Bypassing large 1000uf electrolytic with smaller 10uf electrolytic caps

[Robert2017]

Hello
I just purchased a NAD 306 amp .
I replaced the 2x 1000uf main preamp caps with 2 Nichicon KG III super through .
After that mod the sound has become huge with amazing bass but the high freq was bad with bad detail in high mid-high frequencies .
After that I added 2 x10uf Nichicon KZ electro's in paralel with the 1000uf KG in order to restore the high freq detail and it worked great . Now I have huge bass and also great detail .
Now I was just thinking to add 2 more non-polar 10uf film caps to push the details and clarity even further .
If I add the 2 non-polar film caps there is a danger to kill the amp or something ??.
I have little experience with electronics and I know that this mod in not conventional .
I also added a photo to show exactly how I bypassed the main preamp caps and I also added another photo from the service manual with the 1000uf and 470uf caps location . Please check .
What do you think ?
I know that such mods can add bad resonant freq but until now everithing is great.
I also plan to bypass the 2 x 470uf preamp caps using the same method and also the 2x 220uf main amplifier caps.
I used 10uf smaller electrolytics in conjunction with the 1000uf bigger ones because the 10uf are able to pass the high freq without distortion .
Regards

 

[evsentry3]

Bypassing as you are describing was a technique used in Tektronics oscilloscopes for wide bandwidth circuits. They would have a whole series of caps in descending generation like that. A valid technique when used correctly. I believe the accepted practice is 100th's. Yes, it does create different poles...meaning some timing or phase change.

You don't mention if this is a power supply filter or an output DC blocking cap. Different places in electronic circuits different techniques are used for various reasons.

EV3

 

[Robert2017]

Bypassing as you are describing was a technique used in Tektronics oscilloscopes for wide bandwidth circuits. They would have a whole series of caps in descending generation like that. A valid technique when used correctly. I believe the accepted practice is 100th's. Yes, it does create different poles...meaning some timing or phase change.

You don't mention if this is a power supply filter or an output DC blocking cap. Different places in electronic circuits different techniques are used for various reasons.

EV3

Hi . I also added some photos from the service manual with the exact caps location . I have little experience . Maybe those 1000uf are the main filter preamp caps ???. The sound has improved immensely after adding just those 2x10uf electro's. In fact I can't believe how good it sounds now. The sound have a lot of bass depth and also a lot of detail .Thanks

 

[toxcrusadr]

I'll wait for someone else to comment on where this is in the circuit. Curious, though, where you learned which ones to bypass.

As for whether adding another 10 uf will help or hurt: It may make no further improvement on the sound. Assuming the first did (and I'm not doubting you), more of a good thing is not necessarily helpful in this context. Will it hurt? 10 uf compared to 1000 is 1%. Caps are usually 10-20% precision, right? So a 1-2% error is well within tolerances. Generally speaking a typical amp circuit can handle a lot more variation than that.

 

[petehall347]

the ones you circled are power supply filters .

 

[E-Stat]

My experience is that:
1. The value doesn't have to be very high
2. Use film caps for bypass duty

For a Threshold Stasis 3 amp I used to have, I bypassed the big 25,000 uF Mallory caps with 1 uF films.

 

[Robert2017]

I'll wait for someone else to comment on where this is in the circuit. Curious, though, where you learned which ones to bypass.

As for whether adding another 10 uf will help or hurt: It may make no further improvement on the sound. Assuming the first did (and I'm not doubting you), more of a good thing is not necessarily helpful in this context. Will it hurt? 10 uf compared to 1000 is 1%. Caps are usually 10-20% precision, right? So a 1-2% error is well within tolerances. Generally speaking a typical amp circuit can handle a lot more variation than that.

If the 2x 1000uf caps are dealing with the whole preamp freq means they are unable to also deliver good hights because the bass is distorting the high and mid details inside the cap . So adding 2x 10uf small ones will create a bridge for very clear undistorted details that are able to pass . The difference is huge . Maybe in theory is not worth such mod but I know that 10uf are abble to pass a lot of high and mid-high freq information . So I was just thinking to add another 2x10uf caps but film caps in order to pass even more information . I;m just afraid that something bad can happen because those are non-polar and not a very small value.

 

[Robert2017]

My experience is that:
1. The value doesn't have to be very high
2. Use film caps for bypass duty

For a Threshold Stasis 3 amp I used to have, I bypassed the big 25,000 uF Mallory caps with 1 uF films.

What about using some bigger values like 10uf or 20uf for the film caps ??. There is any danger ?. A 10uf or 20uf cap will be able to pass a lot more information compared to a 1uf cap that is passing only a small high freq band .

 

[evsentry3]

In a power supply what the caps do is filtering "noise". It's not the same as what audio is passing through when they are DC blocks between stages. A filter cap is pointed at being an effective low pass filter. The "problem" with a big reservoir electrolytic is they aren't as effective at higher frequencies. The idea of the bypass is to keep the noise "blocking" more effective at higher frequencies.

And as bypass caps, larger is not necessary the goal. It's to keep the impedance uniform when the audio stage pulls more current quicker (higher freq content). So you're better off to have them be fairly high quality (decent film) caps close to that fractional value.

EV3

 

[Robert2017]

Is ok to bypass the other preamp 470uf caps with 4,7uf film cap and the main amplifier 2x220uf electrolytic caps with 2,2uf film cap ?. I have some mkp audiophiler caps made by Mundorf .

I just want to increase and expand every detail music have in order to sound best.

 

[DaveVoorhis]

Is ok to bypass the other preamp 470uf caps with 4,7uf film cap and the main amplifier 2x220uf electrolytic caps with 2,2uf film cap ?. I have some mkp audiophiler caps made by Mundorf .

I just want to increase and expand every detail music have in order to sound best.

You won't be "bypassing" the caps as such. What you're doing is increasing their capacity. A 470mfd cap with a 4.7mfd cap in parallel is nominally a 474.7mfd cap, which is well within the tolerance range of the original 470mfd cap. It's effectively harmless, though paralleling large caps with small caps is a valid technique for lowering ESR at higher frequencies and more effectively filtering RF noise, particularly where digital circuits are used.

Digital circuits change value (from 0 to 1 and back) rapidly, generating RF noise. To filter out the noise, a small cap -- say 0.01uf -- will typically be physically placed as close as possible to the supply pin(s) of the integrated circuit that produces the noise. This is called a bypass capacitor. They are sometimes shown on schematics as small capacitors in parallel with the large power supply capacitors -- and electrically they are -- but each one is physically located next to an IC, to filter RF at its source and avoid broadcasting it to the rest of the circuit.

Therefore, putting a 4.7mfd cap in parallel with a 470mfd cap probably won't increase and expand musical detail per se, but it might slightly aid noise filtering at relatively higher frequencies, depending on the role of the 470mfd cap in the circuit.

 

[DaveVoorhis]

If the 2x 1000uf caps are dealing with the whole preamp freq means they are unable to also deliver good hights because the bass is distorting the high and mid details inside the cap . So adding 2x 10uf small ones will create a bridge for very clear undistorted details that are able to pass . The difference is huge . Maybe in theory is not worth such mod but I know that 10uf are abble to pass a lot of high and mid-high freq information . So I was just thinking to add another 2x10uf caps but film caps in order to pass even more information . I;m just afraid that something bad can happen because those are non-polar and not a very small value.

This doesn't really make sense. Power supply filter capacitors do not "pass information". They filter the unfiltered DC emitted by the power supply rectifier that converts AC to DC, so that it's smooth DC. There is some argument that small non-electrolytic caps might discharge faster than large ones and thus effectively maintain supply levels for rapid transients, but this is unlikely to have any effect in the range of audio frequencies -- it might be a concern if you've built a power supply to provide high-current pulses of nanosecond duration, which puts them in the gigahertz frequency range. Not relevant for audio.

 

[Robert2017]

This doesn't really make sense. Power supply filter capacitors do not "pass information". They filter the unfiltered DC emitted by the power supply rectifier that converts AC to DC, so that it's smooth DC. There is some argument that small non-electrolytic caps might discharge faster than large ones and thus effectively maintain supply levels for rapid transients, but this is unlikely to have any effect in the range of audio frequencies -- it might be a concern if you've built a power supply to provide high-current pulses of nanosecond duration, which puts them in the gigahertz frequency range. Not relevant for audio.

Ok I understand this theory but the difference in sound now is really huge . I added to more 10uf Nichicon KZ electrolytics per 1000uf cap . Now there are 2x10uf KZs for every 1000uf KG cap. Evsentry3 already mentioned at the beginning of this thread that is a " A valid technique when used correctly". Someone told me that the power supply caps made the sound . I come with those ideas also after I replaced some 2x 1000uf caps but from another amp . So I replaced some 1000uf Nichicon KZs with 2x old 1000uf Black Gates and after that everything changed . The sound has taken a new dimension from 2d to extra 3d. So I don't think is all about smoothing DC from the mains. Maybe there are some other parameters involved that can't be measured or explained ?.

 

[Robert2017]

You won't be "bypassing" the caps as such. What you're doing is increasing their capacity. A 470mfd cap with a 4.7mfd cap in parallel is nominally a 474.7mfd cap, which is well within the tolerance range of the original 470mfd cap. It's effectively harmless, though paralleling large caps with small caps is a valid technique for lowering ESR at higher frequencies and more effectively filtering RF noise, particularly where digital circuits are used.

Digital circuits change value (from 0 to 1 and back) rapidly, generating RF noise. To filter out the noise, a small cap -- say 0.01uf -- will typically be physically placed as close as possible to the supply pin(s) of the integrated circuit that produces the noise. This is called a bypass capacitor. They are sometimes shown on schematics as small capacitors in parallel with the large power supply capacitors -- and electrically they are -- but each one is physically located next to an IC, to filter RF at its source and avoid broadcasting it to the rest of the circuit.

Therefore, putting a 4.7mfd cap in parallel with a 470mfd cap probably won't increase and expand musical detail per se, but it might slightly aid noise filtering at relatively higher frequencies, depending on the role of the 470mfd cap in the circuit.

OK but in speakers crossovers in order to pass the high freq details usually a 4,7uf film cap is used for the tweeter .

In a speaker the mids speaker freqency filtering capacitor usually is from 20uf to 50uf and such caps are not able to pass ultra clear hights maybe because inside the cap there are to much frequencies and maybe inside the cap the high freq are now distorted so their duty is to pass their freq to a mid freq speaker that is exposing only the mid freq by hiding the distorted highs ?.

So in theory if a 330uf cap is passing the whole signal in such case the highs will be able to pass undistorted ?. Or because the cap is dealing with the whole freq (intense current) the highs and crisp detail are in a big procent crushed and mushed out in the big cap ??

I know that in amps caps are acting different compared to a speaker crossover but if a singnal cap inside the amp is having large UF passing band, such cap will be abble to also pass the very high freq details undistorted ?.

 

[Beau Geste]

Forgive my intrusion into the discussion but for me the real news is that the Nichinon KG III has high frequency leaks and should not be so. www.nichicon.co.jp/english/products/pdf/e-kg.pdf
Regards.

 

[DaveVoorhis]

Ok I understand this theory but the difference in sound now is really huge . I added to more 10uf Nichicon KZ electrolytics per 1000uf cap . Now there are 2x10uf KZs for every 1000uf KG cap. Evsentry3 already mentioned at the beginning of this thread that is a " A valid technique when used correctly". Someone told me that the power supply caps made the sound . I come with those ideas also after I replaced some 2x 1000uf caps but from another amp . So I replaced some 1000uf Nichicon KZs with 2x old 1000uf Black Gates and after that everything changed . The sound has taken a new dimension from 2d to extra 3d. So I don't think is all about smoothing DC from the mains. Maybe there are some other parameters involved that can't be measured or explained ?.

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.

 

[DaveVoorhis]

OK but in speakers crossovers in order to pass the high freq details usually a 4,7uf film cap is used for the tweeter .

Speaker crossovers and power supply filters are very different circuits.

A speaker crossover is often configured as a high-pass filter; the signal passes through the crossover capacitor on its way to a speaker.

A filter capacitor in a power supply provides nominally ripple-free DC to power the system. The signal does not pass through it.

However, the generation of amplifier output does load the power supply, sometimes in synchronisation with the output signal. A power supply with inadequate filtering can degrade the output signal by superimposing ripple. A power supply with inadequate current capacity can degrade the signal by causing it to be clipped.

 

[DaveVoorhis]

Forgive my intrusion into the discussion but for me the real news is that the Nichinon KG III has high frequency leaks and should not be so. www.nichicon.co.jp/english/products/pdf/e-kg.pdf
Regards.

What do you mean by "high frequency leaks"?

 

[Robert2017]

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.

You didn't answer this question .Regarding amplifiers if a 330uf cap is passing the whole signal in such case the highs will be able to pass undistorted ?. Also before installing the 1000uf KG caps I made some measurements and ESR was 0.01 and Vloss 1.2% so I presume they are ok because de bass and sound depth was much higher compared with KZ. The problem was only with the bad high freq that was restored using those 10uf KZ caps in parallel.

 

[Beau Geste]

What do you mean by "high frequency leaks"?

I mean this :"After that mod the sound has become huge with amazing bass but the high freq was bad with bad detail in high mid-high frequencies"
Regards.