• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

PP PSU: How much ripple is too much ripple?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
While PSUD II enables us to prototype PSU's to get ludicrous low levels of ripple, how far should we really take things?

I'm building a PP EL84 amp (triode connected) and trying not to compromise it but still be realistic.

Power tranny secondaries are 240v and I'm planning on using SS bridges, not tube rectifiers.

I have half a dozen 470uf caps available, 2 with 450v rating and 4 with 350v rating.

There's no space in the chassis for a choke, so can anyone advise on a decently performing pi or CRCRC filter to suit my needs?

Output tube draw will be class A at about 140mA (35mA per EL84)

Input and phase splitter are 6SN7 configured as gain plus concertina.

All advice gratefully accepted.

(I'd prefer not to have to go shopping for new PSU caps if I could help it)
 
i think something like 10% was acceptable on PP and >5% on SE as PP has a certain amount of a cencelling effect. seeing as its triode mode, your going to have ripple on G2 aswell, as the plate which might want less ripple. you should be fine with around 24v ripple. as for those filter caps, they sound way too big for a tube amp. ive never used bigger than 220uF for a reservoir. then with appropriate dropping resistors in the PSU, you should be fine.
 
. As an example, I just prototyped with 6L6 (6P3S-E) stereo power output stage (UL).
The power supply is a full wave rectifier with some 450 Vdc output voltage.
There are two 390 µF/ 350 V capacitors connected in series after the diode bridge forming a 195 µF total capacitance. This voltge is then directly supplied the the center taps of the OPT's. The total standing current from the power supply is some 250 mA.

Measured hum level was -72 dB (ref = 30 W).

When an extra (single) 180 µF was added in parallel with existing capacitors, the hum level dropped to -80 dB.

These hum levels were created at the output stage completely. The filtering of the phase splitters and voltage amplifying stages were sufficient and did not have any effect to the results.
 
An emitter follower transistor on a heatsink and insulated from the metal of the heatsink in this case one with high voltage capability will reduce ripple values, and be far more effective than simple passive components. At 450vDc your choices of transistor are quite a few as these parts are common in TVs- Suggest bias to begin at 6 ma ie 2x 33k resistors and.arrange the mid point of the two resistors to have a .33uf cap to ground then another .33 uf from the base to ground.

hope this helps cheers / chris
 
Last edited:
I am using a single 250uF cap to power a stereo pair of 5-20 (EL34 ultralinear). The output balance reduces hum, and negative feedback reduces it further.

With the benefit of hindsight I might have aimed at lower ripple. This is not because of hum, but hum intermodulation in the output stage. Hum cancels. Hum IM does not. Fortunately NFB reduces IM too. Back of envelope calculations suggest that hum IM will be sufficiently low, but without much margin for error.
 
Hmm, this is rapidly getting complex!

Basic gist from responses seems to be that with the caps I'm using a PI filter ought to be more than meeting the needs if I don't want to go to more complex arrangements such as regulators.

I'll need a series resistor to drop the 340v peak down to a stable 310 or so anyhow.

I'll start with 470u 150R 470u and see how we go.
 
artosalo, out of interest did you have matched output tube idle currents, or better still were you able to check output hum levels whilst you varied idle current mismatch?

There is bias adjustment for each output tube and idle currents were set identical ( about 50 mA).
I did not test the effect of current mismatch. The global negative feedback was 16 dB.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.