.
If you repair a crossover, you should always measure the drivers in the box to do, and use the old crossover as inspiration to get close. But as drivers (especially cone drivers) age over decades, the t/s parameters and frequency response changes and you need often adapt the crossover to that. Just replacing an old elco with a modern film cap without measuring and simulating is mostly a receipe for bad sound.
Well, if the T/S parameters shift on the woofer, it may well be new woofer or new box time. With mids and tweeters we avoid their resonance, so efficiency is probably more important. That would be magnet strength. Not unless they used ferrofluid (unfortunate). You can clean and put more fluid in. I heard of a gas filled tweeter (stupid beyond belief), have fun with those. They'll just leak out again anyway.
Woofer mass won't change much, I guess a paper cone could really dry out, but then it's ready to literally fall apart. The only two things that will change is magnet strength, and the spider in the woofer. Either one is bad news, a full recone fixes the spider. Remagnetizing a woofer is possible if you know what the original flux was, but then it will probably fall as it did before. New woofer, or entirely new speakers.
Woofer mass won't change much, I guess a paper cone could really dry out, but then it's ready to literally fall apart. The only two things that will change is magnet strength, and the spider in the woofer. Either one is bad news, a full recone fixes the spider. Remagnetizing a woofer is possible if you know what the original flux was, but then it will probably fall as it did before. New woofer, or entirely new speakers.
How is plagued capacitance for bipolar in loudspeaker filter when agging, please?
I read it lowish when old, then when very too old, capacitance has a tendancy to increase !
Same aging speed if in serie or shunt as well as same capacitance changes behaviors ?
It is always a mess to refurbish loudspeakers, after 30 yo, you never know about the exact values whatever the manual of the speaper. I once refurbished a Kef 104/2 filter... nigthmare, when with such filter capacitance are often sorted out less than 2% in some of them !
I read it lowish when old, then when very too old, capacitance has a tendancy to increase !
Same aging speed if in serie or shunt as well as same capacitance changes behaviors ?
It is always a mess to refurbish loudspeakers, after 30 yo, you never know about the exact values whatever the manual of the speaper. I once refurbished a Kef 104/2 filter... nigthmare, when with such filter capacitance are often sorted out less than 2% in some of them !
Electrolytic capacitors tend to "dry out" through the rubber end seal. When you lose electrolyte the capacitance goes down.
I remember testing capacitors "in the old days". They were often high in value. These days to save money, manufacturers often make parts at the bottom end of tolerance. Sometimes they are out of tolerance since few of us check new parts going in. I do since this takes less time than troubleshooting problems later.
I remember testing capacitors "in the old days". They were often high in value. These days to save money, manufacturers often make parts at the bottom end of tolerance. Sometimes they are out of tolerance since few of us check new parts going in. I do since this takes less time than troubleshooting problems later.
Very interesting,
could you please share some theory-to-practice guide to perform such measurements? I use REW for room correction measurements (but apply and fine-tune parametric EQ manually by ear), have no idea how to perform impedance test with that software.
Best,
Mike
Well 😉
in case of my ReVox Emporium B speakers it was exactly the matter of resonance in the midrange driver. Hardened rubber suspension. I had an idea of applying some "clever" chemicals like brake fluid or tire dressing but finally decided to replace the suspension with the new one. But the new made thinks only worse as resonance moved even more into driver's working range. It ended up with something insane but working: to move resonance frequency, the speaker has now two rubber suspension rings glued one on another. Of course the glue was applied on the outer and inner surface only, the working area remains unglued. It's more of a MacGyver method than purist's choice, but it works. The alternatives were to spray the cone with some heavy painting to add mass or to give up as it was pure abstraction to unglue new rubber from the paper cone. Moreover - I found stiffer rubber suspension, but the size was not exactly right as the speaker was made of plastic formed precisely for the original suspension diameter.
This all sounds as crazy as it is but since then I haven't managed to find a second hand or new-old-stock midrange of this type. To my very luck, one midrange had pretty elastic rubber suspension and only the second hardened, probably because one speaker had been placed on sunny place for years (direct exposure - e.g. window). So I had something to compare the wrong driver with. Compare "more wrong" with "less wrong". After fitting the bad performing one with new double rubber suspension, both drivers achieved their resonance at almost the same frequency in a close mic free air test. But still it is pure joke as I do not have the new original to compare with. For now, both midrange drivers have double rubber suspension. It can cause additional friction or even generate some noise or distortion but that's still better than having a resonance within the driver working range.
And now the best part...
Before all the work on Emporium B I used to think that most vintage speaker servicing guys do extensive testing of every altered component to at least avoid a complete mess in cabinet characteristics. But sadly, they often replace old or very old rubber suspension with the generic new one, put the driver back to the cabinet and sale the pair of speakers as "fully serviced vintage classics". As I discovered, new generic rubber suspension itself can ruin the parameters of the speaker in the cabinet. And what if we add new spider, new cone?
That all means: getting old speakers to work like new is virtually impossible task when we need to replace driver's parts with new generic ones. It's more of a lottery than an honest fix. We can always say that we try "getting close". But close to what? Most vintage speakers that I know are far from being linear. They have a strong character and add their own sound signature. So searching for linearity is not enough 😉
What I'd learned the most from the Emporium B recapping/resealing/rerubbering and re-everything was the importance of the pneumatic domain inside the speaker cabinet. Before that I used to think "it's sealed enough". But step by step, searching for every place suspected of even the smallest airflow, the work ended with a low end so powerful and precise that's worth every working hour and research.
It's a closed cabinet. After pushing or connecting to small direct current, the bass driver cones returned lazy to the idle position. Now they return "a bit more" lazy. But acoustically that makes a huge difference. I encourage every owner of closed cabinet speakers to seal them precisely wherever possible, inside and outside. Cabinet from the inside, cables, terminals. Midrange resonance chamber. Some cabinets are divided as a whole for two bass drivers, and the dividing wall must be perfectly sealed. Even drivers can be prone to parasite airflow, e.g. between dust cup and cone. Pumping compressed air through the smallest hole is a loose of energy as the pressure is not being returned to the driver. Do seal everything. It can do miracles.
That's a kind of recommendation I can share as a beginner to all the beginners.
What a great forum is that 😉
Last edited:
For sure!
I had to send so many recones back for rework it wasn't funny. They got to know I actually checked their work. Most times a recone is installed and the customer may not notice a problem until past the warranty period.
Always check the resonance, free air resonance and linearity. You will hear a buzz if you have other problems doing these tests.
Now here is a shocker for you. Most equipment, electronic or mechanical (incl speakers) are very poorly serviced. More often than not, just made to "make noise" and sold as rebuilt or checked okay by a tech. I can't tell you how much equipment I see "just serviced" with serious problems. Many of those problems induced by the "technicians" that certified the equipment working well.
I had to send so many recones back for rework it wasn't funny. They got to know I actually checked their work. Most times a recone is installed and the customer may not notice a problem until past the warranty period.
Always check the resonance, free air resonance and linearity. You will hear a buzz if you have other problems doing these tests.
Now here is a shocker for you. Most equipment, electronic or mechanical (incl speakers) are very poorly serviced. More often than not, just made to "make noise" and sold as rebuilt or checked okay by a tech. I can't tell you how much equipment I see "just serviced" with serious problems. Many of those problems induced by the "technicians" that certified the equipment working well.
I am running an accelerated agening test of X and Y caps as we speak, they are biased to their rated voltage and they are placed in a humiditu chamber and tested according to ISO 16750 (with a supplement that defines some parameters).
After a couple of hundred hours the ESR starts to increase, capacitance is not changed. After 450 h the capacitance is down 5% and ESR is up >1000%!
So measuring ESR is the best way to determine the agening of film capacitor since the degradation starts with oxidation of the metal, but the metal area stays the same and not affecting the capacitance.
So conclusion is that ESR is the best parameter to check in old capacitors.
After a couple of hundred hours the ESR starts to increase, capacitance is not changed. After 450 h the capacitance is down 5% and ESR is up >1000%!
So measuring ESR is the best way to determine the agening of film capacitor since the degradation starts with oxidation of the metal, but the metal area stays the same and not affecting the capacitance.
So conclusion is that ESR is the best parameter to check in old capacitors.