Speaker Recap - Electrolytic or Film?

[corbin]

Any thoughts on electrolytic vs film caps in speaker crossovers? I know the conventional wisdom is to go with the higher quality film cap, but I'm hesitant to do so because the reduced ESR will pass more high frequencies, altering the tonal balance of the speaker.

The speakers in question are EPI 70s, which have the same single 10uF cap as the EPI 100s.

Here are the options:

1) 10uF electrolytic, same as original
2) 10uF film cap
3) 10uF film cap with 2 ohm resistor (?) to match original ESR (edited)

It would seem #3 is the best option as it preserves the original ESR while taking advantage of the superior characteristics of film caps.

Any thoughts? There was a long thread on this issue a while back but I can't seem to find it now.

My EPI 70s have the original electrolytic caps and sound veiled in comparison to my BA A150s - I think it's the caps. Looking forward to hearing what difference recapping makes.

 

[nfafan]

Good thread, watching. Always wondered about this myself; should we try to retain the tone/voice that won our hearts and minds years ago when we first heard them in the "stereo store" - or replace the caps with new/modern which could change all that.

 

[corbin]

To me there's no question that the original voice should be retained, not because it's nostalgic, but because it's how the designer specified it to sound as a whole - with that specific cabinet material, internal damping, driver characteristics, etc. In the case of EPI speakers, where a lot of effort was put into how they sound by their designer, it seems unlikely that a significant improvement would result from casually altering critical design parameters such as capacitor value.

 

[TPettenati]

When I replaced the stock electrolytic with a Dayton poly on my EPI 100V's, I found them to be too bright.
So I added a pot for tweeter level control, as can be seen in one of the crossover schematics on Human Speakers.

Human Speakers EPI 100

 

[StimpyWan]

3) 10uF film cap with another .3uF (?) film cap to match original ESR

Why would paralleled 10.3uF capacitors match the ESR of the original design? Did you meant to say that by adding a 0.3 Ohm resistor, to the 10uF film cap, you would then match to stock?

Personally, I like film. But, if I went with NPE's, I'd add a 0.1uF film bypass cap. To me a bypass capacitor improves the sound. Some feel different, though I haven't heard anything negative from using bypass caps, on my speakers.

 

[corbin]

Why would paralleled 10.3uF capacitors match the ESR of the original design? Did you meant to say that by adding a 0.3 Ohm resistor, to the 10uF film cap, you would then match to stock?

Thanks for correction, I changed the post. Frankly I know next to nothing about electrical engineering, I'm not sure how to retain original ESR with a film cap, but it would seem that would be a worthy goal. Open to more suggestions/thoughts.

 

[drbiggles]

Hey man,

If you want to use a film cap and simulate the ESR of the original electrolytics, add a 1 or 2ohm (I've found 2ohms work well) 10 watt resistor in series with the driver, closest to the driver, on the + side.

Biggles

 

[StimpyWan]

Thanks for correction, I changed the post. Frankly I know next to nothing about electrical engineering, I'm not sure how to retain original ESR with a film cap, but it would seem that would be a worthy goal. Open to more suggestions/thoughts.

No problem. And the good Dr is correct, the resistor would be in series with the new capacitor. But, from everything I've read and studied, ESR differences typically test less than 1 Ohm. Usually between 0.3 to 0.5 Ohms. So, a huge resistor isn't needed to compensate for ESR differences. Regardless, any capacitor used in the refresh, should make for a nice audible improvement, whether NPE or poly.

 

[canuckaudiog]

@corbin What you've said strikes right at the heart of the issue with replacing capacitors in a crossover - the ESR properties.

Obviously I don't need to explain what can happen when you use a different type of capacitor from the original. Here are my thoughts on the subject (which I have done a lot of reading, experimenting and thinking about).

Crossovers are inherently tricky circuits. Some are more simple than others. With a mass produced, lower quality product where the crossover is more an afterthought than it is a part of the design, especially in simple first or second order crossovers, replacing with polypropylene capacitors can sometimes result in an improvement. But, in cases where the balance of the speaker was carefully planned, with the crossover included in that equation (because let's face it, the crossover is not only a frequency splitter, but also an EQ), it would be detrimental to the designer's intentions to replace it with a wildly different part. And it's not as simple as electrolytic vs. polypropylene/mylar film - even within electrolytics there are different types and they too have different ESR properties. For example, low loss electrolytics have, on average, 1ohm lower ESR than other types.

Could it result in an improvement and be in fact better? The school of thought is that electrolytics are chosen simply because they are lower in cost and the bean counters made the call on that. Yes, lowering the cost of a product when you plan to produce thousands of them is important. However, what most people don't realize is that those choices are incorporated into the designer's plans. The company will go to the designer and tell them, "We want you to build a crossover for X speaker. However, we need to maintain at least this much of a profit margin, or we will lose money selling them, so the overall cost of the crossover must be $X". And then, the designer will incorporate the lower cost parts into their design. When you design a crossover, you consider all the parts first, not after you have designed it. The values of the capacitors, resistors and coils need to change if you decide to use different parts, because the balance will be different.

Also, as an affirmation, when asked about this very subject (restoring crossovers), Andrew Jones also stated that the ESR properties of the original capacitors should be adhered to, for the very same reasons mentioned you cited. The ESR is an important specification in a capacitor when designing a crossover and it needs to be accounted for.

In short, I generally do not recommend replacing capacitors with different types. If there is an electrolytic in there, replace it with an electrolytic. Electrolytics are just fine and these days are made with higher quality than they ever have been. By replacing with a different part, you run the risk of changing the tonal balance, and straying away from the designer's intentions.

To answer your more specific question regarding the EPI 70s, if it were me, I would just use a quality electrolytic and call it a day.

 

[canuckaudiog]

Hey man,

If you want to use a film cap and simulate the ESR of the original electrolytics, add a 1 or 2ohm (I've found 2ohms work well) 10 watt resistor in series with the driver, closest to the driver, on the + side.

Biggles

I've tried doing this but the problem with it, is the resistor only affects the frequency of the crossover point, and not the entire bandwidth. It doesn't properly simulate ESR (as you know, ESR is over a variety of frequencies), so the resultant sound is sub par at best.

 

[Retrovert]

Points to remember.

(1) Electrolytics have a rectification effect which causes distortion, and the dipole alignment is different than film.
(2) The higher ESR of an electrolytic is not a feature in a crossover, it is a defect.
(3) If poly film capacitors had existed at the right price point when these speakers were designed this would have been used.
(4) Electrolytics are inexpensive and compact, which is why they are used in crossovers, not because of superior sonic characteristics.
(5) Electrolytics have a short lifespan, degrade over time, and degrade if not regularly used.​

 

[toxcrusadr]

I've tried doing this but the problem with it, is the resistor only affects the frequency of the crossover point, and not the entire bandwidth. It doesn't properly simulate ESR (as you know, ESR is over a variety of frequencies), so the resultant sound is sub par at best.

I don't quite follow the first part of this. A resistor in series with the driver would affect the entire signal going through it.

As for ESR over the frequency band, a couple months ago I saw a link to a page where someone tested old and new caps in detail, including ESR and frequency sweep tests, and found some very interesting results. I can't find it now, but in any case it was clear that the old electrolytics he tested, sucked big time. Anyway I seem to remember that ESR did in fact vary somewhat with frequency. Obviously the impedance of a cap would anyway, even a 100% efficient one, because that's what they do.

 

[toxcrusadr]

Really the bottom line is whether you like the sound. I've used film in place of those mid-sized electrolytics and they sound good to me, but I like a little more treble maybe than some people. Especially the top end.

 

[canuckaudiog]

I don't quite follow the first part of this. A resistor in series with the driver would affect the entire signal going through it.

As for ESR over the frequency band, a couple months ago I saw a link to a page where someone tested old and new caps in detail, including ESR and frequency sweep tests, and found some very interesting results. I can't find it now, but in any case it was clear that the old electrolytics he tested, sucked big time. Anyway I seem to remember that ESR did in fact vary somewhat with frequency. Obviously the impedance of a cap would anyway, even a 100% efficient one, because that's what they do.

Perhaps it was the way I had it wired up, but what I did was wired a resistor right after the polypropylene capacitor. When I took measurements of the speaker, I noticed that the frequencies around the crossover point were lowered, not the entire bandwidth.

 

[Retrovert]

Resistors in series with drivers affect the damping factor. Any impedance in the circuit, including inductor DCR, does that.

 

[StimpyWan]

What concerns me about ESR, is how do we truly measure it accurately? In old degraded capacitors, has ESR shifted the same way capacitance has? If so, how do we know what the original ESR was, in order to match that measurement? Is it possible? And does it really matter, if we are talking about tenths of an Ohm? Can we hear that small of a difference, or do we only think we can?

 

[Retrovert]

ESR in a degraded electrolytic can be a few Ω, enough to be about 2 dB if you work out the math.

The other problem is that the frequency response of the electrolytic is non-linear at best, and it worsens with age. Electrolytic capacitors are intended for use in DC power supplies, not in AC circuits.

 

[toxcrusadr]

The above two posts.

To recap p) :

1. Figuring out what it WAS when the speaker was new is the problem
2. It's a small effect (assuming undegraded caps)
3. How it sounds to YOU after the recap is the important thing.

 

[Retrovert]

Perhaps it was the way I had it wired up, but what I did was wired a resistor right after the polypropylene capacitor. When I took measurements of the speaker, I noticed that the frequencies around the crossover point were lowered, not the entire bandwidth.

Follow the math:

f = 1 / (2Pi x R x C)​

So if you increase R while holding C constant, this becomes f ~ 1 / R. So as R increases from RO (R Old) to RN (R New) the frequency will correspondingly shift by 1 / (RN / RO).

Practical example. Suppose an 8 Ω midrange is crossed at 800 Hz using a 25 uF capacitor. If R is now increased to 10 Ω via a series resistor, an increase of 25%, the crossover point will drop by 25% to 637 Hz.

But it pays to remember that nominal speaker impedance is a nudge-nudge, wink-wink value. Sanity will otherwise suffer.

Edit: fixed typo.

 

[Oerets]

I have found both types of caps in EPI's when under the hood. So the manufacture didn't change value or add parts when different types of 10uf caps were used.

Go with a poly and forget about it.