Please explain me what does mean "Bypass" caps. I have seen they are usually very low capasitance, like 0.0...yF.
Why I am asking is that, I just got some Vishay caps MKT 373 and would like to get your opinion-can I use those in my crossover?! Are they for audio..?!
http://www.vishay.com/docs/28111/mkt373.pdf
And one thing I am intrested in - I need 150yF, is it reasonable to use 10x 15yF caps..parallel?! Does it has some effect to the sound, if I am using 10 small values caps except 1 big value...
Why I am asking is that, I just got some Vishay caps MKT 373 and would like to get your opinion-can I use those in my crossover?! Are they for audio..?!
http://www.vishay.com/docs/28111/mkt373.pdf
And one thing I am intrested in - I need 150yF, is it reasonable to use 10x 15yF caps..parallel?! Does it has some effect to the sound, if I am using 10 small values caps except 1 big value...
By using multiple cups to achieve a final value, technically you are cutting down negative qualities of the capacitors such as inductance and resistance. By the book: it is a better way of doing a crossover.
However, it is usually not necessary from the power handling point of view.
It is also very questionable if it improves sound. Some people believe it does, some people believe it creates parasitic harmonics.
Bypass capacitor is usually more expensive capacitor (presumably of a better quality) used in parallel with main capacitor on the tweeter. It is done in order to save money and have a “high-end” sound.
I've done it both ways and if you use good capacitors such as Solen or Clarity Cap, you will be either way.
So, there's no straight answer as you see.
However, it is usually not necessary from the power handling point of view.
It is also very questionable if it improves sound. Some people believe it does, some people believe it creates parasitic harmonics.
Bypass capacitor is usually more expensive capacitor (presumably of a better quality) used in parallel with main capacitor on the tweeter. It is done in order to save money and have a “high-end” sound.
I've done it both ways and if you use good capacitors such as Solen or Clarity Cap, you will be either way.
So, there's no straight answer as you see.
Member
Joined 2004
It's a common trick in loudspeaker crossovers:
The Vishay costs 0.50 cents a piece. Use them in series with a capacitor for the tweeter. Don't mind the small 0.01 uF difference. The quality of the bypassed cap is no indicator for this trick to work or not.
It's a kind of love it or hate "tweak", results may vary.
Regards
Roland
Caps&Coils
Humble Home Made Capreview
The Vishay costs 0.50 cents a piece. Use them in series with a capacitor for the tweeter. Don't mind the small 0.01 uF difference. The quality of the bypassed cap is no indicator for this trick to work or not.
It's a kind of love it or hate "tweak", results may vary.
Regards
Roland
Caps&Coils
Any capacitor can be used as a bypass capacitor. It is called a bypass because its typically of a much smaller value, lets say .1uf or .01uf, and gives an alternative electrical path to the main cap. Since caps behave differently at different frequencies, and this is dependant on the value rating, this means that a smaller value will give a by-pass, or alternate route at certain frequencies. The negative side of a bypass is that these electrons will flow through each capacitor differently at the same frequency, and thus creates something known as different time constants. In other words, the amount of time it takes an electron to pass through one capacitor will be different than the other. This is thought to create potentially audible distortion, and so some people argue that the improved resolution is nothing more than distortion, and is actually a bad thing.
Member
Joined 2004
pjpoes said:
On a objective basis, sure. But if you listen, you will not get the same "vishay mkp 1837" effect with a 0.10 uF mkp from Solen, Clarity, Mundorff or other mkp caps. I have tried other bypass caps, not the same effect.
Did anyone here try this bypass cap?
An externally hosted image should be here but it was not working when we last tested it.
Never heard it. It's quite expensive.
My personal believe is that you need to use cups of the same value. For example if you need to achieve 10mfd value, use 2 of 5mfd or 3 3.3mfd. I think the main reason would be to cut down DCR and inductance of the capacitors, especially if you need to get to rather large value.
The whole idea of a bypass capacitor reminds me of a bypass drain pipe. Unfortunately it doesn't exactly work the same way.
The common practice would be to use, let's say Solen cap and parallel it with Mundorf of a very small value and there for smaller price sticker. It is believed by some that this way you could have best of both worlds: great sound of a very expensive boutique Mundorf and high MFD using cheap Solen.
IN the Double blind listening test the audible difference is yet to be heard.
BTW, capacitors are parallel to each other and you only need to be concern (if?) with capacitors that are in a series with tweeter. Zobel networks and woofer part of the crossover will not benefit from it at all.
Itertechnic parts aren't available in US, so I am not sure but I am quite skeptical of a $$$ parts unless it's Scan Speak or Seas.
The whole idea of a bypass capacitor reminds me of a bypass drain pipe. Unfortunately it doesn't exactly work the same way.
The common practice would be to use, let's say Solen cap and parallel it with Mundorf of a very small value and there for smaller price sticker. It is believed by some that this way you could have best of both worlds: great sound of a very expensive boutique Mundorf and high MFD using cheap Solen.
IN the Double blind listening test the audible difference is yet to be heard.
BTW, capacitors are parallel to each other and you only need to be concern (if?) with capacitors that are in a series with tweeter. Zobel networks and woofer part of the crossover will not benefit from it at all.
Itertechnic parts aren't available in US, so I am not sure but I am quite skeptical of a $$$ parts unless it's Scan Speak or Seas.
The timing thing does not make any sense to me as voltage at certain frequency would pass a large cap or a small cap for the same delay. Various frequencies have various time delays, not with various capacitance of the capacitors. So parallelling capacitors of different capacitance does not introduce time error.
The idea of bypassing a cap is to try to obtain a flatter impedance at all frequencies. e.g. Electrolytic caps usually have a high inductance (impedance) towards higher frequencies. By passing a high impedance capacitor with a small film capacitor (which has a much lower impedance at high frequencies) lowers the impedance at high frequencies to come up with a flatter impedance.
Because the following or related stage do not usually have infinite impedance or power supply rejection ratio, any variations on the impedance of the caps cause errors.
For example, bypassing a power supply cap will make high frequencies passing through the cap much easily due to the lower impedance. This translates to possibly higher volume of sound at higher frequencies (sparkle) if the amplifier / buffer does not have infinite power supply rejection ratio. This may improve your system, or make it sound worse.
So there is no golden rule like "100uF + 1uF + 0.01uF", etc. They all depend on the specific capacitors. For example, if you bypass a 2200uF Rubycon ZL (which has a very low impedance of 0.015R at 100kHz) with a 1uF or 0.1uF film, you will find the higher treble to be terribly exaggerated. A 0.01uF may work better.
The difficulty is to find the caps that compliment each other to come up with a flat impedance. Manufacturers' datasheets do not give you sufficient data. Blindly bypassing capactors have much better chance to fail than to succedd. Experiments are needed to come up with the best combination. This process is very tedious and time consuming.
Check my latest thread regarding caps at the Electronic Parts section of this forum. It was only posted today.
Regards,
Bill
The idea of bypassing a cap is to try to obtain a flatter impedance at all frequencies. e.g. Electrolytic caps usually have a high inductance (impedance) towards higher frequencies. By passing a high impedance capacitor with a small film capacitor (which has a much lower impedance at high frequencies) lowers the impedance at high frequencies to come up with a flatter impedance.
Because the following or related stage do not usually have infinite impedance or power supply rejection ratio, any variations on the impedance of the caps cause errors.
For example, bypassing a power supply cap will make high frequencies passing through the cap much easily due to the lower impedance. This translates to possibly higher volume of sound at higher frequencies (sparkle) if the amplifier / buffer does not have infinite power supply rejection ratio. This may improve your system, or make it sound worse.
So there is no golden rule like "100uF + 1uF + 0.01uF", etc. They all depend on the specific capacitors. For example, if you bypass a 2200uF Rubycon ZL (which has a very low impedance of 0.015R at 100kHz) with a 1uF or 0.1uF film, you will find the higher treble to be terribly exaggerated. A 0.01uF may work better.
The difficulty is to find the caps that compliment each other to come up with a flat impedance. Manufacturers' datasheets do not give you sufficient data. Blindly bypassing capactors have much better chance to fail than to succedd. Experiments are needed to come up with the best combination. This process is very tedious and time consuming.
Check my latest thread regarding caps at the Electronic Parts section of this forum. It was only posted today.
Regards,
Bill
Member
Joined 2004
owdi said:
You are right...Excuse me, it's like mixing up left and right. In a crossover the place to bypass is indeed the tweeter section.
I mixed a lot of caps when younger to get needed values not at hand. I never heard something special happen when comparing 6.80 uF and some 2.20 and a 0.22 uf (all the same brand) to get to another 6.80 uF cap for speaker nr 2. A small silver mica or vishay parallel i do hear. (don't like silver mica)
I agree there's no golden rule like 10.0 +1.0 + 0.10 uF. Especially considering taste and the speaker (stereo, etc, etc) it's used in as extra variables.
HiFiNutNut said:
You seem to forget that DCR changes with frequency and is variable with capacitor value. This among other issues are the fundamental way in which a capacitor creates the phase shift we take advantage of is crossovers. Another example is the use of a cap to create the necassary phase shift for an AC motor.
http://www.hobby-electronics.info/course/html/ch04s05.html this might also help you understand the relationship between capacitor values, frequency, and the phase shift point (Delay).
By the way, I was trying to give the most objective explanation of what bypass caps are and what is meant by that. When I said that technically any cap could be a bypass cap, that was true.
Of all companies, Sony did some research into paralleling multiple power supply capacitors in a sort of cascade arrangement to minimize ESL and ESR within the working frequencies. While I don't have the equipment to properly test this at the moment, it can be modeled in spice software with equivalent circuits and appears effective.
Also, while I agree that bypass caps can sometimes appear to improve the sound, the claimed negative can in fact be modeled. I would imagine it can be measured as well. If you model paralleled capacitors in a cascade arrangement with huge differences, lets say 100uf and .01uf's, and this is used in the series element of a linkwitz type crossover slope, and looked at the phase plot, you see it is changed, even though total capacitance is the same.
By the way, I was trying to give the most objective explanation of what bypass caps are and what is meant by that. When I said that technically any cap could be a bypass cap, that was true.
Of all companies, Sony did some research into paralleling multiple power supply capacitors in a sort of cascade arrangement to minimize ESL and ESR within the working frequencies. While I don't have the equipment to properly test this at the moment, it can be modeled in spice software with equivalent circuits and appears effective.
Also, while I agree that bypass caps can sometimes appear to improve the sound, the claimed negative can in fact be modeled. I would imagine it can be measured as well. If you model paralleled capacitors in a cascade arrangement with huge differences, lets say 100uf and .01uf's, and this is used in the series element of a linkwitz type crossover slope, and looked at the phase plot, you see it is changed, even though total capacitance is the same.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.