Regulated power supplies and capacitance

[Greg Erskine]

I've been following with great interest the trend towards using regulated PSU on our GCs. I have been trying to understand a technical reason why a regulated PSU is generally regarded as an improvement over a basic PSU. I can understand why a regulated PSU is an advantage for class A amps where a constant power and low ripple is advantageous, but the NS chipamps are class AB and have very high PSRR.

From reading many threads, the consensus is the NS chipamps don't like too much capacitance. Adding a regulated supply seems to allow the adding of significant amounts on capacitance without the drawbacks. With my limited knowledge, are the regulators really only acting as a buffer, hiding the caps from the chips? I know low power buffers are used on the signal input but are there high power buffers suitable for the power rails.

Also has anyone tried a capacitor multiplier PSU on a GC? I was wondering if a capacitor multiplier has the right qualities to be an ideal PSU for a GC.

[jackinnj]

"generally regarded" -- not by me -- I've yet to see one iota of factual information to support this view in the case of the Overture series. this isn't to say that I would alter my view, but it hasn't been demonstrated. regulation will help an amplifier with poor PSRR, but this isn't the case with the Overture series.

all you do with regulation is burn watts that would otherwise produce SPL's.

[Greg Erskine]

hi jackinnj,

I agree with what you say, but there are quite a few accomplished GC builders here that prefer a regulated PSU. The traditional reasons for this don't apply so I'm wondering is there another reason.

I haven't tried a regulated PSU myself because I'm a little hesitant to break away from original GC philosophy, short signal and feedback paths and minimal component count, but its getting to the point where I must try it for myself.

[carlosfm]

Let me put it this way: would you guys use a good signal op-amp (like an OPA627 or whatever) with an unregulated, low capacitance PSU?

This is even more serious, it's a power amp.

Regarding regulated PSU with the Overture series chips from National, I could demonstrate the difference with only one CD track, you would be addicted.
But you are far away, nothing much I can say.
The proof of the puding is listening, at least for me.

If you have big, very sensitive, easy to drive, horn speakers (preferably full-range or with simple crossovers), unregulated low capacitance will do.
Unregulated high capacitance sounds bad with these chips.
For real-world use, regulate.

[sam9]

I built a LM3886 amp quite some time ago. I could detect (instrument or listening) no advantage one way or the other except that with the same transformer, the regulated version would clip a little sooner. You could see this on a 'scope but in actual listening I never became aware of it.

The amswer to your question may differ according to the choice of amp-IC. If the app notes say anything about it, I would go with them a the most likely alternative.

I think there is a "coolness" factor with regard to regulated supplies. Nothing wrong with that, of course.

The logical course for you to follow would be to try both kinds of power supplies and decide for yourself. In fact I can't think of any reason why out can't build a single PS and just have regulated and unregulated terminal blocks so you could change over from one to the other with minimal effort.

[carlosfm]

Originally posted by sam9
I built a LM3886 amp quite some time ago. I could detect (instrument or listening) no advantage one way or the other except that with the same transformer, the regulated version would clip a little sooner.

With the same trafo you were regulating to a lower voltage.

Originally posted by sam9
In fact I can't think of any reason why out can't build a single PS and just have regulated and unregulated terminal blocks so you could change over from one to the other with minimal effort.

That would be nice if the chip sounded decent with 10,000uf caps.

[sam9]

With the same trafo you were regulating to a lower voltage.

Yes, this is a slight "trade-off" with a regulated supply. The otherway to look at it would be to note that for the same rail voltages, a regulated supply needs a slightly higher voltage transformer - which may or may not be readily available.

[sam9]

quote:
Originally posted by sam9
"In fact I can't think of any reason why out can't build a single PS and just have regulated and unregulated terminal blocks so you could change over from one to the other with minimal effort."


That would be nice if the chip sounded decent with 10,000uf caps.

Since I don't think there is any circumstance where more capacitance will sound worse, just use the capacitance you prefer for regulated or unregulated (which ever is greater).

FWIW, my LM3886 amp use 2x4700uF per rail and I have detected no short comming. Keep in mind it is my practice to select amps (purchased or constructed) such that I'm unlikely to ever overtax them.

[Greg Erskine]

Hi sam9

my LM3886 amp use 2x4700uF per rail and I have detected no short comming

Have you tried a direct comparision using only 1000uF or 2000uF per rail?

Since I don't think there is any circumstance where more capacitance will sound worse

I think this is where a lot of Gaincloners will disagree with you.

I did do a test 18 months ago bwtween 1000uF and 10000uF and couldn't really detect a significant difference one way or the other. So I haven't used large caps on my GCs since. I must revisit this again, now I know a little bit more about what I am doing.

No one have commented about using a capacitance multiplier. From what I can work out it gives you virtual capacitance which is what I think these chipamps like. Again, I must try it.

[Nuuk]

In fact I can't think of any reason why out can't build a single PS and just have regulated and unregulated terminal blocks so you could change over from one to the other with minimal effort.

Not quite so simple as the regulated GC apparently prefers much smaller caps on its pins than the unregulated GC.