Use of parallel caps in a speaker crossover?

[brucew268]

I'm a little unclear on the degree and character of effect of bypass capacitors in a speaker crossover, and am trying to predict the results of various options. For example:


1. If I replace a Solen PP cap with a Mudorf Silver/Oil cap, I would assume the effect of the Mundorf would be much greater than if a small Mundorf were used as a bypass on Solen. Or not?




If I find Solen PP to be a bit thick and syrupy, thus seeming slightly 'slow', lacking a little detail that I want...
And Ansar PP in that same application seems better at detail, seems quicker, still somewhat smooth but just on the edge of hard in the upper mids/lower treble at times...

2. If I paralleled each with the Mundorf Silver/Oil, would the Mundorf push the sound in Mundorf's direction whilst keeping much of the bigger cap's character? Thus the sound in the Solen/Mundorf app would be a bit different to the Ansar/Mundorf app?

[awkwardbydesign]

Is this any help? http://www.humblehomemadehifi.com/Cap.html

[YNWaN]

Yes, that is the theory. However, when I say theory I should point out that mixing caps in this way is a contentious issue and there are those who claim it makes no difference.

I feel I should also point out, although you may well know this, that adding caps in parallel and using a bypass cap are not quite the same thing. They are both adding caps in parallel, but what is termed a bypass cap is usually of very low value (0.1uF or even 0.01uF). What you are describing is more a blending of capacitor values - it is common to find the cap value you want is not available and it is also common to create that value by paralleling up two, or more, caps to make the correct value. The smaller values used to make up the whole are often of a higher quality for the reasons you give.

[Reffc]

You can only really try it and see for yourself as the results are often unpredictable and there are no hard and fast rules to SQ with bypassing. One reason it was originally done was to improve linearity (with frequency) of 'lytics by using a smaller polyester or polyprop as a bypass in a small value as Mark says. Sometimes, this works with other caps but for the most part, most of todays polyprops are all linear and not affected to the same degree by changes in temperature or frequency compared with older 'lyics.

Another good reason for placing capacitors in parallel is where you may have values of say 5% tolerance and if we take say a 20uF as an example, one 20uF at 5% could be 21uF on one side and 19uF on the other if bought as one 20uF value at 5% whereas putting two 10uF caps in parallel is more likely to result in a tighter average tolerance between left and right channels. In fact, it's a good idea to do this if you want to keep things nice and tight for larger values (anything much above 10uF really). The alternative is to pay more and get 1 or 2% tolerance matched caps for more critical applications.

As to one poly sounding slow and syrupy, the description may well be what you hear as a difference, but in electrical terms, almost all polys made today have very low impulse time and low ESR and also have low dielectric absorption so tend to be quite "clean" sounding in the bass for example as the caps don't have the same charge memory as something like older 'lytics.

It's an ongoing (circular) debate as to "which is best" but everyone will have their favourites. I tend to think that the values are more important and getting the crossover points right makes a bigger difference than the name on the capacitor and using parallelled caps is one way of keeping a reign on tolerances.

[brucew268]

Paul: Yes, when I spoke of syrupy, slow, or fast I was meaning to speak of the experienced sound character not of the electrical performance in the circuit. Your information is good for me to stock away for future consideration. The statement about variable results and the need to try and see... ah well. Maybe I'll have to do that. Serial experimentation with several expensive caps was what I was hoping to avoid, but maybe will have to be ready for.

Mark: I guess some seemingly knowledgeable folks in my experience have used 'parallel' and 'bypass' as synonymous, and others make a distinction. I wonder if the terms are used slightly differently in speaker applications from amplifier applications. The terminology is not core to my question... more interested in functionality. One person years ago said they thought that having the second cap of about 5% of the larger cap's value should result in the smaller cap influencing the sound significantly (whatever that means) with its character. But then, at least in preamp/amplifier applications bypass caps were recommended to me in the much smaller values that you mentioned.

Richard, I have read the Humble Homemade reviews a few times over the years and found it more useful than most reviews, but never looking for this specific question perviously. Interestingly, today I found these potentially relevant quotes buried in some reviews:

"Although Crescendo Harmonically Balanced film-and-foil capacitors were designed to operate as a single capacitor in most tweeter circuits, they also benefit from paralleling and cascading with the higher voltage Crescendo capacitors. They also make an excellent parallel cap when paired with most other base caps for high performance midrange circuits, and can be configured in large pure film-and-foil stacks for the ultimate in midrange performance”.

"Sonic Craft Sonicap... I didn't find them to mix well with other caps - a 50/50 mix with a standard Mundorf M-Cap gave me the funny sensation of listening to two different cap's at the same time."


The Solen caps below are were in my Proac Response 1sc speakers at purchase, and perhaps explains the sound mix that interested me for a while and then felt too confining, and so I switched to Ansar, also described below (by Humble Homade). I haven;t found them thin and distant per se, but that could relate to the little bit of lower HF hardness that sometimes peeks its head up:

"Solen PB-MKP-FC: ...Compared with electrolytics they have much more detail, smoothness and a wider stereo image. Compared to a Mundorf Supreme Cap they sound closed-in. Midrange is slightly forward compared to more neutral caps, they can be a little nasal sounding in some configurations. Due to this character they would make a good match with Focal tweeters for example. "

"Ansar CPA: has a vibrant, clean and clear nature that seems to be concentrated more on the top-end of the spectrum. Not completely neutral but therefore they can be usefull if you want to give a more open sound to a system that sounds a bit congested in the midrange and / or lower treble, in such a situation the Ansar CPA open's things up nicely. Used in a tonally neutral system they could do with a bit more presence and body, they tend to lean towards a slightly thin nature."


I still have to decide whether to spring for replacing the caps @ £274-374
or to try smaller caps in parallel, whether 5% of value is too large to be a bypass cap, I don't know. Adding smaller caps in parallel would be more like an investment of £104-164.

If the second option would get me most of the way there, I'd much rather save the dosh. But if I really need the full caps to get the results I want... I'd need to sharpen my pencil. Worst would be to not get significant results from the £104-164 and then have to lay out the £274-374! A fare bit of Scottish blood in me... and in my wife!

[Reffc]

I have my doubts Bruce as to whether bypassing will bring the seemingly significant changes you're after. The issue with using 5% or even 2 or 5% as a bypass value on existing caps is that you will shift the crossover point and change the sound if applying to existing caps, unless you meant using the same overall value but mixing new Solens with Mundorf etc to arrive at the correct values?

Whilst I have never experienced a cap per-se to influence a speaker's sound to be "shut in" (as speaker off axis response shouldn't change no matter what cap is used), I have experienced some caps (eg Jantzen Z Cap Supremes) to add a rather unpleasant brightness to upper frequencies, and in more than just one circuit. Just why this should be, I don't know. I tend to stick with Claritycap ESAs for signal path applications and use solens in shunt or Claritycap SAs. Others I have liked in HF sections have included Mundorf Evo oil caps or Mundorf M-Cap Supremes. Again, why they "sound" slightly different for the same values in the HF, I don't know, but they seem to. I'm pretty wary of that Humble Homemade review as it is purely subjective to the eyes and ears of the author in an unknown system and with unknown measured tolerances (dont forget that two caps varying by say 5% will sound different depending on where in the speaker circuit they're being used). A very healthy dose of salt should be taken. Then again, there's very little meaningful on the web about this subject. Fundamentally, it's the values and tolerances that matter most as at the end of the day, the drivers operating within their frequency bands and within their cabinet design can only be influenced so far by what make of cap they use to roll on or off, so the most significant change will always be a speaker change ;)

[brucew268]

Warning well taken not to confuse a subjective review with objective results or with results that will translate to the same results in my context. BTW: several different reviews by different people seem to give consistent description to Mundorf Silver/Oil caps. Also, I just noticed that I can get Audyn True Copper as bypass caps for significantly less than the Mundorfs.

It seems good to remind ourselves that most of these PP caps are 5% tolerance anyway. So adding a 1%, 2%, or possibly even a 5% bypass value cap I would think would not be much of an issue with crossover point.

[Reffc]

Warning well taken not to confuse a subjective review with objective results or with results that will translate to the same results in my context. BTW: several different reviews by different people seem to give consistent description to Mundorf Silver/Oil caps. Also, I just noticed that I can get Audyn True Copper as bypass caps for significantly less than the Mundorfs.

It seems good to remind ourselves that most of these PP caps are 5% tolerance anyway. So adding a 1%, 2%, or possibly even a 5% bypass value cap I would think would not be much of an issue with crossover point.

The Audyns are good caps.

a few percent may not matter but 5% may well matter dpending on the driver and the filter design (given that a cap may be at say plus 5% to start with). For a typical 6 Ohm tweeter driver which is designed say to cross over at 3KHz, increasing by 5% drops crossover (2nd order) point by around 100 to 150Hz and 10% by 250Hz so it can make quite a difference especially on the summed response. If looking at say an 8 Ohm woofer crossing at 2500 Hz, it can increase crossover point by several hundred Hz by dropping 5% on value (2nd order again). This can matter very much depending on the driver and the results imho. From my own work and results it would suggest that such variances are more audible for woofers/mids than the branding of the cap. Another way of looking at this is that it would seem good to remind ourselves that 5% can be audibly significant! You pays your money and makes your choice though...

[YNWaN]

I would agree that 5% of the total is too high be be considered as a 'bypass cap' and should be considered as part of the whole valure required.

[brucew268]

...so the most significant change will always be a speaker change ;)

If my pockets were willing, I'd be at your front door for an audition!:violin::)

[Reffc]

If my pockets were willing, I'd be at your front door for an audition!:violin::)

:lol:

You'd be welcome to come and listen anyway.

[brucew268]

Fundamentally, it's the values and tolerances that matter most as at the end of the day, the drivers operating within their frequency bands and within their cabinet design can only be influenced so far by what make of cap they use to roll on or off

I did get well enough difference between Solens and Ansars to satisfy me for a few years. However your thought, which I've heard from a couple others, raises for me this question. The tweeter on my 1sc uses a 3.3uF and a 5uF. However, when I replaced with Ansars, a 5uF was not available and so I used a 4.7uF. Could that difference be contributing to the slight hardness that I hear in women's vocals and piano on some CD's? BTW: I do show a hump at 6.6 khz.

[Reffc]

I did get well enough difference between Solens and Ansars to satisfy me for a few years. However your thought, which I've heard from a couple others, raises for me this question. The tweeter on my 1sc uses a 3.3uF and a 5uF. However, when I replaced with Ansars, a 5uF was not available and so I used a 4.7uF. Could that difference be contributing to the slight hardness that I hear in women's vocals and piano on some CD's? BTW: I do show a hump at 6.6 khz.

Hi Bruce,

The 4.7 would have a slight effect but it shouldn't really introduce a peak in response. It would just affect (slightly) the slope of the filter. Some of the early crossovers used a 4.7 as standard I think. Women's vocals I would have thought were more likely to be slightly lower, so it might be cone resonance from woofer roll-off (ie a spike in woofer acoustic response at resonance lifting the summed response above the crossover point) which is quite common. However, Proac engineered the crossovers well on these and I think they used a notch filter on the woofer to control resonance and acoustic response. Have you tried one speaker at a time to see if it might be a specific driver issue? Sometimes these things are quite difficult to pin down without measurement.

[brucew268]

I'll check it out. I've been assuming that the hardness was not actually the fundamental sounds of the voice or piano, but some sort of harmonics above it that affect the perceived hardness. I do note that on Tony Bennett's Duets II, there is a good bit of HF energy in the mix, and not just because it has an exuberant brass section. I have to be careful with my use of AC filters and time of day or it gets hard and difficult to listen to. But when all of the aforementioned is right, the nuances, shape of the sounds, and harmonics all flesh it out very well, even if it is in your face just a bit.

[brucew268]

The driver behaviour was the same in both channels. But I've found a few items which I'm trying. The existing caps were not as secure as I'd like, I had a little Dynamat (soft bitumen) on the woofer cages, and I had a power conditioner having some effect, which seemed negative at the time. The Dynamat on the woofer cage did seem to deepen the base and give a little better definition, not quite so fat either. So maybe the extension was making them a bit hard (or shouty?) at the upper end of their range, since they were operating in range wider than they work best. Removing the bitumen and securing the caps better (at the same time) made the HF feel a bit muted, but not quite as simple as that...trying to get my head around what was happening to the sound. I removed the power conditioner and got a bit of my midrange presence back (restoration of HF cues and harmonics). But the design of the power conditioner is solid, correction of PFC.

So what, if prior to the power conditioner and recent changes, my speakers had over the top harmonics due to AC power distortions? I had adjusted the toe in significantly, crossing three feet in front of my head. So maybe that orientation (related to axis and reflective surfaces) was working well compensating for the abnormalities of my context (ringing AC power, marginally secured xover caps)? I'm now starting over with speaker placement to see what my results are, and may take a week or more to be sure of what I'm hearing.

[YNWaN]

None of the power conditioners I have heard have persuaded me that they do not introduce more compromise than they solve.

To be honest, your problem sounds a bit like someone complaining about a cars performance at high revs and hoping to solve it by adjusting the torque of the wheel nuts. Having said that, bolt tension may help you a little in this case. Tighten the base driver as tight as you can (without breaking of bending anything) - do the same for the tweeter - this will make more difference than strips of bitumen.

If you have the speakers toed in significantly in front of the listener you will be listening significantly off-axis to the tweeters - some tweeters have a smooth off-axis response but many do not and have a quite spiky frequency response - this may well be true of yours.

[YNWaN]

I've had a brief Google of the ProAc 1SC and the Stereophile review is quite illuminating:


Somewhat compensating for this on-axis behavior, the lateral dispersion has a slight off-axis energy excess at the bottom of the tweeter's passband, as can be seen from fig.4. (This graph shows just the response changes, which means that the on-axis trace looks like a straight line.) In a small, sparsely furnished room, the Response One SC might sound just a little bright. But in a fairly large room like WP's, the speaker sounds neutral. In the vertical plane (fig.5), a large suckout develops in the crossover region for listeners who sit above the tweeter axis. As WP found, fairly high stands work best with the little ProAcs.

So as you move off axis it is rather as I suggested and the top end rises relative to the rest of the spectrum. It is also the case that the ProAc does not have (intentionally) a very flat frequency response in an attempt to balance it's limited bass output - this may well not suit your room.

If you have room issues it would be much more effective to add some acoustic panels at the points of first reflection and behind you:

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/3DCAB744-471D-4A96-AD53-DA6B9588A286_zpsmc7qnfjm.jpg

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/0E32EC09-F642-4539-8524-0BAC51EB7073_zpsnfycxjyq.jpg

[brucew268]

My room already has a fair bit of acoustic treatment, derived both through strategy and through trial and error.

At first when the mfr told me to consider adjusting the speakers, I thought what a load...! But I asked for further explanation and am considering his answer. Yes, I had also looked at the Stereophile measurements the off-axis behaviour of the Proac 1sc. It shows a fairly even drop off, except is a little peaky in the the 4khz-8khz range, especially in one direction off axis. And then the very high end looks a bit peaky between 30-45 degrees. So if the small size of my room and the bit bright sounding treble straight resulted in a soundstage a little less than I liked, I adjusted them enough off-axis that the soundstage fleshed out better and tamed some of the HF just a little.

So if cleaning up the power supply of its harmonic electrical ringing has decreased some of the apparent HF, then my speaker position might now result in music losing some presence, or the reduction in HF might make the upper midrange feel a bit more prominent and thus 'hard sounding'. So I'll experiment with speaker placement back toward a more normal position. However, from experience I know that this particular mains unit does not have its full effect until about 7-10 days after plugging in. So I won't know if it is bogus theory or true for some days... and in the meanwhile stay away from the main hifi for a week. BTW: probably was not appropriate for me to refer to it as as power conditioner as it takes a rather different approach and is actually trying to correct things electrically towards PFC in the ground floor mains, which is the house circuit that the hifi is wired into.

[brucew268]

Are those wooden frames with polyfill or fibreglass insulation inside? They look nicely done; but with the insignia in the corner, I guess they are commercially made.


If you have room issues it would be much more effective to add some acoustic panels at the points of first reflection and behind you:

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/3DCAB744-471D-4A96-AD53-DA6B9588A286_zpsmc7qnfjm.jpg

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/0E32EC09-F642-4539-8524-0BAC51EB7073_zpsnfycxjyq.jpg

[YNWaN]

That's my logo - most of the stuff I design or make had it - but they aren't a commercial proposition. As you say, they are a stuffed wooden frame with a perforated hardboard back and spaced from the wall by 2,5" standoffs.

Speaker positioning is the single most important aspect of optimising speakers (position relative to boundaries, stand height and toe in) everything else is just icing on the cake compared to these key aspects.

The only way a mains conditioner should take days to come of its best is if it charges a massive bank of capacitors - my least favoured approach (sorry).

At the end of the day the ProAc is a really small speaker and it's balance is partly faked to lend it a touch of weight - I'm afraid that you probably just need bigger speakers (or a decent sub-woofer).

[brucew268]

It is not a bank of capacitors, and it does not need to feed directly the mains leads for the hifi. It feeds into one of the house AC circuits: 1. It controls the (house AC) line frequency which reduces the harmonic created by the voltage flux. 2. It controls the ringing frequency of the power supply. 3. Stores energy inside it's magnetic core and returns the energy during a dip. 4. Provides active power factor correction for the component/circuit.

[YNWaN]

Hmm... 'The ringing of the power supply' - I think I've been as much help as I can be.