• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Capacitor upgrade in crossover - Is it audible? - Part 2

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
Here is the second part on the topic "Capacitor upgrade in crossover - Is it audible?
This time it is exclusively about the measurable differences of electrolytic capacitors with film capacitors in crossovers of loudspeakers.

The first part about film capacitors can be found here and who likes can read about the measurable differences between sand-cast and low-inductance resistors here.
The third and final part on this topic: Capacitor upgrade in crossover - You Can't Handle the Truth - part-3


The full mini series:
Capacitor upgrade in crossover - Is it audible?
Capacitor upgrade - part two
Capacitor upgrade - part three
Replace resistor by low-inductance resistor - Is it audible?
Different Binding Posts - is it audible?
Audible difference in high-end capacitors? - ABX samples


The test method and test setup were described in detail in the first part. If you're not familiar with what a capacitor (in series with the driver) does, you can read this in part one either - I'm too lazy to repeat all this. ;)
1586217477281.png

As in part 1, the capacitor is connected in series to the chassis and then measured in the near field.

With almost the same capacitance of the capacitors, the deviations then measured are solely due to the different design/type of the capacitors or of minimal deviation in capacity - there is also a section on this in Part 1.

Electrolytic capacitor versus film capacitor
To obtain more certainty about the accuracy of the measurements, comparisons were made with three different capacitance values - 8.2µF, 47µF and 100µF.
High capacities in series to the driver are mostly used in the mid-range crossover. Therefore the measurement for 100µF was carried out with a mid-range driver.

a) Comparison with 8.2µF capacity
1586159791108.png

The film capacitor has a capacitance of 8.47µF, the electrolytic capacitor one of 8.46µF. The deviation of the capacitors from each other is therefore only 0.1%. According to the investigations of capacitors with low capacitance deviations made in part 1, the measured deviation should be only minimal.

noname MKT
1586220067595.png


Audyn-Cap electrolytic capacitor
1586220094413.png


It is easy to see that the resistive part of the electrolytic capacitor is four times as high as that of the film capacitor.

Let's take a look at the frequency response measurements.
1648418508006.png


A very small difference can be seen. To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
1648418552693.png


The deviation of the electrolytic capacitor is greater than we would expect due to the small difference in capacitance - for the effects of minimal capacitance deviations see part 1.
Even at high frequencies the deviation is still almost 0.1dB. Between 3 - 6kHz the deviation is almost about 0.15dB.

Here, the high ESR of the electrolytic capacitor has a clear effect on the frequency response of the tweeter.



b) Comparison with 47µF capacity
1586177345584.png

The MKT 250V film capacitor has a capacity of 46.7µF, the electrolytic capacitor with 100V dielectric strength has a capacity of 46.9µF.
The capacitance deviation is therefore about 0.4%.

electrolytic capacitor
1586221247620.png


noname MKT film capacitor
1586221296957.png

The resistive part of the electrolytic capacitor is five times as high as that of the film capacitor.

Due to the differences in capacitance, one would expect that measured frequency response of the electrolytic capacitor would be slightly above that of the film capacitor. However, due to the high ESR of the electrolytic capacitor, this is not the case here either.
1648420005335.png

To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
1648420061590.png

Film capacitor and electrolytic capacitor behave as in the example above.
The maximum sound pressure difference in this case is 0.2dB.



c) Comparison with 100µF capacity
Here, an additional comparison of an old and a new electrolytic capacitor is to be carried out.
1586206031477.png

Starting point is a Visaton electrolytic capacitor from 1985 (+-3 years) with 100µF nominal capacity and +-10% tolerance. The measured capacitance of the Visaton electrolytic capacitor is about 111.5µF.
1586219609319.png


Next are two new electrolytic capacitors (100µF + 8.2µF, +-10% tolerance capacitors) which together have a capacity of 111.5µF.
1586219631425.png


Finally a combination of film capacitors (100µF + 10µF + 0.82µF, +-5% tolerance capacitors) with a total of 111.3µF.
1586219645685.png

The deviation of the capacitors compared to the film capacitor is only 0.2%.
The resistive part of the 35 year old Visaton electrolytic capacitor is only two times as high as that of the film capacitor.
The resistive part of the new electrolytic capacitor is three times as high as that of the film capacitor.

As already mentioned above, such high capacitance values in series to a chassis actually only occur with bass/midrange drivers.
For this reason, this test setup uses a midrange driver instead of a tweeter - the BMS 6S117 midrange driver. The driver was chosen because it has triple aluminum demodulating rings and therefore low distortion. This should make it easier to detect possible distortion caused by capacitors.
The rest of the test setup remains unchanged.

For the evaluation we consider the reasonable frequency range of the chassis (before cone-breakup). Therefore all frequency response diagrams consider the range 100Hz to 3kHz.
1648418727610.png


You know the drill... To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
1648418795274.png

The result is interesting because it shows that the quality of the electrolytic capacitor is also very important.
The 35 year old Visaton electrolytic capacitor has a lower ESR than the combination of two newly purchased electrolytic capacitors. Accordingly, the Visaton cap performs better than the newly purchased electrolytic capacitors compared to the film capacitor.

Nevertheless, the deviations in the frequency response with a maximum of 0.1dB are so small that this should be practically inaudible when replacing the capacitor in the crossover in e.g. a 3-way loudspeaker.



d) Multiton Distortion - Film versus Electrolytic capacitor
Because people keep asking for it, here is the multitone distortion, which contain the harmonic distortion and IMD, of all three 111µF capacitor combinations from section c) in series to the mid range driver.
1586224803464.png


The distortion attenuation is sometimes more than -60dB (which corresponds to a distortion factor of less than 0.1%), and yet no different contribution of the capacitors can be seen in the multitone distortion.



Conclusion

For the same capacitance, there are measurable differences when comparing electrolytic capacitors against film capacitors (in series with the chassis), which depend strongly on the quality, especially the ESR, of the electrolytic capacitor.

However, the difference in sound pressure in the measurements shown is only 0.2dB in the worst case.
This is higher than in the first part when comparing normal film capacitors against a high-end capacitor, but in the end still insignificant (in most cases).

If the electrolytic capacitor is of good quality, durability does not play a decisive role either.

Only the high tolerance values of electrolytic capacitors and the possible change of the capacitance value over time could pose a problem. Especially when used in the tweeter crossover - this should be avoided.

Very important: according to these measurements a film capacitor does not make a better sound when used in the crossover compared to an electrolytic capacitor.

If you replace capacitors, especially electrolytic capacitors, with film capacitors in the crossover, without checking with an LCR meter, you will most likely notice a change in sound. However, 99% of the change in sound is only due to the different capacitances caused by the capacitor tolerances and not because the film or high-end capacitor "sounds better".



UPDATE:

e) Aged electrolytic capacitors

In my stock are four 100µF (+-10% tolerance, 35 VAC) Visaton electrolytic capacitors from 1985 (+-3 years). These were installed in a crossover for over twenty years and were in regular use. The last ten years they were stored in my stock and only used sporadically.
1586261713582.png


One of these capacitors was already measured in part c). Out of curiosity I measured all four capacitors. I am a bit surprised that at first sight all four are still working.
1586261651692.png

The measured capacitances@1kHz were between 111µF and 116µF. Thus all electrolytic capacitors slightly exceeded the permissible upper tolerance range. This is most likely due to ageing effects.

I can't say if the full dielectric strength of the capacitors is still given, but I haven't noticed any disadvantages even at high sound pressure levels when they were used in a crossover.

The third and final part on this topic: Capacitor upgrade in crossover - You Can't Handle the Truth - part-3


Update 2022-03-23: Fixed diagrams with corrupted scaling.
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
It is also clear that compounding capacitors doesn't help.

I often see it mentioned that coupling caps (those in series with the audio signal) that compounding (paralleling) helps.
It can't and this test shows it as well.
So if someone has a 10μ capacitor and parallels 0.1uF film cap this does not help for higher frequencies because it's impedance will always be 100x higher than that of the larger cap (in the considered audible range) so the vast majority of the current will always pass through the larger value cap.

Compounding caps DOES help with decoupling power pins but this is because this works in the MHz range (smaller value caps have lower ESR at higher frequencies).
 

sonci

Active Member
Joined
Feb 18, 2018
Messages
233
Likes
112
Whenever I replaced electrolytics capacitors to film, speakers became overly bright, after some time I have to put back the original caps.
I guess every serious manufacture account for the original ESR of the cap, if replacing you should add a small resistor in series with the film cap, but IMO its better to design a new crossover than to just upgrade components.
 
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU

VeerK

Active Member
Joined
Jan 29, 2019
Messages
257
Likes
318
Location
NYC
Added a small section on aged electrolytic capacitors.
Hope this covers most of your wishes now - I've had enough!;)

Hell of a job, thanks for taking the time to put in all this work. I’ve been DIYing speakers for a few years now so I’ve seen all the hubbub over different capacitors and their magical abilities. Nice to see some measurements :)
 

Ashley Salmond

Member
Forum Donor
Joined
Jan 17, 2021
Messages
58
Likes
12
Location
Toowoomba , Australia
Here is the second part on the topic "Capacitor upgrade in crossover - Is it audible?
This time it is exclusively about the measurable differences of electrolytic capacitors with film capacitors in crossovers of loudspeakers.

The first part about film capacitors can be found here and who likes can read about the measurable differences between sand-cast and low-inductance resistors here.
The third and final part on this topic: Capacitor upgrade in crossover - You Can't Handle the Truth - part-3

The test method and test setup were described in detail in the first part. If you're not familiar with what a capacitor (in series with the driver) does, you can read this in part one either - I'm too lazy to repeat all this. ;)
View attachment 57559
As in part 1, the capacitor is connected in series to the chassis and then measured in the near field.

With almost the same capacitance of the capacitors, the deviations then measured are solely due to the different design/type of the capacitors or of minimal deviation in capacity - there is also a section on this in Part 1.

Electrolytic capacitor versus film capacitor
To obtain more certainty about the accuracy of the measurements, comparisons were made with three different capacitance values - 8.2µF, 47µF and 100µF.
High capacities in series to the driver are mostly used in the mid-range crossover. Therefore the measurement for 100µF was carried out with a mid-range driver.

a) Comparison with 8.2µF capacity
View attachment 57386
The film capacitor has a capacitance of 8.47µF, the electrolytic capacitor one of 8.46µF. The deviation of the capacitors from each other is therefore only 0.1%. According to the investigations of capacitors with low capacitance deviations made in part 1, the measured deviation should be only minimal.

noname MKT
View attachment 57567

Audyn-Cap electrolytic capacitor
View attachment 57568

It is easy to see that the resistive part of the electrolytic capacitor is four times as high as that of the film capacitor.

Let's take a look at the frequency response measurements.
View attachment 57569

A very small difference can be seen. To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
View attachment 57570


The deviation of the electrolytic capacitor is greater than we would expect due to the small difference in capacitance - for the effects of minimal capacitance deviations see part 1.
Even at high frequencies the deviation is still almost 0.1dB. Between 3 - 6kHz the deviation is almost 0.15dB.

Here, the high ESR of the electrolytic capacitor has a clear effect on the frequency response of the tweeter.



b) Comparison with 47µF capacity
View attachment 57421
The MKT 250V film capacitor has a capacity of 46.7µF, the electrolytic capacitor with 100V dielectric strength has a capacity of 46.9µF.
The capacitance deviation is therefore about 0.4%.

electrolytic capacitor
View attachment 57572

noname MKT film capacitor
View attachment 57573
The resistive part of the electrolytic capacitor is five times as high as that of the film capacitor.

Due to the differences in capacitance, one would expect that measured frequency response of the electrolytic capacitor would be slightly above that of the film capacitor. However, due to the high ESR of the electrolytic capacitor, this is not the case here either.
View attachment 57575
To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
View attachment 57576
Film capacitor and electrolytic capacitor behave as in the example above.
The maximum sound pressure difference in this case is 0.2dB.



c) Comparison with 100µF capacity
Here, an additional comparison of an old and a new electrolytic capacitor is to be carried out.
View attachment 57531
Starting point is a Visaton electrolytic capacitor from 1985 (+-3 years) with 100µF nominal capacity and +-10% tolerance. The measured capacitance of the Visaton electrolytic capacitor is about 111.5µF.
View attachment 57564

Next are two new electrolytic capacitors (100µF + 8.2µF, +-10% tolerance capacitors) which together have a capacity of 111.5µF.
View attachment 57565

Finally a combination of film capacitors (100µF + 10µF + 0.82µF, +-5% tolerance capacitors) with a total of 111.3µF.
View attachment 57566
The deviation of the capacitors compared to the film capacitor is only 0.2%.
The resistive part of the 35 year old Visaton electrolytic capacitor is only two times as high as that of the film capacitor.
The resistive part of the new electrolytic capacitor is three times as high as that of the film capacitor.

As already mentioned above, such high capacitance values in series to a chassis actually only occur with bass/midrange drivers.
For this reason, this test setup uses a midrange driver instead of a tweeter - the BMS 6S117 midrange driver. The driver was chosen because it has triple aluminum demodulating rings and therefore low distortion. This should make it easier to detect possible distortion caused by capacitors.
The rest of the test setup remains unchanged.

For the evaluation we consider the reasonable frequency range of the chassis (before cone-breakup). Therefore all frequency response diagrams consider the range 100Hz to 3kHz.
View attachment 57579

You know the drill... To be able to see the difference better, the diagram is normalized to the frequency response of the MKT capacitor and displayed with a scaling of 0.1dB.
View attachment 57580
The result is interesting because it shows that the quality of the electrolytic capacitor is also very important.
The 35 year old Visaton electrolytic capacitor has a lower ESR than the combination of two newly purchased electrolytic capacitors. Accordingly, the Visaton cap performs better than the newly purchased electrolytic capacitors compared to the film capacitor.

Nevertheless, the deviations in the frequency response with a maximum of 0.1dB are so small that this should be practically inaudible when replacing the capacitor in the crossover in e.g. a 3-way loudspeaker.



d) Multiton Distortion - Film versus Electrolytic capacitor
Because people keep asking for it, here is the multitone distortion, which contain the harmonic distortion and IMD, of all three 111µF capacitor combinations from section c) in series to the mid range driver.
View attachment 57582

The distortion attenuation is sometimes more than -60dB (which corresponds to a distortion factor of less than 0.1%), and yet no different contribution of the capacitors can be seen in the multitone distortion.



Conclusion

For the same capacitance, there are measurable differences when comparing electrolytic capacitors against film capacitors (in series with the chassis), which depend strongly on the quality, especially the ESR, of the electrolytic capacitor.

However, the difference in sound pressure in the measurements shown is only 0.2dB in the worst case.
This is higher than in the first part when comparing normal film capacitors against a high-end capacitor, but in the end still insignificant (in most cases).

If the electrolytic capacitor is of good quality, durability does not play a decisive role either.

Only the high tolerance values of electrolytic capacitors and the possible change of the capacitance value over time could pose a problem. Especially when used in the tweeter crossover - this should be avoided.

Very important: according to these measurements a film capacitor does not make a better sound when used in the crossover compared to an electrolytic capacitor.

If you replace capacitors, especially electrolytic capacitors, with film capacitors in the crossover, without checking with an LCR meter, you will most likely notice a change in sound. However, 99% of the change in sound is only due to the different capacitances caused by the capacitor tolerances and not because the film or high-end capacitor "sounds better".



UPDATE:

e) Aged electrolytic capacitors

In my stock are four 100µF (+-10% tolerance, 35 VAC) Visaton electrolytic capacitors from 1985 (+-3 years). These were installed in a crossover for over twenty years and were in regular use. The last ten years they were stored in my stock and only used sporadically.
View attachment 57640

One of these capacitors was already measured in part c). Out of curiosity I measured all four capacitors. I am a bit surprised that at first sight all four are still working.
View attachment 57639
The measured capacitances@1kHz were between 111µF and 116µF. Thus all electrolytic capacitors slightly exceeded the permissible upper tolerance range. This is most likely due to ageing effects.

I can't say if the full dielectric strength of the capacitors is still given, but I haven't noticed any disadvantages even at high sound pressure levels when they were used in a crossover.

The third and final part on this topic: Capacitor upgrade in crossover - You Can't Handle the Truth - part-3
Hi, so you are saying there is no difference between expensive and cheaper capacitors.
 
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
Hi, so you are saying there is no difference between expensive and cheaper capacitors.
So summarized, no I do not claim that.

There are clearly measurable differences, so the resistive part of the electrolytic capacitor (very cheap) is up to three times as high as that of the film capacitor (more expensive) in my examples.

However, the effects of these measurable differences hardly make a measurable "acoustic" difference when used in the crossover of loudspeakers.
The impulse response, the frequency response and the IMD change only minimally (hardly measurable) in the application range of the capacitors - that is for large capacitances (>80µF) in the bass/midrange and capacitances <20µF in the high frequency range.

When moving from the low kHz range to the MHz range, the differences can be dramatic, but this is irrelevant for LS crossover.

The "acoustically" measurable changes caused by possible component tolerances (mostly at +-5% or +-10%, rarely +-2%), on the other hand, are significantly higher and can be so large that audible differences can occur when capacitors are replaced - for more detail see part one, section"2) Capacitors - low capacitance deviations"
.
 

Ashley Salmond

Member
Forum Donor
Joined
Jan 17, 2021
Messages
58
Likes
12
Location
Toowoomba , Australia
So then Geert, you are saying there is a noticeable difference in sound with the film capacitors. I am getting ready to build some speakers with the GR Research A/V2. And since I was going to the trouble I thought I may improve upon the quality of the erse capacitors. What would you recommend in the film type.
 

Ashley Salmond

Member
Forum Donor
Joined
Jan 17, 2021
Messages
58
Likes
12
Location
Toowoomba , Australia
So summarized, no I do not claim that.

There are clearly measurable differences, so the resistive part of the electrolytic capacitor (very cheap) is up to three times as high as that of the film capacitor (more expensive) in my examples.

However, the effects of these measurable differences hardly make a measurable "acoustic" difference when used in the crossover of loudspeakers.
The impulse response, the frequency response and the IMD change only minimally (hardly measurable) in the application range of the capacitors - that is for large capacitances (>80µF) in the bass/midrange and capacitances <20µF in the high frequency range.

When moving from the low kHz range to the MHz range, the differences can be dramatic, but this is irrelevant for LS crossover.

The "acoustically" measurable changes caused by possible component tolerances (mostly at +-5% or +-10%, rarely +-2%), on the other hand, are significantly higher and can be so large that audible differences can occur when capacitors are replaced - for more detail see part one, section"2) Capacitors - low capacitance deviations"
.
Thank you for that informative reply. So as far as sound goes there is little difference between the electrolytic and film capacitors. It's just that the more expensive are more stable and consistent and hold there values longer
So summarized, no I do not claim that.

There are clearly measurable differences, so the resistive part of the electrolytic capacitor (very cheap) is up to three times as high as that of the film capacitor (more expensive) in my examples.

However, the effects of these measurable differences hardly make a measurable "acoustic" difference when used in the crossover of loudspeakers.
The impulse response, the frequency response and the IMD change only minimally (hardly measurable) in the application range of the capacitors - that is for large capacitances (>80µF) in the bass/midrange and capacitances <20µF in the high frequency range.

When moving from the low kHz range to the MHz range, the differences can be dramatic, but this is irrelevant for LS crossover.

The "acoustically" measurable changes caused by possible component tolerances (mostly at +-5% or +-10%, rarely +-2%), on the other hand, are significantly higher and can be so large that audible differences can occur when capacitors are replaced - for more detail see part one, section"2) Capacitors - low capacitance deviations"
.
So summarized, no I do not claim that.

There are clearly measurable differences, so the resistive part of the electrolytic capacitor (very cheap) is up to three times as high as that of the film capacitor (more expensive) in my examples.

However, the effects of these measurable differences hardly make a measurable "acoustic" difference when used in the crossover of loudspeakers.
The impulse response, the frequency response and the IMD change only minimally (hardly measurable) in the application range of the capacitors - that is for large capacitances (>80µF) in the bass/midrange and capacitances <20µF in the high frequency range.

When moving from the low kHz range to the MHz range, the differences can be dramatic, but this is irrelevant for LS crossover.

The "acoustically" measurable changes caused by possible component tolerances (mostly at +-5% or +-10%, rarely +-2%), on the other hand, are significantly higher and can be so large that audible differences can occur when capacitors are replaced - for more detail see part one, section"2) Capacitors - low capacitance deviations"
.
Thank you Greet, that is a very good and clear information. So to sum up, the more expensive film capacitors don't so any measurable difference just anecdotal. That the electrolytic don't last as well and change there values.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
May I show a different measurement, please. As the capacitors discussed are used in a crossover, they have to be investigated for distortion when loaded.

Two quickly made measurements with
1) Hitano electrolytic capacitor 22uF/63V loaded with 6.8 ohm resistor
2) Monacor MKT capacitor 22uF/250V loaded with 6.8 ohm resistor

Distortion was measured at 1kHz and 9.5V amplifier output, distortion measured at 6.8 ohm load.

Electrolytic capacitor + resistor 6.8 ohm (sorry for the typo in the graph)
elyt22uF_6R8_highpass.png



MKT capacitor + resistor 6.8 ohm (sorry for the typo in the graph)
MKT22uF_6R8_highpass.png


With MKT capacitor the distortion is almost 10x lower. In my opinion, no serious designer would use electrolytic capacitors in a high quality speaker design. The electrolytic capacitor is stressed and shows quite high odd harmonics distortion.
Please forgive the mains spuriae, the setup was done very quickly.

Edit: I have just measured the distortion with the capacitor shorted and it is the same as with the MKT capacitor. So the MKT capacitor has not increased the harmonic distortion in this test setup.
 
Last edited:

Ashley Salmond

Member
Forum Donor
Joined
Jan 17, 2021
Messages
58
Likes
12
Location
Toowoomba , Australia
May I show a different measurement, please. As the capacitors discussed are used in a crossover, they have to be investigated for distortion when loaded.

Two quickly made measurements with
1) electrolytic capacitor 22uF/63V loaded with 6.8 resistor
2) MKT capacitor 22uF/250V loaded with 6.8 resistor

Distortion was measured at 1kHz and 9.5V amplifier output, distortion measured at 6.8 ohm load.

Electrolytic capacitor
View attachment 136777


MKT capacitor
View attachment 136778

With MKT capacitor the distortion is almost 10x lower. In my opinion, no serious designer would use electrolytic capacitors in a high quality speaker design. The electrolytic capacitor is stressed and shows quite high odd harmonics distortion.
Please forgive the mains spuriae, the setup was done very quickly.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
PMA, thank you so much for that detailed insight. I am so impressed by your knowledge and generous nature.

Thanks for your kind words. I have done these tests many times, with different parts, so I knew the result beforehand. Capacitor distortion is to be examined when the capacitor is loaded, otherwise one would measure no difference against foil capacitors.

I suggest to google for Cyril Bateman capacitor measurements, he has done a lot of work in this field.
It may be suggested that the distortion shown might be inaudible, however the measurement clearly shows the improper use of the electrolytic capacitor. They are not built to handle AC voltage across their terminals.
 
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
Distortion was measured at 1kHz and 9.5V amplifier output, distortion measured at 6.8 ohm load.
Can you please do the test under more realistic conditions and measure at 2.83V.
A normal 5-6'' bass-midrange driver already delivers 86-93dB@1m SPL at 2.83V voltage.

Your test conditions (9.5V) reflect a sound pressure level of 97-104dB@1m SPL. This is a sound pressure level range where harmonic distortion is almost irrelevant due to the distortion created by the human ear. Anything below single digit percentages is irrelevant.
So whether the electrolytic capacitor shows 0.05% or 0.5% HD3 is completely irrelevant at such sound pressure levels.

As you said, these are similar measurements that Cyril Bateman made and are probably the cause of the "commercialism" and uproar about capacitor replacement in loudspeaker crossovers.

The todays best 5-6'' drivers show about 0.01%-0.03% third order harmonic distortion at 87-90dB@1m@1kHz - that means at 2.83V.
An average 5-6'' driver is more like 0.5% to 1% HD3 at 87-90dB@1m@1kHz.
Even your unrealistic measurements at 9.5V with 0.05% HD3 for the electrolytic capacitor do not play a role when using an average driver.
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Can you please do the test under more realistic conditions and measure at 2.83V.
A normal 5-6'' bass-midrange driver already delivers 86-93dB@1m SPL at 2.83V voltage.

Yes, as long as the primitive test setup has not been disconnected yet.

Electrolytic capacitor 22uF + 6.8 ohm
elyt22uF_6R8_highpass_2.8V.png


MKT 22uF + 6.8 ohm
MKT22uF_6R8_highpass_2.8V.png


Some remarks: as the MKT capacitor is much bigger and has long leads it catches more mains spuriae in my primitive provisional setup.
Second, again the MKT does not add distortion to the test setup. With the electrolytic cap, we can see 3rd harmonic rising. 2nd is lower not because it is not there, but because of the distortion cancellation of the H2 between the capacitor and the imperfect test setup.

Last, I understand your reasoning but I do not agree. To me, electrolytic capacitors should not be used in speaker crossover circuits. It is no marketing, they just are not good and not intended to be used in this way. See the rated ripple current of the capacitors. And they are nonlinear and have reduced life time under such conditions. When I see electrolytic capacitors in speaker crossovers, my first impression is that someone wants to save money and increase profit. But this forum is mostly about cheap solutions, I understand. Each to his own.

Last2, I do not think that 9.5V amplifier output is something unusual. Many speakers are about 85dB/2.83V/m.
 
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,086
Location
.de, DE, DEU
@pma much thanks for the new measurements at 2.8V

So, now we are at a sound pressure level of 86-93dB for 5-6'' drivers (depending on how the driver was designed).

As mentioned before, the world's best drivers show 0.01% to 0.03% HD3 at 2.83V@1kHz (roughly 86-90dB).
The electrolytic capacitor causes another 0.01% HD3 (the film capacitor much less), which is completely irrelevant - whether 0.01% or 0.02% HD3, no human can perceive that while listening to normal music.

One should not forget that "normal" good drivers show 0.3% to 0.7% HD3 at 2.83V or roughly 86-90dB@1m.
So by a factor of 30 to 70 higher distortion values than the electrolytic capacitor.

This is now also in good agreement with my measurements, which were also performed at 2-2.8V and one can see no difference in HD measurements or IMD measurements between electrolytic capacitors and film capacitors when measuring via microphone directly at the driver.

Actually, one would have to go even lower with the voltage, since the human ear, at even lower sound pressure levels, shows less masking and is therefore more sensitive to harmonic distortion.


Last, I understand your reasoning but I do not agree. To me, electrolytic capacitors should not be used in speaker crossover circuits. It is no marketing, they just are not good and not intended to be used in this way. See the rated ripple current of the capacitors. And they are nonlinear and have reduced life time under such conditions. When I see electrolytic capacitors in speaker crossovers, my first impression is that someone wants to save money and increase profit. But this forum is mostly about cheap solutions, I understand. Each to his own.
If you want only the best of the best, you naturally do not use electrolytic capacitors.
And as I already wrote in the start post, I would also not use electrolytic capacitors in the tweeter filter branch of the crossover.

But to replace, in the midrange and low frequency range, electrolytic capacitors with large capacitances with film capacitors makes little sense for 99% of current speakers, in terms of harmonic distortion, because the "normal" drivers in LS have many times more distortion - apart from the world's best drivers at the moment, where it can be argued that the use of film capacitors further optimizes the already excellent harmonic distortion.

Capacitances around 100-200µF in the low and midrange are not uncommon for 3-way loudspeakers.
In terms of price, bipolar electrolytic capacitors are 4-5€, the film capacitors with capacitances of 100-200µF, however, cost 30-90€ each (mind you standard film capacitors, if you want "high-end" you have to sacrifice your pension).
For 0.01% less HD3 this is a very high price.
 
Last edited:
Top Bottom