They will run quieter with a regulated supply, and possibly
there are other sonic benefits, possibly not.
there are other sonic benefits, possibly not.
Regulated power supplies for something light, like a preamp, are easy. Regulated power supplies for a power amp are a booger. Most people cheap out and build something that runs fine at idle, but cannot pass sufficient current under load. They figure that caps at the output of the regulator will take the load and everything will be fine. The amp sounds like an amplified car wreck and those listening to it say,"Well, gee, regulated power supplies must not be that good after all." And, that, of course, becomes the received wisdom.
A couple of points:
--The regulator itself must be able to supply the entire current demand of the circuit, and plus a little more. For a class A amp, this can be considerable. For a class AB amp, it can be less demanding--until the music gets loud--which brings us to...
--The regulator must behave well when faced with transients
--The rest of the power supply must be built as heavily as though it were going to support the circuit without the regulator. In other words, don't try to use the regulator to take out the sawtooth remaining after a poor job of filtration coming off the bridge. For example, let's say that some fool put an amp together with, say, a nominal 45V off the caps after the bridge. There's not enough filtration, so you've got something like a 2Vp-p sawtooth going into the regulator. The regulator puts out a nice, steady 42V. (Let's pretend for the moment that there's no overhead for the regulator.) Sounds good, right? 45V less the 2V of the sawtooth leaves 43V and all you want is 42V. Everything is peachy. Until you play music. Then the power supply gets drawn down, the 45V becomes 43V, the 2V actually increases, and the regulator starts passing the sawtooth onto the rail. Ugh.
Yes, that's oversimplified (the sawtooth would actually be centered on the 45V, for instance, not below it), but the general idea remains the same. You simply must have a robust power supply behind the regulator. Those who are against regulators would point out that this is true anyway. Well, duh! There's no substitute for heavy iron. Well, then what does a regulator get you that a really big power supply doesn't? Rock steady rail voltages. Isolation from line nasties. I'm always amused to see people do calculations to three or four decimal points for the operating points for their devices, then use an unregulated power supply. Ahem. AC line variations are going to swallow all those careful calculations whole. The guy designed for 32V, but what he actually got was 30.5V. Why? Because it's July and everybody's air conditioning is running, or because it's January and everybody's heat is running. Or because it's May and everybody's class A stereo is running. The AC at my house runs about 122 to 124V. That's pretty good. Roughly +-1%. But even taking one of my circuits over to a buddy's house who lives 6 miles away drops the line voltage down to 119V or so. A lot of people don't realize that each neighborhood transformer is set by the power company. Ignore air conditioning for the moment and contemplate the idea that a hot, sweaty lineman may not feel like tuning your street's AC that last little bit. Things vary from place to place even without having wild fluctuations in the AC due to load variations. A well designed regulator can take that variable out of the equation. You no longer have to guess what your rail voltage will be.
--Although other people feel differently, I'm a big proponent of having a heavy capacitor bank after the regulator. Not only will it help with transients, but it will quiet things even further.
--You might as well get used to it--if you intend to regulate the rails for a power amp, you're going to end up with a power supply the size of an entire amp. That's not just a roundabout way of saying that the power supply represents the majority of the bulk of an amplifier. You're going to need more heatsinks, more circuitry, and more caps. Oh, and don't forget more money. It might help to think of a regulator as an amplifier in its own right...an amplifier of DC. It literally amplifies the DC reference voltage up to the rail voltage.
There was something else I was going to throw in, but it slipped out of my head while I was typing and I've got to go earn my keep. If I remember, I'll stick it in later.
Grey
A couple of points:
--The regulator itself must be able to supply the entire current demand of the circuit, and plus a little more. For a class A amp, this can be considerable. For a class AB amp, it can be less demanding--until the music gets loud--which brings us to...
--The regulator must behave well when faced with transients
--The rest of the power supply must be built as heavily as though it were going to support the circuit without the regulator. In other words, don't try to use the regulator to take out the sawtooth remaining after a poor job of filtration coming off the bridge. For example, let's say that some fool put an amp together with, say, a nominal 45V off the caps after the bridge. There's not enough filtration, so you've got something like a 2Vp-p sawtooth going into the regulator. The regulator puts out a nice, steady 42V. (Let's pretend for the moment that there's no overhead for the regulator.) Sounds good, right? 45V less the 2V of the sawtooth leaves 43V and all you want is 42V. Everything is peachy. Until you play music. Then the power supply gets drawn down, the 45V becomes 43V, the 2V actually increases, and the regulator starts passing the sawtooth onto the rail. Ugh.
Yes, that's oversimplified (the sawtooth would actually be centered on the 45V, for instance, not below it), but the general idea remains the same. You simply must have a robust power supply behind the regulator. Those who are against regulators would point out that this is true anyway. Well, duh! There's no substitute for heavy iron. Well, then what does a regulator get you that a really big power supply doesn't? Rock steady rail voltages. Isolation from line nasties. I'm always amused to see people do calculations to three or four decimal points for the operating points for their devices, then use an unregulated power supply. Ahem. AC line variations are going to swallow all those careful calculations whole. The guy designed for 32V, but what he actually got was 30.5V. Why? Because it's July and everybody's air conditioning is running, or because it's January and everybody's heat is running. Or because it's May and everybody's class A stereo is running. The AC at my house runs about 122 to 124V. That's pretty good. Roughly +-1%. But even taking one of my circuits over to a buddy's house who lives 6 miles away drops the line voltage down to 119V or so. A lot of people don't realize that each neighborhood transformer is set by the power company. Ignore air conditioning for the moment and contemplate the idea that a hot, sweaty lineman may not feel like tuning your street's AC that last little bit. Things vary from place to place even without having wild fluctuations in the AC due to load variations. A well designed regulator can take that variable out of the equation. You no longer have to guess what your rail voltage will be.
--Although other people feel differently, I'm a big proponent of having a heavy capacitor bank after the regulator. Not only will it help with transients, but it will quiet things even further.
--You might as well get used to it--if you intend to regulate the rails for a power amp, you're going to end up with a power supply the size of an entire amp. That's not just a roundabout way of saying that the power supply represents the majority of the bulk of an amplifier. You're going to need more heatsinks, more circuitry, and more caps. Oh, and don't forget more money. It might help to think of a regulator as an amplifier in its own right...an amplifier of DC. It literally amplifies the DC reference voltage up to the rail voltage.
There was something else I was going to throw in, but it slipped out of my head while I was typing and I've got to go earn my keep. If I remember, I'll stick it in later.
Grey
Member
Joined 2002
I was reading some where in a mag / online that it is useless to run a regulated psu for a class a amp. It was talking about it being not good because the amp is always drawing current and thus there is no need for a regulated psu. ? Nelson can you comment at all ?
Member
Joined 2002
Silly me, i read the suggestion as regulating the voltage for the small devices. That could lower the rails for the output the 5 additonal volts NP favors and save some cash.
A number of years ago i copied a fully regulated class A PP design of the French Audiophile club, with a separate powersupply at 5 feet distance, posted a picture of the contraption on some thread. (industrial 9-pole connectors with gold contacts, good for +100 amps)
The thought of an Aleph 5 with full regulation is as mindblowing as reading Casey's words on the Monster thread after the injections they gave him in his new biotech job.
A number of years ago i copied a fully regulated class A PP design of the French Audiophile club, with a separate powersupply at 5 feet distance, posted a picture of the contraption on some thread. (industrial 9-pole connectors with gold contacts, good for +100 amps)
The thought of an Aleph 5 with full regulation is as mindblowing as reading Casey's words on the Monster thread after the injections they gave him in his new biotech job.
jleaman,
Some topologies, such as the original Zen and the Son of Zen, have essentially DC current draw (albeit for different reasons). That doesn't mean that they wouldn't benefit from a steady rail voltage. Noise on the rail could be a nuisance, particularly on the SOZ, which doesn't have all that great a CMRR, even though it's a differential.
Other topologies vary quite a bit in their current draw. Class A means simply that the signal is present in the active devices at all times, not that the circuit's current draw sums to DC.
The Mini-A will vary its current draw, moment-to-moment. The Aleph-X sums to DC.
That said, Nelson tends to design circuits with a certain amount of 'give' to them. I had one pair of (slightly modified) Aleph 2s on a big variac for a while, just so I could play with the rail voltage. None of the Alephs are going to come unglued if the rail increases or decreases by 10 or even 15%, although the measured performance would suffer a bit.
You can get rid of a great deal of noise with passive components. A CRC or CLC filter bank will do a good job. But you can't lock the voltage that way.
Grey
Some topologies, such as the original Zen and the Son of Zen, have essentially DC current draw (albeit for different reasons). That doesn't mean that they wouldn't benefit from a steady rail voltage. Noise on the rail could be a nuisance, particularly on the SOZ, which doesn't have all that great a CMRR, even though it's a differential.
Other topologies vary quite a bit in their current draw. Class A means simply that the signal is present in the active devices at all times, not that the circuit's current draw sums to DC.
The Mini-A will vary its current draw, moment-to-moment. The Aleph-X sums to DC.
That said, Nelson tends to design circuits with a certain amount of 'give' to them. I had one pair of (slightly modified) Aleph 2s on a big variac for a while, just so I could play with the rail voltage. None of the Alephs are going to come unglued if the rail increases or decreases by 10 or even 15%, although the measured performance would suffer a bit.
You can get rid of a great deal of noise with passive components. A CRC or CLC filter bank will do a good job. But you can't lock the voltage that way.
Grey
GRollins said:
That's a good substitute for a full blown reg.
I have tested on Aleph4 CRC, CLC and even LC which is my current setting with excellent results, I could add that for my clone CRC was not good enough by a big marging
With LC regulation is outstanding and really cleans the amp by a substantial amount. Lytics are 160mf per bank bypassed with MKP film caps and inductance is arround 15mH on gapped iron cores.
Member
Joined 2002
The only problem with clc is that it takes up so much room 🙁.. I am going to see about doing a group buy on the zen regulator boards> i have to just email nelson and as for permission to make the boards they are designed. I wonder if it will improve any thing or not but it's worth a try.
digi01 said:
Does it . . . ?
Anyhow, I include the regulated PSU into my "nice-to-have" shopping list.
Regards
- Status
- Not open for further replies.