I've always preferred regulation with heavy caps on the output side. The regulator locks the voltage, and the caps take on the task of supplying the current pulses. The amount of capacitance downstream from the regulator will then determine the speed required of the regulator (though I confess to leaning towards the "fast" end of the spectrum, myself), since it will act as a filter for anything coming back up the line towards the regulator. Working backwards from that, people who might want, for instance, 1000uF, are going to need a very fast regulator, indeed. People like me, who dream of silly things like 10 Farads, can go for much slower circuits.
Why go for heavy capacitance? It's been my experience that people who "don't like the sound of regulators" are almost invariably from one of two camps:
--They dislike regulation on a philosophical basis or,
--They're speaking from experience with circuits that have small amounts of capacitance.
Can't do anything about the first group, they're entitled to their opinions. But the second group can be swayed...perhaps. Given that caps (choose your favorite brand) are less likely to have a 'sound' than a circuit, I'd rather listen to caps than transistors or tubes. (N.B. Some peoples' preference for tube regulators over solid state may be a reflection of this same concept.)
Circlotron's use of an amp to do the job strikes me as entirely appropriate. A regulator is nothing but a DC amplifier, anyway. I just wish I had a few unused amps sitting around to assign to the task.
Grey
Why go for heavy capacitance? It's been my experience that people who "don't like the sound of regulators" are almost invariably from one of two camps:
--They dislike regulation on a philosophical basis or,
--They're speaking from experience with circuits that have small amounts of capacitance.
Can't do anything about the first group, they're entitled to their opinions. But the second group can be swayed...perhaps. Given that caps (choose your favorite brand) are less likely to have a 'sound' than a circuit, I'd rather listen to caps than transistors or tubes. (N.B. Some peoples' preference for tube regulators over solid state may be a reflection of this same concept.)
Circlotron's use of an amp to do the job strikes me as entirely appropriate. A regulator is nothing but a DC amplifier, anyway. I just wish I had a few unused amps sitting around to assign to the task.
Grey
If you have 1000uF filter caps on a fairly modest rail voltage of +/- 20 volts, and along comes a 1 amp current drain.... Guess what 10v ripple is going to do to a 20 volt rail? Eeeyeew. 
You will only be able to swing about +/- 8 volts before you hit clipping.
You will only be able to swing about +/- 8 volts before you hit clipping.
When more is less and less is more, more or less.
How one reconciles those two assertions with the known laws of physics I'm not quite sure.
Me? Big caps and ordinary power cable.
The reason for mentioning this is that we have one group of people saying that 30-50% PSU ripple is ok, the chip will flatten it all out, and on the other hand the power cord crowd say that what the object of their passion does may be so small as to be unmeasurable but it makes a world of difference to the sound.Circlotron said:*** This would be an excellent question to ask the Magic Power Cord worshipers. Ding! -> Checkmate!
How one reconciles those two assertions with the known laws of physics I'm not quite sure.Me? Big caps and ordinary power cable.
Circlotron said:If you have 1000uF filter caps on a fairly modest rail voltage of +/- 20 volts, and along comes a 1 amp current drain.... Guess what 10v ripple is going to do to a 20 volt rail? Eeeyeew.
You're right, but consider some things:
-The 10V ripple will only be present if you draw 1A DC, or if the frequency is under 50Hz (and not 100Hz, because the second half period will draw current from the other rail); this is far from being impossible, I agree
-With a 10V rillpe, and a 1A current on a 8R loudspeaker, you will need a 8V output. Not far from clipping, but still ok.
And this was done with one aproximation: that you draw a DC current. A 50Hz AC current won't cerate a 10V ripple because the 1A will be the peak value.
But yes, in this case, the chip is pushed close to it's limit. But this limit (in this case, the PSSR) is so big...
Frank,
You mean you've got a regulator circuit that can't do DC?
Yet regulates a DC rail?
Cool.
Tube amps, bless 'em, need not apply for this application. Being transformer coupled will, by definition, keep them from passing DC. If they can't pass DC, then there's no rail, and if there's no rail...
Ditto for transistor amps with cap-coupled outputs, such as the original ZEN. That's not to say that you couldn't remove the cap and get it to work as long as all other parameters fell into place.
A regulator is nothing but an amp that amplifies DC. Given a reference voltage, it will do its dead level best to reproduce that DC reference. Incorporation of a feedback loop can trick the amp into reproducing a higher or lower voltage, with the added benefit of correcting perturbations in the rail. But when all is said and done, it's just an amplifier amplifying a DC signal. What could be simpler?
I've seen people claim that a regulator amplifies ground, but I don't agree with that viewpoint. Granted, a dirty ground can influence the output of the regulator, but that's true for any amp. I doubt those same people would claim that an audio amp is amplifying ground; else how do they intend to get music out of it? It's the DC reference that gets amplified.
Grey
You mean you've got a regulator circuit that can't do DC?
Yet regulates a DC rail?
Cool.
Tube amps, bless 'em, need not apply for this application. Being transformer coupled will, by definition, keep them from passing DC. If they can't pass DC, then there's no rail, and if there's no rail...
Ditto for transistor amps with cap-coupled outputs, such as the original ZEN. That's not to say that you couldn't remove the cap and get it to work as long as all other parameters fell into place.
A regulator is nothing but an amp that amplifies DC. Given a reference voltage, it will do its dead level best to reproduce that DC reference. Incorporation of a feedback loop can trick the amp into reproducing a higher or lower voltage, with the added benefit of correcting perturbations in the rail. But when all is said and done, it's just an amplifier amplifying a DC signal. What could be simpler?
I've seen people claim that a regulator amplifies ground, but I don't agree with that viewpoint. Granted, a dirty ground can influence the output of the regulator, but that's true for any amp. I doubt those same people would claim that an audio amp is amplifying ground; else how do they intend to get music out of it? It's the DC reference that gets amplified.
Grey
More or Less ?......................
I strongly question those saying that they prefer the sound of GC amplifiers with minimal supply capacitance - ie 1000 uF.
I have not read any proper reviews (opinions) of the differences except for comments of more tuneful (?) bass with less capacitance.
Anybody here have experience with GC amplifiers and different amounts of supply capacitance ?.
Eric / - Big caps and magic power cable for me.
Circlotron said:
The reason for mentioning this is that we have one group of people saying that 30-50% PSU ripple is ok, the chip will flatten it all out, and on the other hand the power cord crowd say that what the object of their passion does may be so small as to be unmeasurable but it makes a world of difference to the sound.
Me? Big caps and ordinary power cable.
I strongly question those saying that they prefer the sound of GC amplifiers with minimal supply capacitance - ie 1000 uF.
I have not read any proper reviews (opinions) of the differences except for comments of more tuneful (?) bass with less capacitance.
Anybody here have experience with GC amplifiers and different amounts of supply capacitance ?.
Eric / - Big caps and magic power cable for me.
GRollins said:
For Jung regs, 317s, 780xs... yes
but can you consider a basid reg made of a zener and a pass transistor as a DC amplifier?
and how about a shunt reg?
Hi,
What does "can do DC" mean?
Sure the regs handle DC, what I took exception with is the term "AMPLIFY DC" which is a rather ill chosen term.
As I know you to be familiar with tube regs, especially series regs, I find this a rather odd way of phrasing this matter as the input DC voltage isn't amplified at all, i.e. it does not surface as being an order of magnitude higher in value.
The way you put it lets to believe we now have a higher, amplified, voltage on the output of the regulator compared to what we put in.
That would be rather impossible to accomplish IMHO.
Cheers,😉
What does "can do DC" mean?
Sure the regs handle DC, what I took exception with is the term "AMPLIFY DC" which is a rather ill chosen term.
As I know you to be familiar with tube regs, especially series regs, I find this a rather odd way of phrasing this matter as the input DC voltage isn't amplified at all, i.e. it does not surface as being an order of magnitude higher in value.
The way you put it lets to believe we now have a higher, amplified, voltage on the output of the regulator compared to what we put in.
That would be rather impossible to accomplish IMHO.
Cheers,😉
Frank,
'Can do DC' meaning able to pass DC. There's no real requirement for really, really good high end response as long as there are large-ish caps at the output of the regulator.
That's not to say that everyone uses caps at the output. Other folks feel differently than I do about such things.
True, you can't pull more voltage out of the regulator than is available in the raw DC, but that's not the "signal" being amplified any more than the rail voltage in a "normal" amplifier is what's being amplified. The signal is the reference voltage. And yes, I've seen schematics where the DC reference is actually amplified, although I wouldn't care to go that route, myself--you'd also amplify any noise from the reference (though the reference voltage was invariably well filtered). Yuk. But there's no need to assume voltage amplification, since most regulators provide current as output.
Bricolo,
Hmmm...good point. Do shunt regulators fit the DC amplifier model? Lemme ponder on that. Shooting from the hip, I'd say yes. Treat the regulator as a black box and watch the voltage and current. An ideal regulator would hold the voltage fixed and let the current vary according to the needs of the moment. Whether the black box contains a series or shunt regulator shouldn't be an issue. I'll try to get a spare moment and think that one through.
Is a pass transistor and a Zener an amp? Sure. The input signal is the--let's say 9V--from the Zener. But there's zip for current, right? However, the output for the regulator can potentially be huge, current-wise. Did we amplify the voltage? No. But we sure amplified the current.
Grey
'Can do DC' meaning able to pass DC. There's no real requirement for really, really good high end response as long as there are large-ish caps at the output of the regulator.
That's not to say that everyone uses caps at the output. Other folks feel differently than I do about such things.
True, you can't pull more voltage out of the regulator than is available in the raw DC, but that's not the "signal" being amplified any more than the rail voltage in a "normal" amplifier is what's being amplified. The signal is the reference voltage. And yes, I've seen schematics where the DC reference is actually amplified, although I wouldn't care to go that route, myself--you'd also amplify any noise from the reference (though the reference voltage was invariably well filtered). Yuk. But there's no need to assume voltage amplification, since most regulators provide current as output.
Bricolo,
Hmmm...good point. Do shunt regulators fit the DC amplifier model? Lemme ponder on that. Shooting from the hip, I'd say yes. Treat the regulator as a black box and watch the voltage and current. An ideal regulator would hold the voltage fixed and let the current vary according to the needs of the moment. Whether the black box contains a series or shunt regulator shouldn't be an issue. I'll try to get a spare moment and think that one through.
Is a pass transistor and a Zener an amp? Sure. The input signal is the--let's say 9V--from the Zener. But there's zip for current, right? However, the output for the regulator can potentially be huge, current-wise. Did we amplify the voltage? No. But we sure amplified the current.
Grey
The AVA Mosfet 120B has a series pass regulator that is just a FET with a zener on the gate, no feedback.
The amp sounded fine.
The amp sounded fine.
We lowered the output impedance by the use of negative feedback. Sounds like an amplifier to me.GRollins said:
Or maybe it's a torque wrench? 😉
- Status
- Not open for further replies.