Arguing measurements dont tell the whole story is one thing, arguing that measurements are inaccurate or problems don't exist simply because many people do something and/or they think it benefits the sound is another.
I used to think bypassing was good, I think it can have a minor benefit to sound, but slowly came to the realisation that the dominant and consistent effect is adding glare and harshness, easily mistaken for 'detail'.
If you favour a more balanced and smooth sound, that does not become fatiguing or tiring to listen after days/weeks/months, I wouldn't recommend bypassing.
Better to use as few caps as possible to get the job done, best thing you can try is a small electrolytic (low capacitance, high ESR and low inductance ) as close to the load as possible, if supply voltage permits it maybe a solid polymer as they are known to have better high frequency characteristics, watch out for the ESR.
If a circuit sounds dull without a bypass there is probably some other problem you need to address.
I used to think bypassing was good, I think it can have a minor benefit to sound, but slowly came to the realisation that the dominant and consistent effect is adding glare and harshness, easily mistaken for 'detail'.
If you favour a more balanced and smooth sound, that does not become fatiguing or tiring to listen after days/weeks/months, I wouldn't recommend bypassing.
Better to use as few caps as possible to get the job done, best thing you can try is a small electrolytic (low capacitance, high ESR and low inductance ) as close to the load as possible, if supply voltage permits it maybe a solid polymer as they are known to have better high frequency characteristics, watch out for the ESR.
If a circuit sounds dull without a bypass there is probably some other problem you need to address.
Last edited:
The guy just want to bypass all his 220 uF/400C caps on the B+ an amp ! He should better read or ask first about what is ripple and noise... inductence will come after imho.
"Want" is a strong word. I just came to the realization that there are a million different takes on this, scientific not-so-scientific, theoretical. subjective. Some say inductance is an issue, others say it's not the inductance the issue is something else, others say there will be no additional inductance. And if someone comes forward who has built amps for years and says it's a good thing to do it's automatically snake oil because they are biased towards their own product.
Personally, I was convinced it would bring a big improvement. But now after talking with you guys about it, I am not so sure anymore.
Like some people say it's science that it brings an improvement others say it's science that it doesn't.
We will see
Last edited:
here we have a typical amp circuit with 100u and 100n for each power rail.
are we saying that the 100nF is not needed and/or detrimental?
are we saying that the 100nF is not needed and/or detrimental?
Yes and no.
I said earlier: That said, this kind of bad practice rarely has directly noticeable adverse consequences: the overall supply impedance generally remains low enough to avoid serious problems
You talk about amplifiers, and most regular amplifiers and opamps generally have a good PSRR and are relatively insensitive to the impedance of the supply lines, but more "exotic" topologies might not be that tolerant.
Amplifiers are not the only breed of loads though: some high-perf ADC and DACs require an absolutely perfect bypass at all frequencies, up to tens of MHz to deliver their advertised performance.
You can also see the problem from another perspective: many modern, high-perf LDO's do not tolerate output caps falling outside certain values, types or combinations: they will simply oscillate.
Combinations are particularly problematic, because they generate multiple peaks and dips in the impedance.
The denoiser is an example of high-perf, composite regulator. It has to be properly bypassed too, and doesn't tolerate fancy combinations. Member RickTH learned it the hard way: despite my explicit advice not to double up the bypasses, he did it anyway because "Elektor did it all the time". As a result, he spent a few hours scratching his head, wondering what was wrong (he has no oscilloscope).
Even with a mundane and placid 317 or 337, the wrong output bypass causes a peak in the spectrum of the output noise. Generally unnoticeable and unnoticed, but the result can be a puzzling loss of performance which has no apparent explanation unless you know what to look for
Not needed, certainly, detrimental not in this case: the impedance variation will be harmless there
OF COURSE this is science but it is complex enough to doubt for wanting to find too much simplier answers. I do the same btw before crossing a little level step by step...30 y ago I was bypassing everything because in the magazines I was seeing this cool red caps...our brains want to make shortcuts because we are lazy...I mean it is not as easy we want it to be.
In the shematic that doesn't show the pcb layout btw, I think it is not necessary but who knows without probing ? If the 1000 uf is already local it should be enough. This capacitance parralleling seems here a bad practice. Now not sure you can hear it...you migth or not. Just the tools can say if needed ir if it is ringing.
In the shematic that doesn't show the pcb layout btw, I think it is not necessary but who knows without probing ? If the 1000 uf is already local it should be enough. This capacitance parralleling seems here a bad practice. Now not sure you can hear it...you migth or not. Just the tools can say if needed ir if it is ringing.
Last edited:
Well ,you're missing the accountant. Every dollar spent turns into ten dollars retail.
If so, then also a tiny bit of capacitance doesn't have that big of an impact either.
Sorry, I mistakenly thought we were in DIY Audio, not DIY RADIO
Cost, size, and endurance. Those caps you linked to have an endurance rating of 500 hours + 500 hours at 1.3 x VNDC and 70°C and they wouldn't fit in the chassis 🙂
Same thing goes for the internet.....
Unless you're a professional, schooled in and certified, those online "he said to do this" will make you into a nutcase.
Electrolytic caps have improved a lot since 1940. I do remember when an e-cap got useless by 1MCPS, but today they are no more inductive than a wire a little longer than their lead spacing.
But some old notions never die.
There are still people who believe that a bypass capacitor has to be capacitive and that electrolytic capacitors are therefore unusable from the frequency where the capacitive reactance drops below the ESR. That's false, of course, the impedance just has to be low enough. In fact a small resistive impedance damps resonances and is in that sense better than a capacitive impedance.
No, you can not question a magazine while with internet you can question. The learning experience is not the same. In fact this istotally different.
It's operational life expectancy unquestionable exceeds mine.
Attachments
So I'd like to see for myself and let my ears be the judge.
With what capacitance do you guys suggest I'd try? 0.1uf - 1uF - 10uf?
Or maybe 0.47uf instead of 10 for a middle ground between 0.1 and 1?
3x 10uf caps are a bit pricey
What do you guys think?
With what capacitance do you guys suggest I'd try? 0.1uf - 1uF - 10uf?
Or maybe 0.47uf instead of 10 for a middle ground between 0.1 and 1?
3x 10uf caps are a bit pricey
What do you guys think?
I'm assuming the amp, unlike me, isn't powered up continuously for 22 years. You can never know though.