Power off question

[catalin gramada]

Thanks Gents. That.s my PS how is wired now

[vintagekiki]

catalin gramada ...

Instead of switch S2 (which prevents an unpleasant sound sensation when amp start on) elegant solution is some circuit from the soft start gallery.

https://320volt.com/en/simple-soft-start-circuit-220v/
Simple soft start Circuit 220V

http://electronics-diy.com/soft-start-for-power-supply.php
Soft Start for Power Supply

https://www.schematicsforfree.com/archive/dir/Audio/Circuits/Power+Supplies+-+Audio%2C+Amplifiers.+etc/Softstart+circuits
Audio Amplifiers/Softstart circuits

[catalin gramada]

Instead of switch S2 (which prevents an unpleasant sound sensation when amp start on) elegant solution is some circuit from the soft start gallery.

https://320volt.com/en/simple-soft-start-circuit-220v/
Simple soft start Circuit 220V

http://electronics-diy.com/soft-start-for-power-supply.php
Soft Start for Power Supply

https://www.schematicsforfree.com/archive/dir/Audio/Circuits/Power+Supplies+-+Audio%2C+Amplifiers.+etc/Softstart+circuits
Audio Amplifiers/Softstart circuits

Thanks. The limiter resistor is essential to do this power supply usable. It is a 800VA PT and big bank of capacitors. Without it the startup inrush current will be huge (is 1.2 ohm primary winding). Limiting at 8-10A at startup make possible to use fast fusible protection close rated at max admissible circuit condition for more effective protection. Is not a standby sw but a two stroke power on. I still wonder if I can mechanical modify a DPDT toggle switch contact from a row in a way to ensure a row to do a faster contact before the other- sort of mechanical misalignment or some- in one stroke? Otherwise the delay have not to much sense to be greater than 1/50 sec I think...A relay time controled circuit is more elegant indeed and is some I'll consider.

[vintagekiki]

catalin gramada... modify a DPDT ...

Try to find Toggle switch with 3 position, DPDT, On-On-On for 10A min.

https://www.steinair.com/product/toggle-switch-3-position-dpdt/

[Helmholtz]

Increasing bias filter capacitance will create a new problem of slow ramp-up of bias voltage, won't it? So you will be missing bias at power up rather than power down.

Yes, slow ramp-up could be a problem - depending on cap value and circuit. That's why I put it as a question.

And does it not also create problems with LF response?

I don't see how, as the bias cap is not in the signal path. In fact its negative end can be considered a signal ground.

[g1]

I don't see how, as the bias cap is not in the signal path. In fact its negative end can be considered a signal ground.

I don't see how (bias filter capacitance affecting LF response) either, it's just something I though I recalled from previous discussions about practical limits of bias filter size. I'm probably not remembering correctly.

[Chuck H]

The bias filter having an influence on LF response is a subject originally argued here by "he who shall not be mentioned" (paraphrase: audio wizard male). It was many years ago and there were some higher minds involved in the discussion. As I recall there was some credibility to it. Though it was was still arguable as to how much it could actually matter. In the end it was still up in the air as to whether or not an astute player would be sensitive to the difference. I'll need to dig deep and hope to find the discussion if I am put to siting chapter and verse for why it's a possibility at all since I've relegated it to the "doesn't matter enough" file. But just because "I" put it there doesn't mean it belongs there. It just means I had bigger fish frying at the time.

[Helmholtz]

The bias filter having an influence on LF response is a subject originally argued here by "he who shall not be mentioned" (paraphrase: audio wizard male).

The underlying theory might have been like this:

The bias cap is in series with the grid reference resistors, so its impedance seemingly adds to the PI load.
If its impedance was high enough, it might cause some bass boost as cap impedance increases at LF.
(Cap impedance being high enough translates to cap value being low enough.)

But such a low value bias cap couldn't do its main job, i.e. filter mains frequency.


But the main fallacy of this theory is that the PI signals are opposite phase, making the junction of the grid ref. resistors a virtual signal ground*.
I.e. there there can't be signal voltage - even without a cap to circuit ground.
Meaning that there is actually no net PI load current flowing through the bias cap, so it can't have an influence on the signal.

* A signal ground is a circuit point where there can't be any signal voltage wrt to circuit ground.


BTW, is there a list of "not to be mentioned" persons?

[Enzo]

Not really, there was one character here for a while, and he spewed complete baloney, including sometimes even dangerous advice.

[vintagekiki]

"he who shall not be mentioned"

I am happy to feel
that I am not like others

[g1]

The bias filter having an influence on LF response is a subject originally argued here by "he who shall not be mentioned" (paraphrase: audio wizard male). It was many years ago and there were some higher minds involved in the discussion. As I recall there was some credibility to it.

This was my recollection as well. I just couldn't bring myself to mention the name and 'credibility' in the same statement.
I'm at least glad that my recall is still (mostly) ok.

[g1]

The bias filter having an influence on LF response is a subject originally argued here by "he who shall not be mentioned" (paraphrase: audio wizard male). It was many years ago and there were some higher minds involved in the discussion. As I recall there was some credibility to it. Though it was was still arguable as to how much it could actually matter.

The underlying theory might have been like this:

The bias cap is in series with the grid reference resistors, so its impedance seemingly adds to the PI load.
If its impedance was high enough, it might cause some bass boost as cap impedance increases at LF.
(Cap impedance being high enough translates to cap value being low enough.)

I dug up a couple of the related posts. Although there may be a point that there will be some change in impedance of the bias supply, and thus affecting AC ground there, SGM's idea about it seems backwards. He thinks stiffening the bias supply will act the same as stiffening the main supply, and increase bass response. That does not seem to make sense here. If anything, I think the opposite would be true, as Helmholtz outlined (lower capacitance would increase LF response).
Chuck, maybe these will trigger some recollection as you seem to have run some sims related to the issue.

https://music-electronics-forum.com/...173#post506173
https://music-electronics-forum.com/...272#post514272
https://music-electronics-forum.com/...278#post514278

[Chuck H]

Ok... I read over the links (thanks g1) and it's pretty much just as Helmholtz surmised on BOTH counts. Yes, it was assumed that larger cap values would bypass more LF and mitigate phase cancellation at the junction of the grid loads. AND that would seem to be incorrect since the signal voltage from either output of the PI is opposite phase and both share the bias circuit series resistance. My original sims were run as an individual bias circuit just to see if values typically used could be increased to make a significant difference in the bypassed frequency. And it seemed that could be the case. These tests didn't include the PI circuits they serve. New sims I just did included the PI circuit and though a significant difference in LF can be registered RIGHT AT the node for the grid loads and bias circuit series resistance when using different value caps, there was absolutely NO change whatever at the opposing ends of the grid load resistors. So... It's an absolute myth. Just more perceived reality based on fractional knowledge spewed as lore by a self important blowhard... Who shall not be mentioned.

[Helmholtz]

I think it's essential to understand that the PI load currents through the grid ref. reference resistors do not flow to ground but rather between the opposite phase PI outputs - this assumes a well balanced PI. Any differential current will flow to ground, though.

BTW, bias ripple (50/60Hz) is a common mode signal for the power tubes. That's why the bias cap can be relatively small with a well balanced output stage.