Remember that there are two factors influencing the (possible) differences in the sound of components such as capacitors. They are electrical and mechanical.
We can't do much about the electrical factor but has anyone experimented with the mechanical? eg adding mass.
I have just removed 8 capacitors from the PSU for my GC buffer circuit.
The PSU had two stages of regulation using LM317/337 regs and I had read somewhere about adding some extra 220 uF caps to lower the noise from the regulators. Unfortunately, they only added some distortion and now I am wondering if this was because I used some very cheap caps that have been in my parts box for five or six years!
Still on the subject of 'do caps make a difference to the sound', it is always worth remembering that
2 + 3 +11 +1 +6 = 23 as does 5 + 10 + 2 + 6!
In other words, there are many combinations of numbers that arrive at the same total.
We can't do much about the electrical factor but has anyone experimented with the mechanical? eg adding mass.
I have just removed 8 capacitors from the PSU for my GC buffer circuit.
The PSU had two stages of regulation using LM317/337 regs and I had read somewhere about adding some extra 220 uF caps to lower the noise from the regulators. Unfortunately, they only added some distortion and now I am wondering if this was because I used some very cheap caps that have been in my parts box for five or six years!
Still on the subject of 'do caps make a difference to the sound', it is always worth remembering that
2 + 3 +11 +1 +6 = 23 as does 5 + 10 + 2 + 6!
In other words, there are many combinations of numbers that arrive at the same total.
There's another very good option: Ansar capacitors.
http://www.cricklewoodelectronics.co.uk/Cricklewood/customer/home.php?cat=155
They seem to be sold just by Cricklewood, but they come in up to 80uF and are 400v. For higher voltages (?) you can always do a series.
Kuei Yang Wang is probably listening and can comment about them.
Carlos
http://www.cricklewoodelectronics.co.uk/Cricklewood/customer/home.php?cat=155
They seem to be sold just by Cricklewood, but they come in up to 80uF and are 400v. For higher voltages (?) you can always do a series.
Kuei Yang Wang is probably listening and can comment about them.
Carlos
Nuuk said:
Any particular reason Mr Carlos?
No.
Just because I like and I use them.
Seams the right value for me, not too much, not too little.
100uf it is.
Anyway, it's the most common value I see on commercial products after an LM317/337.
I know this doesn't mean anything but...
Nuuk, just try it and tell us, it seams you have there a particularly sensitive circuit to test this, so...
Christer said:
Sounds like real engineering.
I'm not an engineer, thanks.
Something against that?
Engineers measure and get satisfied when it measures right.
I only get satisfied when it sounds good to me.
Something against that?
If you don't have anything useful to say, shut up.
Or else, what's the ideal capacitance after an LM317/337 regulator?
And why?
Tell us, Mr. Engineer.
fdegrove said:
Does it matter what class the amp is running in for the powersupply caps to matter?
Not really, they're always there those nasty electrolytics, aren't they?
So yes, it would matter more in a class B or class AB amp than in a pure class A design but it matters nonetheless.
Nobody ever built an amp with nothing but polyprops in the PSU?
Guess not.
I'll be the first to admit it's easier to do with tubes due to the much higher rail voltages used but the difference is easily heard...
Thankfully electrolytics have much improved over the years but filmcaps haven't exactly been sitting still either.
Give it a try guys,
Well, there are a couple of ways of avoiding electrolytics feeding output transistors directly.
1) Capacitance multiplier
2) Switching power supply
I've heard good things about CMs although I haven't tried them. There are persistent rumours that SPSs have exceptional characteristics - the PanasonicXR45 uses one to feed its class D output stages, which have negligeable PSRR, and people seem to like the thing. Comments?
Francois.
An engineer (or engineering wannabe like me ) would
have taken measurements with a distortion analyzer
and an oscilloscope if it didn't 'sound right'.
Might have caught some oscillation. Simply tinkering
blind without testing to confirm normal operation is
risky. I know most DIYers don't have any test gear
beyond a DMM nor much technical knowledge, but
that's the chance you take.
To mangle an old saying, measure twice, then
listen. Lots of engineers do this, too. I did extensive
modifications on my all-JFET Hafler SE-100 preamp,
in the process discovering my mods had caused
one of those wretched 7815s to go unstable, which
a local 10 uF cap on the output cleared up.
Call it double confirmation. And it did sound better
afterwards.
(Hey, Carlos! You know what they say about
'French military victorys'...Cinco de Mayo! )
) wouldhave taken measurements with a distortion analyzer
and an oscilloscope if it didn't 'sound right'.
Might have caught some oscillation. Simply tinkering
blind without testing to confirm normal operation is
risky. I know most DIYers don't have any test gear
beyond a DMM nor much technical knowledge, but
that's the chance you take.
To mangle an old saying, measure twice, then
listen. Lots of engineers do this, too. I did extensive
modifications on my all-JFET Hafler SE-100 preamp,
in the process discovering my mods had caused
one of those wretched 7815s to go unstable, which
a local 10 uF cap on the output cleared up.
Call it double confirmation. And it did sound better
afterwards.
(Hey, Carlos! You know what they say about
'French military victorys'...Cinco de Mayo!
)
Hey Carlos,
You have a funny idea about engineers. Basically an engineer tries to get the best performance by intelligently combining techniques, parts, layout, etc. Measurements are a great help to see if you are going the right way.
That doesn't mean that it must replace listening, of course. But saying hey, I tried 100uF, and it sounds good doesn't seem a very smart way to try to get the best from what you have available. I mean, maybe 50uF sounds even better! Measurements and knowing what you are doing can help you see in which direction and how far to go.
Jan Didden
You have a funny idea about engineers. Basically an engineer tries to get the best performance by intelligently combining techniques, parts, layout, etc. Measurements are a great help to see if you are going the right way.
That doesn't mean that it must replace listening, of course. But saying hey, I tried 100uF, and it sounds good doesn't seem a very smart way to try to get the best from what you have available. I mean, maybe 50uF sounds even better! Measurements and knowing what you are doing can help you see in which direction and how far to go.
Jan Didden
Christer said:
Well then, I'm sorry too.
This net thing is too slow and the smile didn't appear on your post.
analog_sa said:
Man, I just gave a suggestion for Nuuk to try it.
I didn't say it's the only value.
I said it to try this on HIS circuit, as it seams that HIS circuit is sensitive to these small changes.
BTW I think that, as always, the cap must be NEAR the component (in this case the regulator).
If that's not practical, at least a small 0.1uf near the regulator.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.