Hi. Is there any advantage to using multiple small capacitors, instead of one large capacitor, in a crossover network.
I'm wanting a 3uf capacitor for a tweeter, should i use one 3uf, or 3 x 1uf. Also, is there any point to adding a small 0.2uf teflon to the parallel stack. Some people are reporting good results.
The capacitor i am thinking about is the k71-4 polystyrene nos from russia, and ft-3 teflon
Thanks
I'm wanting a 3uf capacitor for a tweeter, should i use one 3uf, or 3 x 1uf. Also, is there any point to adding a small 0.2uf teflon to the parallel stack. Some people are reporting good results.
The capacitor i am thinking about is the k71-4 polystyrene nos from russia, and ft-3 teflon
Thanks
Save your money and buy a polypropylene for speaker crossover duty.
It's only when the polypropylenes become too big and/or too expensive that you then swap to Bi-Polar electrolytics.
It's possible that you may hear no deterioration using one of the lesser plastic films like MKT, MKS, PPS etc.
This is especially so in a low pass duty. High pass tends to be more critical and may expose the very slight distortions that the lesser plastics create.
Have you read any of the many capacitor reports from C.Bateman?
Here's a starter.
All the others are over the 1MB Forum limit.
It's only when the polypropylenes become too big and/or too expensive that you then swap to Bi-Polar electrolytics.
It's possible that you may hear no deterioration using one of the lesser plastic films like MKT, MKS, PPS etc.
This is especially so in a low pass duty. High pass tends to be more critical and may expose the very slight distortions that the lesser plastics create.
Have you read any of the many capacitor reports from C.Bateman?
Here's a starter.
All the others are over the 1MB Forum limit.
Last edited:
I would advise 2 x 1.5uF, preferably Film and Foil type, such as Jantzen Alumen Z-Cap or if you like oil, the Arizona Blue Cactus. These two are the rare breed of film and foil caps which are of reasonable size with large capacitance. Rated for 200V.
I used to mix polypropylene, teflon, oil etc. but now prefer to stay with one type of dielectric for each purpose.
Last edited:
The point is whether is the same of having one cap or many caps. I prefer single cap, same as in amp's PSU. Some advocate parallel is better.
The same consideration could be made for resistors : for, say, 3.3 Ω resistor, should be ok to use 3 X 10 Ω resistors paralleled ? But at this stage ( perceived sound correlated to the materials/technique used ) everyone is on his own. I prefer single component...I don't like "the signal to be splitted then re-converged in one point"
The same consideration could be made for resistors : for, say, 3.3 Ω resistor, should be ok to use 3 X 10 Ω resistors paralleled ? But at this stage ( perceived sound correlated to the materials/technique used ) everyone is on his own. I prefer single component...I don't like "the signal to be splitted then re-converged in one point"
Alternatively one can use resistors in series. That avoids your splitting and re-converging...............I don't like "the signal to be splitted then re-converged in one point"
Theory would prove that inductance is usually too low to affect audible frequencies, and ESR changes very little with addition of small bypass caps (relative to their "weight" in overall capacitance).
Practical observation is that bypass caps influence the sound and it is usually audible as shine, sparkle, detail in high frequencies. More often then not it also leads to listening fatigue IMHO. My observation with russian teflons is that they generally add dry detail, no glare. Try them out, they are cheap. But don't be afraid of critical listening. Revert the change and compare. Not all "improvements" are indeed for the better.
One good quality cap is better than medium quality one with a high end bypass. It's also much easier to get the sound right (coherent) with a single cap or a cap battery made of same type capacitors.
Reduced ESR might make a difference - but only for caps with significant ESR such as bipolar electrolytics. For film caps it makes no difference as ESR is already very low. Your aim is to roughly match what the designer intended.goodguys said:
Reduced inductance makes no difference as this is audio, not radio.
'Bypass' caps are almost always unnecessary in audio, and sometimes harmful. They either do nothing, or add an unwanted resonance. If the latter occurs you may hear it, and may at first mispercieve it as an improvement.
In my opinion the only reason for paralleling capacitors is if you need a certain value that you don't have, instead decide to replace them by capacitors that are laying around somewhere in your stashes. I did this in order to replace the 25 µF bipolar electrolytics that I've found in my 1970ies Wega speakers by a combination of 15 µF + 4µ7 + 4µ7 foil capacitors. I've shown pictures here: Restoration
You can get measurably lower ESR by paralleling a cap, as well as inductance.
However, most film caps have really really low ESR to begin with. The days when low ESR was considered more transparent are long gone, but it does contribute to the crossover balance.
My advice to anyone playing around with bypass caps: Try an Audyn True Copper cap in the 0.1uF range. I have found it helps warm up larger caps ( > 5uF) and they're not unreasonably priced.
Best,
E
However, most film caps have really really low ESR to begin with. The days when low ESR was considered more transparent are long gone, but it does contribute to the crossover balance.
My advice to anyone playing around with bypass caps: Try an Audyn True Copper cap in the 0.1uF range. I have found it helps warm up larger caps ( > 5uF) and they're not unreasonably priced.
Best,
E
I've always found that using high quality bypass caps across larger medium/low quality larger caps works a treat so long as they are not in series with the signal.
With caps that are in series with the signal, I found just one cap of the highest quality you can use is the best way to go, any bypass caps on these seem to smear the sound at frequencies where the two overlap.
Cheers George
With caps that are in series with the signal, I found just one cap of the highest quality you can use is the best way to go, any bypass caps on these seem to smear the sound at frequencies where the two overlap.
Cheers George
Where the smaller "bypass" one starts to roll off and the bigger main one is still doing it's job.
Cheers George
I talking about series caps, or coupling caps that have the audio signal going through them.
A series bypass signal cap (coupling) which is too small by itself may LF roll off at say 2khz, the main cap that it's bypassing will go down to below 20hz and go above to over 20khz.
The over lap of two caps with totally different construction and sounds qualities will be overlapped from 2khz to 20khz, giving two different characteristics which to me smear the sound rather than making it better. As I said best to use one very high quality cap in these series (coupling) positions.
Cheers George
Last edited:
There's nothing wrong with paralleling caps, and at very high frequencies there are some benefits. By very high I don't mean audio frequencies. Those benefits happen more when bypassing high speed logic and wide bandwidth opamps.
There's also no benefit in dissipation factor (or esr). Why? Let's say the target is 3 uF. Imagine a 3 uF cap stretched out as a big square of dielectric with metallization on both sides. The losses, DF or esr as you please, are whatever they are. Now, get out your scissors and cut the cap in half. Each half now measures 1.5 uF because it's half the area. The DF remains unchanged because it's a ratio with capacitance. The esr, however, doubles because it's not a ratio but an absolute number. IMO, a good reason not to use esr unless you're a switching supply designer, but I digress.
So, you've got these two 1.5 uF sheet capacitors. You get two pieces of wire and connect the top and bottom electrodes of the caps together. You've paralleled them, more or less the same as before you cut them in half. You now have the original 3 uF, with the same DF and esr you started with.
The lesson is if you use the same materials, say Mylar or polypropylene, there's little gain or penalty for paralleling capacitors.
Thing get more complicated if one capacitor is of poorer quality than the other, but we try not to use poor quality capacitors! Hint- it's impossible to improve a poor quality capacitor (by that I mean high losses, DF or esr, as you please) by putting some smaller capacitor in parallel with it, so if you want good, use a single good cap, as has been said. No penalty for using a couple good caps, if you need to hit a value.
There's also no benefit in dissipation factor (or esr). Why? Let's say the target is 3 uF. Imagine a 3 uF cap stretched out as a big square of dielectric with metallization on both sides. The losses, DF or esr as you please, are whatever they are. Now, get out your scissors and cut the cap in half. Each half now measures 1.5 uF because it's half the area. The DF remains unchanged because it's a ratio with capacitance. The esr, however, doubles because it's not a ratio but an absolute number. IMO, a good reason not to use esr unless you're a switching supply designer, but I digress.
So, you've got these two 1.5 uF sheet capacitors. You get two pieces of wire and connect the top and bottom electrodes of the caps together. You've paralleled them, more or less the same as before you cut them in half. You now have the original 3 uF, with the same DF and esr you started with.
The lesson is if you use the same materials, say Mylar or polypropylene, there's little gain or penalty for paralleling capacitors.
Thing get more complicated if one capacitor is of poorer quality than the other, but we try not to use poor quality capacitors! Hint- it's impossible to improve a poor quality capacitor (by that I mean high losses, DF or esr, as you please) by putting some smaller capacitor in parallel with it, so if you want good, use a single good cap, as has been said. No penalty for using a couple good caps, if you need to hit a value.
Last edited:
That's fine if the series bypass cap is so small (pf) that it's only bypassing above the audio band, then it's doing nothing to the sound.
A 10uf "mid" quality series coupling cap looking at a load of 10kohm has a LF frequency response of -3db at 1.5hz to 100's of khz.
A typical high quality bypass cap say .1uf has a frequency response into that same 10kom load of -3db at 160hz to 100's of khz.
The over lap of the two very different caps from 160hz up can be a problem because they are both working at those frequencies both with totally different characteristics.
Best to use one only, very good quality 10uf cap in the first place, and forget about the bypass cap.
Cheers George
A 10uf "mid" quality series coupling cap looking at a load of 10kohm has a LF frequency response of -3db at 1.5hz to 100's of khz.
A typical high quality bypass cap say .1uf has a frequency response into that same 10kom load of -3db at 160hz to 100's of khz.
The over lap of the two very different caps from 160hz up can be a problem because they are both working at those frequencies both with totally different characteristics.
Best to use one only, very good quality 10uf cap in the first place, and forget about the bypass cap.
Cheers George
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.