Bypassing large 1000uf electrolytic with smaller 10uf electrolytic caps

[petehall347]

Shouldn't matter, really. Indeed, you should be able to pull the caps entirely and replace the power supply with a battery and it will sound the same.

this will go off topic if i talk about batteries and mains power supplies ... i am still looking for a cheap supply of batteries to achieve +- 70v for my mosfet pa amp . its only 400 watts but can annoy folk at least 5 miles away . . everyone close by always has a good time . but thats another story .. its all just something i would like to try and see what happens .

 

[Robert2017]

i would like to see data on these improvements .like set frequencies being run through the amps and pics of scopes and distortion analysers etc .
and also ripple of the power supplies and varying voltages .
i dont doubt these improvements might happen i would really like to see why it happens . i dont have the equipment here to do valid testing or the time to do it .

I think that passing a sine wave thrue a scope will give the same result with a small difference so the electro engineer will be tempted to think that changing caps don;t worth.

But a good piece of music is 1000 times more complex compared to a sine wave so in order to measure the music a 1000x more advanced scope will be required . I don't think that such "music measuring" scope machines are invented or will ever be invented.

 

[petehall347]

I think that passing a sine wave thrue a scope will give the same result with a small difference so the electro engineer will be tempted to think that changing caps don;t worth.

But a good piece of music is 1000 times more complex compared to a sine wave so in order to measure the music a 1000x more advanced scope will be required . I don't think that such "music measuring" scope machines are invented or will ever be invented.

would i distortion analyser detect anything at speakers outputs ?
a scope would see if the voltages and frequencies were changing in the power supplies . after all this is where its coming from .

 

[Robert2017]

Assuming by "main caps" you mean power supply filter caps, there may be slight differences in discharge rate, ESR, thermal noise, terminal resistance between different filter caps, and these might have some (subtle) effect on sound.

I'm not claiming anyone is imagining anything, but such dramatic changes do not occur as result of changing one set of functioning power supply filter capacitors for another set of functioning filter capacitors of the same value, unless some caps are defective or unformed. If they were working properly and fully formed, the change in sound came from something else.

But I made this test also in another amp . This time a Yamaha and again the 1000uf filtering preamp and amp caps have been changed from the original to new ones back and forth and they where all formed and the difference was huge . And again no cap was deffective.

 

[petehall347]

any chance you could record from pre outs next time you do this and upload somewhere so i can listen to what you are ?

 

[Robert2017]

any chance you could record from pre outs next time you do this and upload somewhere so i can listen to what you are ?

I don't have the nerves to repeat the caps switch all aver again but the difference is real and I did the experiment also with a second Yamaha amp .Is all about the caps . Some caps sound pretty the same and some are very special . You should try switching the preamp filtering section caps from Nichicon KZ to KG and see what happens .

 

[HiFiThor]

Power supply caps don't "make" sound; they produce filtered DC to power the amplifier circuitry. If anything "makes" the sound, it's the amplifier transistors modulating the power supply voltage to produce the output signal, though it's really all components working as a circuit that makes the sound. Obviously, if you take out the filter caps, it will sound bad -- a loud 50 or 60hz hum, mostly.

If you've noticed a dramatic difference in sound after replacing caps, it's likely that the original caps failed and your replacement caps are now providing appropriate filtering.

It could also be a psychoacoustic effect; much like cleaning a car makes it seem to run better.

If there are parameters involved that can't be measured or explained, then we'll have to rewrite over 150 years of rigorous physics, not just electronics. Since suggestions of unknown parameters only appear in certain audio contexts that simultaneously discount known properties of human hearing, we can safely assume the perceived effects are not due to mysterious electrical effects.

I have to agree. Power supply caps are for developing a smooth and stable DC voltage for the amp rails. Good design also includes bypass caps across the rectifier bridge. In some designs I have seen small disc caps across the filter caps, AR amps come to mind. Normally you will see in a well designed amp bypass caps and inductors placed at strategic places in the amplification stages to minimize noise and induced ripple.

 

[DaveVoorhis]

I think that passing a sine wave thrue a scope will give the same result with a small difference so the electro engineer will be tempted to think that changing caps don;t worth.

But a good piece of music is 1000 times more complex compared to a sine wave so in order to measure the music a 1000x more advanced scope will be required . I don't think that such "music measuring" scope machines are invented or will ever be invented.

The classic way to test this with an oscilloscope is with a square wave, not a sine wave. Differences in amplifier performance that are invisible with a sine wave are easily visible with a square wave. It's easy to show the input to the amplifier on one 'scope channel and the output of the amplifier on another 'scope channel, and overlay them (or subtract them, with appropriate phase adjustment) to highlight differences.

However, the usual classic way to test this is not with an oscilloscope but with a distortion meter, but it can be done even more easily these days by inputting a suitable music, sine wave, square(ish) wave, white noise, etc., source and recording the output of the device under test with a good quality digital recorder or computer both before and after replacing components.

Take two recordings prior to replacing components, invert one and add it to the other. The result is the difference between the two recordings and establishes the baseline level of background noise.

Then take a recording after replacing components, invert it and add it to one of the recordings made previously. The result is the difference between the two recordings, which is the effect of replacing the components.

You can also perform spectrum analysis -- using low-cost or free software for the purpose -- on the original and after-replacement recordings, in order to clearly identify changes in frequency response.

 

[DaveVoorhis]

I have to agree. Power supply caps are for developing a smooth and stable DC voltage for the amp rails. Good design also includes bypass caps across the rectifier bridge. In some designs I have seen small disc caps across the filter caps, AR amps come to mind. Normally you will see in a well designed amp bypass caps and inductors placed at strategic places in the amplification stages to minimize noise and induced ripple.

Yup. The small caps across the rectifier bridge filter RF noise generated by the rectifier, and the small caps across the main filter caps filter low-level thermal noise generated by the filter caps. No effect on the sound of the amp, though, and many good designs don't bother.

 

[Robert2017]

Yup. The small caps across the rectifier bridge filter RF noise generated by the rectifier, and the small caps across the main filter caps filter low-level thermal noise generated by the filter caps. No effect on the sound of the amp, though, and many good designs don't bother.

ok but I readed regarding the power supply filtering caps that the higher the ripple the better it will sound .

 

[Robert2017]

Yup. The small caps across the rectifier bridge filter RF noise generated by the rectifier, and the small caps across the main filter caps filter low-level thermal noise generated by the filter caps. No effect on the sound of the amp, though, and many good designs don't bother.

Ok but if the caps have very low ESR in low frew and very very bad ERS in the highs for the power supply such ESR cap will make the preamp or amp sound the same like a cap that have good ESR on all freq ?.

I made the experiment of changing the 1000uf filtering caps in more than one amp so I know for sure that the reality kills the theory .But if you replace a Pansonic FR with a Nichichon KW maybe there will be no difference .

What do you think about this . This guy say that in the power supply a large electrolytic becomes inductive at high frequencies and its impedance rises so it needs to be shunted with a smaller solid electrolytic....

" Regardless of whether the claim of 'better sound quality' is real or not, there are basic engineering issues involved here. A power supply capacitor for a DAC will be presented with high frequency ripple from the DAC chip, probably somewhere in the 10MHz region. So, the impedance of the cap at 10MHz may be far more important than the actual value of capacitance, usually quoted for 120Hz. If the cap has a higher impedance, then the ripple generated by the DAC will not be attenuated as much.

The original cap, an FW 2200uF/35V is packaged in a case that is 16mm x 25mm. The new KW 10,000uF/50V cap is packaged into a case that is 25mm x 50mm. Without actually measuring the cap or seeing published impedance vs. frequency curves, it is still safe to say that the KW cap will have a higher impedance above 1MHz than the original FW simply because the electrodes are larger. In theory, this could lead to worse performance.

What would be a better approach is to try a 2200uF KW series cap, and parallel a smaller 10-50uF (at twice the actual PS voltage) solid tantalum or solid aluminum chip across that, and also parallel a small monolithic ceramic cap around 0.1uF-1uF. In this way, you are creating a composite capacitor that will behave like the 2200uF cap at low frequencies, but as the large electrolytic becomes inductive at high frequencies and its impedance rises, it is shunted with the smaller solid electrolytic. At still higher frequencies, when the solid electrolytic becomes inductive, the ceramic cap will take over, probably well past the operating frequencies used in the DAC. This approach will make the cap more electrically ideal at all important frequencies, and should improve sound quality in a meaningful way."

 

[DaveVoorhis]

ok but I readed regarding the power supply filtering caps that the higher the ripple the better it will sound .

Something's backwards, there. The higher the ripple, the worse it will sound, unless you like the sound of 50 or 60hz ripple.

 

[DaveVoorhis]

I made the experiment of changing the 1000uf filtering caps in more than one amp so I know for sure that the reality kills the theory .

It sounds like you may have been replacing defective caps with good ones, or there's some other cause for the change in sound. The filter caps in a power supply exist only to provide clean DC, same as you'd get from a battery. The impedance of filter caps at audio frequencies is low enough to be insignificant. There may be some nominal inductance of the leads, body and structure that vary from capacitor to capacitor, but also negligible effect at audio frequencies.

What do you think about this . This guy say that in the power supply a large electrolytic becomes inductive at high frequencies and its impedance rises so it needs to be shunted with a smaller solid electrolytic....

" Regardless of whether the claim of 'better sound quality' is real or not, there are basic engineering issues involved here. A power supply capacitor for a DAC will be presented with high frequency ripple from the DAC chip, probably somewhere in the 10MHz region. So, the impedance of the cap at 10MHz may be far more important than the actual value of capacitance, usually quoted for 120Hz. If the cap has a higher impedance, then the ripple generated by the DAC will not be attenuated as much. ... "

He's right, but he's talking about radio frequencies, and it's why small value ceramic caps are often placed very close to the supply leads of digital integrated circuits. These are called bypass caps, and they filter out RF noise produced by digital or switching circuits. They're electrically in parallel with the power supply filter caps, but they're physically located next to RF noise causing devices to filter out the noise before it's inductively or capacitively transmitted to nearby circuits and prior to the relatively high inductance of long leads or PCB traces.

At the relatively low audio frequencies at which audio amplifiers operate, they're generally not required.

 

[whoaru99]

Something's backwards, there. The higher the ripple, the worse it will sound, unless you like the sound of 50 or 60hz ripple.

I agree, but I'm thinking the intent was to mean caps with higher ripple current rating. Not sure what that would mean, if anything, for the sound but I'd think it might improve longevity.

Ripple specs need to be checked carefully though, sometimes those having big numbers are rated at high frequency and a derating factor applied for 60/120Hz duty.

 

[DaveVoorhis]

I agree, but I'm thinking the intent was to mean caps with higher ripple current rating. Not sure what that would mean, if anything, for the sound but I'd think it might improve longevity.

That makes sense. An inadequate ripple current rating would indeed cause the cap to heat up and shorten its life, but wouldn't affect sound.

 

[Robert2017]

I just absourd.

It sounds like you may have been replacing defective caps with good ones, or there's some other cause for the change in sound. The filter caps in a power supply exist only to provide clean DC, same as you'd get from a battery. The impedance of filter caps at audio frequencies is low enough to be insignificant. There may be some nominal inductance of the leads, body and structure that vary from capacitor to capacitor, but also negligible effect at audio frequencies.
.

Seems that you have a lot of knowledge but I think you miss something . Starting with 80s the big amp manufacturers instead of using normal caps they used more expansive ones.

And in many situations they used the most expansive ones exactelly in the preamp power supply and amp main caps.

I imagine that is not good for their business to use such 10x more expansive caps but they must have a good reason .

Maybe they know something that is beyound the regular conventional theory .

 

[restorer-john]

... Starting with 80s the big amp manufacturers instead of using normal caps they used more expansive ones....

Yes, to justify the higher price and appeal to audiophiles. I think they succeeded, most of us bought it hook line and sinker.

...And in many situations they used the most expansive ones exactelly in the preamp power supply and amp main caps...

Nice big caps with gold lettering and pretty colours, cool names like Fine Gold, Muse, Cerafine etc. Bigger value capacitors equate to better performance in an audiophile's mind.

...I imagine that is not good for their business to use such 10x more expansive caps but they must have a good reason...

I imagine it was very good for their business, and the cap supplier's business too. You think they didn't make a bigger profit on the more expensive components?

...Maybe they know something that is beyound the regular conventional theory .

Conventional theory on capacitors is well understood, has been for literally centuries. It is not understood very well by people who get their 'knowledge' from the internet, subjective reviewer tests, and sites with vested interests in purveying 'magic' capacitors.

Once you've do some objective testing, using methods like DaveV espouses above, you will be surprised.

Have fun.

 

[petehall347]

to be a little fair many older capacitors were specified at +100 percent

 

[Robert2017]

Yes, to justify the higher price and appeal to audiophiles. I think they succeeded, most of us bought it hook line and sinker.



Nice big caps with gold lettering and pretty colours, cool names like Fine Gold, Muse, Cerafine etc. Bigger value capacitors equate to better performance in an audiophile's mind.



I imagine it was very good for their business, and the cap supplier's business too. You think they didn't make a bigger profit on the more expensive components?



Conventional theory on capacitors is well understood, has been for literally centuries. It is not understood very well by people who get their 'knowledge' from the internet, subjective reviewer tests, and sites with vested interests in purveying 'magic' capacitors.

Once you've do some objective testing, using methods like DaveV espouses above, you will be surprised.

Have fun.

You can;t test the difference with such methods using "analizers". The only way is to make accurate recordings befor and after and to use a reference amplifier and to make blind tests .

You have some very fixed ideas . How can you possibly be so wrong and to not be able to see that.

For example a Panasonic FM sounds very different compared to a Silmic 2.

In a speaker crossover a signal electrolitic will sound very bad compared to a paper in oil which sound a lot more musical while the electro sound dry and borring .

It seems that some people have really bad ears and they can;t hear right .

Seems I;m lucky to have such good ears/.

 

[restorer-john]

...Seems I;m lucky to have such good ears...

We are all lucky to have good ears, and even many of the guys here who have some hearing loss still love and enjoy their gear and their music. It doesn't take good ears to hear differences, it takes trained ears and correct interpretation of what you hear.

This place is full of people with varying views, beliefs and points of view- that's what makes it interesting.

Welcome to AK.