How bad is this Electrolytic cap in this xover?

aceinc

Member
2007-05-11 5:18 am
I have powered subwoofers, which I use in conjunction with Magnepan MMGs. Because of a number of reasons I run the combination off of the speaker wires, as opposed to sub out and the Maggies on their own amp. Changing would be a PITA.

The combination sounds quite good. I was thinking they would sound a bit better if I were to limit the low freqs from getting to the Maggies. Towards that end I was thinking about an 80hz 4 ohm high pass filter. Parts Express sells such a beast here;

Parts Express 80 Hz High Pass 4 Ohm Crossover

https://www.parts-express.com/pedocs/tech-diagrams/266-456--crossover--schematic.pdf

It is my understanding that electrolytic capacitor = bad in crossovers. The entire frequency range from 80 hz up would be going through this cap. I thought I would check the assembled wisdom of DIYAudio. To coin a phrase, "How bad is it Johnny?":D

I am not opposed to building a simple high pass filter, but to get something other than electrolytic capacitors with a spec of 100v & 350uf is quite pricey (>$75), at least where I have looked.

I have a second question, if the electrolytic capacitor will be baaad, where can I get something better at a reasonable price?
 

ubergeeknz

Member
2018-08-31 12:32 pm
Bipolar caps differ from polarised electrolytics in terms of their distortion signature. I understand that if a polarised cap isn't adequately biased, it can indeed cause signal distortion due to a "diode" effect.

As to whether bipolar electrolytics sound "bad" probably depends on your imagination and penchant for spending loads of money on huge film capacitors.

But seriously, how a capacitor will affect the sound really comes down to the specifications (and veracity thereof) for the specific model of capacitor being used.
 

aceinc

Member
2007-05-11 5:18 am
Bipolar caps differ from polarised electrolytics in terms of their distortion signature. I understand that if a polarised cap isn't adequately biased, it can indeed cause signal distortion due to a "diode" effect.

As to whether bipolar electrolytics sound "bad" probably depends on your imagination and penchant for spending loads of money on huge film capacitors.

But seriously, how a capacitor will affect the sound really comes down to the specifications (and veracity thereof) for the specific model of capacitor being used.

Based on this information;

  1. How can I tell the quality of the Cap in the above referenced HPF?
  2. Where can I get good quality BP caps?
 

H713

Member
2017-11-24 11:21 pm
Madison WI
I have powered subwoofers, which I use in conjunction with Magnepan MMGs. Because of a number of reasons I run the combination off of the speaker wires, as opposed to sub out and the Maggies on their own amp. Changing would be a PITA.

The combination sounds quite good. I was thinking they would sound a bit better if I were to limit the low freqs from getting to the Maggies. Towards that end I was thinking about an 80hz 4 ohm high pass filter. Parts Express sells such a beast here;

Parts Express 80 Hz High Pass 4 Ohm Crossover

https://www.parts-express.com/pedocs/tech-diagrams/266-456--crossover--schematic.pdf

It is my understanding that electrolytic capacitor = bad in crossovers. The entire frequency range from 80 hz up would be going through this cap. I thought I would check the assembled wisdom of DIYAudio. To coin a phrase, "How bad is it Johnny?":D

I am not opposed to building a simple high pass filter, but to get something other than electrolytic capacitors with a spec of 100v & 350uf is quite pricey (>$75), at least where I have looked.

I have a second question, if the electrolytic capacitor will be baaad, where can I get something better at a reasonable price?

It's fine. I agree that it's greatly preferable to do the filter before the amplifier.

If you insist on using this crossover, there's nothing inherently wrong with using an electrolytic, but pay attention to its maximum current. Given that it's something Parts Express is selling, they probably did their homework and are using that cap within its ratings. The cost will be astronomical for a poly cap of this size- don't even bother. You won't hear a difference.

Edit: The filter you linked to is already using a BP cap.
 

aceinc

Member
2007-05-11 5:18 am
@ Ubergeeks
Oh no no no! It's just that you don't want any reactive component at power level. It's very unpredictable what the filter+load can do. At least with magneto dynamics with the characteristic impedance peak at Fs...

Is what you are suggesting that the compromise of passive crossover networks in general are so fraught with problems there is no place in mid to hi fi systems?
 

aceinc

Member
2007-05-11 5:18 am
It's fine. I agree that it's greatly preferable to do the filter before the amplifier.

If you insist on using this crossover, there's nothing inherently wrong with using an electrolytic, but pay attention to its maximum current. Given that it's something Parts Express is selling, they probably did their homework and are using that cap within its ratings. The cost will be astronomical for a poly cap of this size- don't even bother. You won't hear a difference.

Edit: The filter you linked to is already using a BP cap.

The cap used in the PE Xover is 100v, I am driving a pair of MMGs with Emotiva XPA1s which are rated 1KW at 4 ohms. Is 100v sufficient?

I noticed it was a BP cap, but is it a "good" BP cap?

Edit: The first question is probably superfluous, as once the signal below 80hz is removed, the power requirements will decrease.
 
Last edited:

wg_ski

Member
2007-10-10 5:21 pm
All electrolytics are not created equal. The ones you’ll find on a pre-assembled crossover board are typically junk. Put more than 20 watts of average power through them and you get the can embedded in your forehead and pieces of confetti all over the room. Ones designed for filtering rectified mains voltage for SMPS duty can handle a couple hundred without overheating and exploding. The only way that can happen is with fairly low ESR. The same low ESR you need for audio use. Get a pair of 680 uF/200 V snap-in caps rated for 105C and a couple amps of ripple current and wire back to back. Panasonic, Nichicon, etc. from Digikey. Make up the last 10uF with a decent MKP.

350 uF of MKP sure can get expensive. Not needed anyway, because the INDUCTOR will have a non-trivial DCR and will de-Q the network a bit even if the cap is ideal. There will be some “capacitor distortion”, but will likely be less than the iron core inductor and certainly less than the speaker itself at low frequencies. At higher frequencies where the cap is basically a short, the cap distortion reduces just because the voltage across it gets small.
 

wg_ski

Member
2007-10-10 5:21 pm
It's not at all unpredictable.

Quite predictable and modelable. The crossover Q and frequency will shift but you can easily calculate/simulate what it will be with your measured complex load impedance. Adjust values accordingly if it’s enough to upset you. If you model the acoustic low frequency behavior of the speaker itself as a filter response in cascade with the electrical filter you can get it even more “accurate”, but at some point in-room measurements would be required to improve things.
 
Hi aceinc,

Cyril Bateman showed two non-polar electrolytics were as low distortion as the film cap. See link for his article capacitor coupled vs. direct coupled ouput
also downloads from ProAudio site.

To add two non-polar's to your filter it may be easier to get 4 off 1400uF polar electrolytics at say 100V and wire them as two non-polars in series.

Each non-polar is made from + to + (or - to - if you prefer). It doesn't matter how the two non-polar pairs are wired in series (because they are non-polar's:D).

Select 105C then low ESR, then cost, then the longest life of those. To get 1400uF to +/-10% you may have to measure them. if you buy 8 (for 2 ch) then you can choose the best combination from them for two 350uF sets.

Alternatively, search for 700uF (closest to it) and parallel 2 off 175uF sets to make 350uF. You then get 16 off around 700uF polar's to measure and select.

Sorry if it sounds too complicated. But isn't this DIY?
 

N101N

Banned
2018-05-30 1:21 am
It's better to move the HP at line level, i. e. before the amplifier. Just change the value of the input cap...

Of course.

Oh no no no! It's just that you don't want any reactive component at power level. It's very unpredictable what the filter+load can do. At least with magneto dynamics with the characteristic impedance peak at Fs...

Of course.

¤

The Parts Express 80 Hz High Pass 4 Ohm Crossover is of poor quality.

I would use a 6dB high-pass filter at around 100Hz, determined empirically (and a 80Hz / 24dB active low-pass filter for the subwoofer).
 

diyiggy

Member
2019-01-16 12:22 am
Is a NPE so good in serie impulse behavior above 1000 Hz as the original poster is talking about HPF ?


Why not purchase a safer MKT (which has quite good dielectric insulation enough) instead more expensive MKP ?



Anyway it should be less expensive than 75 USD imho : Jantzen MKP Cross Cap is 17 USD at 100 uF/400Vdc, take 3 with a 47 uF and you should be good as they are 3% acurate... around 60 USD I assume for 350 uF. Now as said MKT will suffice and lytics as well. Mundorf/Jantzen have NPE with a good enough ESR at 0.03 DF /1K Hz (at least perhaps not good enough if the speaker is an ESL?)
100 uF/100 Vdc non etched npe from Mundorf are 5 euros here in CE
 

H713

Member
2017-11-24 11:21 pm
Madison WI
In most cases, throwing expensive parts at a mediocre design does precious little.

My personal opinion is that this kind of thing should ideally be done before the amplifier because it is much easier (and much, much cheaper) to design a good filter for line-level applications than for high-power, low-impedance systems like a loudspeaker.

If you must put the filter after the amplifier, then it will require careful design (not just plugging a few numbers into an online calculator) and component selection. Using an online "crossover calculator" is not a great solution (maybe a good starting point for tweaking). Buying an off-the-shelf generic crossover is a significantly worse solution. Every driver has different characteristics, sometimes drastically so. That's why it's impossible to design a "universal" crossover, so to speak.