Where is the point of diminishing returns with caps in passive crossovers?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm replacing the caps in the crossovers of my well known brand 3-way horn-based speakers. From what I have been reading, I believe that the low pass part of the crossover is a 2nd order, involving a 3mH inductor in series and a 47uF cap in parallel. The original cap is a non-polar electrolytic. My understanding is that it's function is to "bleed off" any high frequency signal that isn't filtered out by the inductor. When I've run the values through various on-line calculators it seems that the crossover point is around 400Hz (still hazy on whether this is the 3, 6 or 12dB point). FWIW, I think the values that result are bsaed on the Litz-Riley type, although I could be wrong.

So, my questions are:
1) Am I understanding this correctly?
2) Does the cap have any other function besides filtering out higher frequencies?
3) Besides "sound quality" and variable reactance with temperature and frequency (but, does this occur at anywhere near audible levels?), is there any specific advantage to replacing the fairly cheap electrolytic cap with a much more expensive poly cap?
4) My reading on the metalized film caps vs. film and metal foil caps (e.g., data sheets) indicate that the plate/electrode layer of the former is on an order of 250 times thinner (.02 vs. 5 or more u) and at least one writer raised the issue of lower resistance in the plates of the metalized film caps (see section 2.4) Capacitor Characteristics:
One thing that may be very important for passive crossover networks is the material used for the 'plates' of the capacitor. Metallised film caps may not be the best choice because of the resistance of the film itself. The film is usually extremely thin, and it may not have a low enough resistance to allow the full current required. I have not experienced any problems with this, but a film and foil type is more suited to high current operation than a metallised film construction. This topic is mentioned on capacitor manufacturers' websites, and I recommend a search if you want more information about current handling capacity.​

Of course, I have not been able to find a sanely priced 47uF 100V film and metal foil cap, so the main question boils down to:

Should I bother spending a minimum of $12.55 per polypropylene metalized foil cap (Dayton), or is replacing the electrolytic cap with film a case of reaching the point of diminishing returns (vs. making listening environment changes, etc.)? Is there a significant (measurable physical performance) difference in how the two types of caps perform in this particular function?
BTW, I'm replacing the other caps in the crossover with NOS non-boutique film and foil caps (Wesco 32M 1.0 100V with 1% tolerance that measured well within 1%) and some 277P 1.5 uF 200V 10% orange drop style caps that I haven't received yet. I couldn't find any references to 277P caps, but the model number was similar to other known foil caps, so I gave it a shot.

Also, I'm still pretty much a newbie with practically zero training in electronics or even electrical theory (I still have trouble with the metaphors I've seen for Voltage - voltage is pressure? So, voltage is metaphorically like the amount of water pressure in a garden hose, pressure washer, and ? Isn't that directly related to resistance, then? What is the force operating? I think I get current - it's the amount of electrons flowing, right?) So, if anyone has any other metaphors I'd be glad to hear them.

Here are some of my audio biases:
  • I don't really buy into the idea that boutique caps make an quantitatively better difference vs. well made generic caps of the same construction.
  • I do have some training in human perception, and I'm very much aware of the ideas of the detectability of differences in signals (my recollection is about 3% for untrained observers and lower for experts) and accommodation (as you grow used to an often heard signal your brain gradually shifts it into a "background" mode ).
  • I think that "different" does not necessarily mean better or worse.
  • I also have fairly extensive training in the cognitive sciences, and am well aware of the various biases the people use in evaluating their environment.
  • I think that these factors explain much of what is being described bu people when talking about the tone qualities of electronic components

Lots of questions here, and yes, I've completed lots of searches already.

Thanks,

Mike
 
Also, I'm still pretty much a newbie with practically zero training in electronics or even electrical theory (I still have trouble with the metaphors I've seen for Voltage - voltage is pressure? So, voltage is metaphorically like the amount of water pressure in a garden hose, pressure washer, and ? Isn't that directly related to resistance, then? What is the force operating? I think I get current - it's the amount of electrons flowing, right?) So, if anyone has any other metaphors I'd be glad to hear them.

Pressure (psi) - Voltage (volts)
Flow rate (gallons per minute) - Curent intensity (ampers)
Diameter of hose - Conductor diameter
Water vane (strangulation that oposes water flow) - Resitor (opposes the curent flow). Across the vane you have a head loss (pressure drop) for a certain flor rate. Across a resitor you have a voltage drop for a certain curent.
A tank with two inlets and a rubber membrane in the middle - capacitor with two terminals.
A heavy paddlewheel in the the flow - coil (rotating mass is like inductance). How good it blocks (or accelerates) the flow is the Q factor.
A check-valve - electronic diode

So yes, they are similar.
 
Last edited:
It sounds like you know enough to make practical purchases and listen to them in order form your own opinion. Unfortunately in this sort of endeavor, unless you're looking for a lot of argument, that's ultimately all you get. I'll say that a good quality and well applied air core coil and film caps are going to sound different (usually better) than some type of iron core and electrolytics. Beyond that you'll have to determine what's best and why.
 
One thing to bear in mind is that the ear is very sensitive to small frequency response variations and especially so between 1-3kHz. As most crossovers to tweeters are placed in this region it is easy to see how a component change could cause the sound to change. Now one could of course assume this to perhaps be related to a better quality or different quality of crossover component, whereas in reality the components come with tolerances that are generally around +-10% to +-5%, not to mention differences in the series resistance.

If you want to do crossover component changes and be sure that you are really hearing the component you need to make sure that the frequency response of the loudspeaker before and after a component change are as close to being the same as is possible. Otherwise that difference in character could simply be a small change in output around the crossover frequency etc.

Generally speaking good quality electrolytics are fine, one can step up to a film cap and get appreciable performance increases that generally ends up meaning better component tolerances with respect to aging and a lower equivalent series resistance. Once you're using a decent quality film cap though (say a nice basic polyprop) paying for a supposedly better film cap is where the laws of diminishing returns really kicks in. Of course even basic poly caps can cost a lot when you're looking at large values of beyond 50uf. Here I tend to parallel a low size poly film cap with a large lytic. One area that can improve with price is the tolerance of the parts. I would much rather pay a small premium for +-1% parts of a standard cap line, then pay more for a supposedly better cap that was still at +-5%.

As to the question of the parallel connected 'shunt' cap in the 2nd order LW, low pass electrical network, the quality of the cap here can make an appreciable difference in the frequency response, but only if the ESR of the chosen cap is gross. The cap is obviously chosen with respect to the impedance of the driver and is 'tuned' one could say to a specific frequency. At low frequencies the caps impedance will appear as very high compared to the driver so all the signal will pass through the driver. As frequency increases the caps impedance begins to fall and you end up with some of the signal being divided between the driver and being returned to ground via the cap. If the caps own ESR is high it could start to dominate such that the response of the filter doesn't do quite what you're expecting it to do. Of course all of this can be modelled and factored into the design.

If you're altering an already bought loudspeaker then the only thing you can do is try it out. The good thing is that decent quality film caps aren't madly expensive unless you start using large values. As the loudspeaker is already using a lytic cap there is always the chance that it could have dried up a little and changed in value. If you can measure the caps and see that they are within spec, then it is up to you as to whether or not you spend the money. If they were in spec then I personally wouldn't change a thing.
 
Thanks all for the metaphors and feedback. Having gone out to the local wine bar and then the local fried chicken and beer place since posting my original thoughts, I will need some time to fully process all the information (i.e., sober up).

re: hearing differences. I have a fairly mediocre short-term/working memory for most things, especially information (signal) presented auditorally. So unless I set the system up so that I can switch back and forth multiple times between two versions of a component within seconds, I don't expect that I will be able to detect which of two audio streams sounds "better" (I've tried this comparing different audio components - too much time switching between the different components to tell). I haven't even attempted to try a "blind" comparison. I'm mainly trying to increase my chances of producing the recorded signal with as much fidelity as possible.

@SonicR1 - I would have thought that across the vane and resistor there would be a pressure increase vs. a decrease. And for the tank/capacitor; the inlets are connected to each other outside of the tank (ie., the negative and positive poles of a circuit)? Do I have that right?

@DF46 And vice versa, right? Like a water tube level?

And what do you all think of the significance of the resistance differences between metalized film and aluminum.tin foil plates in capacitors? Electrolytic vs. metalized film vs. film and foil?

Cheers
 
@SonicR1 - I would have thought that across the vane and resistor there would be a pressure increase vs. a decrease. And for the tank/capacitor; the inlets are connected to each other outside of the tank (ie., the negative and positive poles of a circuit)? Do I have that right?
Well across the vane pressure decrease if you take refference the source pressure, in the direction of flow. This if you have a constant pressure source that feeds the vane.
If your water hose leaks back to ground, that atmospheric pressure would be "0 volts".
An electric power source is like a pump. It will convert a mechanical energy into pressure/flow, but the water that it "sends" away needs to return to the pump if you want the circuit to work continuously.

And yes the capacitor you got it right. Volume of chambers is the capacity. You can send pulsations back and forth via that membrane (AC), but no continuous flow (DC).
 
And what do you all think of the significance of the resistance differences between metalized film and aluminum.tin foil plates in capacitors? Electrolytic vs. metalized film vs. film and foil?

That's an important question. In some sopeakers, the designer is smart enough to make use of the higher ESR (equivalent series resistance) of an electrolytic cap as part of the crossover circuit. Substitution of a film cap will certainly change the sound, but may well make it less accurate than before. In other cases, the electrolytic will be used just because it's cheap, with no real design thought, and the film cap may or may not sound different or better.

Since decent film caps are cheap (avoid audiophile designer brands!) as are bipolar electrolytics, it's easy and educational to buy and try both and see which is the case for your particular speakers.
 
Just another Moderator
Joined 2003
Paid Member
Where I found the biggest difference between electrolytics and film caps was in notch filters. The increased ESR of the electros reduced the effectiveness of the notch filter quite substantially. However a bit of esr can also be a good thing for damping resonances that might otherwise occur.

Tony.
 
That's an important question. In some sopeakers, the designer is smart enough to make use of the higher ESR (equivalent series resistance) of an electrolytic cap as part of the crossover circuit. Substitution of a film cap will certainly change the sound, but may well make it less accurate than before. In other cases, the electrolytic will be used just because it's cheap, with no real design thought, and the film cap may or may not sound different or better.

Since decent film caps are cheap (avoid audiophile designer brands!) as are bipolar electrolytics, it's easy and educational to buy and try both and see which is the case for your particular speakers.

the lower priced mundorf film caps (poly) are excellent designs. A safe bet, (regarding taking a chance on a potential to fit the given application) IMO and IME.

MUNDORF EB GMBH - Capacitors

(Mcap is the one)

As far as being a 'knee' in the cost vs quality sweepstakes curve, this cap is a good choice. I'm sure there are others (contenders), this is just one of them.
 
Crossover point

One thing to bear in mind is that the ear is very sensitive to small frequency response variations and especially so between 1-3kHz. As most crossovers to tweeters are placed in this region it is easy to see how a component change could cause the sound to change.

Interesting point. The speaker is a three-way design with the mid and high ranges covered by horns. It seems to me that the tweeter X-over point is at about 6000Hz. Of course, now that I've shortened your quote I've forgotten the rest of your discussion and can't think of how your points apply to this situation.:rolleyes:
 
I would much rather pay a small premium for +-1% parts of a standard cap line, then pay more for a supposedly better cap that was still at +-5%.

So far I have 1.0uF caps for the midrange section of the network which are rated at (and test at) 1%. Pretty much spot on.

As to the question of the parallel connected 'shunt' cap in the 2nd order LW, low pass electrical network, the quality of the cap here can make an appreciable difference in the frequency response, but only if the ESR of the chosen cap is gross. The cap is obviously chosen with respect to the impedance of the driver and is 'tuned' one could say to a specific frequency. At low frequencies the caps impedance will appear as very high compared to the driver so all the signal will pass through the driver. As frequency increases the caps impedance begins to fall and you end up with some of the signal being divided between the driver and being returned to ground via the cap. If the caps own ESR is high it could start to dominate such that the response of the filter doesn't do quite what you're expecting it to do. Of course all of this can be modelled and factored into the design.

A lot of info to digest here. The low crossover point appears to be about 400Hz (maybe lower). The speaker impedance (12" woofer) is 4.0 ohms. The original cap is a standard 20% tolerance 47uF 100V electrolytic. Any thoughts here? I'm wondering how changing from the to a poly would effect the crossover point and performance.

Some other posters (on another thread on possibly a different forum, I forget at the moment) have emphasized replacing the stock 3mH inductor with what LL at best 18AWG wire (probably smaller) with a coil constructed with a much larger gauge wire to reduce the DCR. However, I did not see a discussion of how much or even what kind of difference a lower DCR (in the < .5 ohm range) would make. Faster response by the woofer to frequency changes? I 'm just guessing here.

Thanks for all your feedback. Any additional thoughts you have would be great. But remember, I don't know much!:)
 
Last edited:
If the caps own ESR is high it could start to dominate such that the response of the filter doesn't do quite what you're expecting it to do. Of course all of this can be modeled and factored into the design.

Hence the danger of any use of electrolytics in a given crossover. Electrolytics age. Sometimes, terribly. Only one of their problems.

Eg, I just pulled a set of 16V 2200uf electrolytic caps from a 30 year old circuit. (amplifier, not speaker)

They measured 100uf at 120hz.
 
The original cap is a standard 20% tolerance 47uF 100V electrolytic. Any thoughts here? I'm wondering how changing from the to a poly would effect the crossover point and performance.

Well I think the question would be, how far out from 47u is the cap? The designer most likely wanted the cap to actually be 47u so anything other then that is a deviation from his or her original intentions. We do not know this for sure, however I doubt they would have quality control such that they would measure each lytic that came through the door, just to make sure they were within spec. From this point of view it would probably be cheaper to use a basic poly cap, so it is quite likely that the cap, being a 20% lytic, could quite easily be not quite what the manufacturer had intended.

If the ESR of the lytic were to have a detrimental effect on the response, then I would expect it to have less effect around the knee of the crossover (where you want it to have less effect) and more effect towards the end where the filter is well into the stop band (where you don't mind it having more effect).

If you can measure the lytic and it is in spec, then I wouldn't change it, however if you cannot, I don't think changing to a poly would be a bad idea as per say. Doing so could very well change the sound, if for the better would be entirely subjective on your part.

Some other posters (on another thread on possibly a different forum, I forget at the moment) have emphasized replacing the stock 3mH inductor with what LL at best 18AWG wire (probably smaller) with a coil constructed with a much larger gauge wire to reduce the DCR. However, I did not see a discussion of how much or even what kind of difference a lower DCR (in the < .5 ohm range) would make. Faster response by the woofer to frequency changes?

Altering the series resistance of the inductor will alter the SPL that the bass driver plays at, whilst altering the alignment of the driver in it's enclosure very slightly too.

The series resistance of the inductor should have been designed into the loudspeaker right from the start, this doesn't mean that it was though. Again you're really experimenting a little here, you might prefer the sound with a lower DCR inductor and you might not.

If you're really going to give the low DCR thing a go then get a good cored inductor such as this to really make the comparison and money spent worth it.

The impedance of the woofer here is of some importance perhaps, in that relatively speaking, a higher DCR inductor will have more of an impact on a 4 ohm driver compared to an 8 ohm driver as the DCR of the inductor represents more of a change in proportion to the drivers resistance.

As to the sound being faster, no this wont happen, but subjectively it could tighten up the bass a little if the driver is already in a box that's slightly too small.

As with the capacitor though, you'd have to try a low DCR inductor to find out if you prefer it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.