How much ripple on B+ is allowed?

[Rundmaus]

Hi,

at the moment I'm playing around with Duncanamps PSUD in order to try out some ideas for a tube based PS for the PP1C...

One main question to me is: How much ripple on the B+ rail is acceptable without hum problems? Searched the forum but found not much about it...

What about the negative rail for the tube CCS? Is the CCS more sensitive against ripple than the amp? How much ripple can one allow here?

Greetings
Andreas

[SY]

A good CCS gives outstanding power supply rejection, as long as its reference is quiet. Push-pull circuitry also has excellent power supply rejection.

To determine what is acceptable, start by defining what your desired noise is for the amplifier, then start working backward to see how much ripple for each stage is allowable to achieve that goal.

[Rundmaus]

...thanks for your fast reply , but I'm afraid that I'm unable to work my way back as you suggest... I'm still a total tube newbie...

Tried to simulate the original PP1C PS in PSUD and came to about 10mV ripple on the B+ rail...

About the CCS, the "standard" layouts for tube based CCS I found in the forum generate the reference from their supply voltage, does this mean I have to clean the negative rail up completely, or is it possible to filter the reference separately, lets say by a sufficiently big cap?

Greetings,
Andreas

[kmtang]

Are you building a PSU for the line or phono amp?

If so, I would suggest you could build a power supply with simple C-L-C-L-C double choke filter which would provide you with almost pure B+ power supply.


Johnny

[SY]

Using a low impedance reference plus bypassing it will take care of noise.

[Rundmaus]

It ist going to be a EL34 PP amplifier, schematics see below...

I think for the B+ rail, I'll try to reach the values reached with the original supply in PSUD... Concerning the CCS i hope for more answers...

kmtang, you suggest a C-L-C-L-C filter, i thought about a L-C-L-C filter chain, as I read somewhere that tube rectification and big input caps in the filter chain aren't so good... What do you think?

Greetings
Andreas

[kmtang]

Hi Andreas,

For EL34 PP amp, it just need C-L-C filter for power supply. The push-pull configuration will cancel out the hum ripple from B+.

I won't worry that much for the B+ ripple at all. I would rather use smaller H choke (say 5H) for lower internal resistance and high quality (low ESR) bypass caps.


Johnny

[SY]

Oh, man, that's an easy one. CC diff amp, p-p output stage... unless you're driving something ridiculously sensitive, a few volts of ripple won't cause you a noise problem. This opens the opportunity to keep the ripple current low by not having to use a very large capacitance in the power supply.

Since this is running AB, make sure you do your simulations at both zero signal and full power current draw.

[Rundmaus]

Hm,

good point concerning the simulations... what load resistance or load current do I use for both situations?

Apart from that, i wonder why the original supply (see below) uses enough filtering to bring the ripple down to a few mV? Is it because of the HF noise of the SS diodes in the rectifier? Or just oversized filtering to provide an "as good as possible" B+?

Another thing, how much voltage do I need on the negative rail for the tube CCS? I have some mains transformers lying around for the project, and there would be a 50-0-50 winding left to be used for the negative rail. But that means that after tube rectification / filtering I have about -70V available? Is this enough for a tube CCS?

Greetings
Andreas

[kmtang]

It is better stay with the original design with use of 12Vac for the CCS -ve volt. If you use higher voltage, you would have to redesign the circuit for it to work properly.

Johnny