"Bypass caps" and their effect on perceived sound

[cnpope]

Originally Posted by 20to20

1-2 ohms is when the filament is cold. Depending on the tube type, it rises to anywhere from 20 -80 ohms when the filament is hot.

Can you present your calculation of 20-80 Ohms? The figures given by andyjevans support the answer of about 3 ohms when hot.

Chris

[TheGimp]

Using the datasheet values of 2.1V at 625mA and 4.2V at 325mA I get:

2.1V - 3.36R
4.2V - 12.923R

[20to20]

Originally Posted by cnpope

Can you present your calculation of 20-80 Ohms? The figures given by andyjevans support the answer of about 3 ohms when hot.

Chris

Sure, before he specified his 2v DHT's I assumed we were discussing typical 6 or 12v tubes with 150ma to 300ma heaters. 12.6/.15 = 84 or 6.3/.3 = 21 These are the tubes customerily used when doing the "filament" biasing trick.

[yagoolar]

I have found the document, in Polish, in which some testing of various electrolytic caps was done. See page 13 - old cap (80s I guess) was tested without and with a bypassing caps. Oscilloscope traces are self explanatory.
http://www.mrelektronik.pl/032.pdf

[Conrad Hoffman]

Originally Posted by disco

Above 10KHz electrolytics behave differently, from brand to brand, from type to type. Wouldn't it be a explanation for differences in decoupling a ps?

Sure, but if that difference is affecting anything, it's better to deal with it at the source or destination, rather than fight the inductance of the wireing. There's no benefit to doing it at the main cap, unless there's some physical constraint like lack of space anywhere else or the lure of lugs in plain sight to attach to, IMO not the best reason to do it that way.

[Nounours18200]

... IMO again, if bypassing electrolytics has an effect, the correct design solution is really to bypass someplace else, either at the rectifiers to kill RF, or at the circuit where you're not fighting with the inductance of the wiring from the main caps. The correct return point for the bypass is also important, and it may not be where you think!

I understand that the bypass should positioned at the rectifiers levels, but I do not see what you mean by "The correct return point for the bypass is also important, and it may not be where you think" ??

Thanks for your extra explanation (and sorry for my English),

[Wavebourn]

Originally Posted by Conrad Hoffman

Sure, but if that difference is affecting anything, it's better to deal with it at the source or destination, rather than fight the inductance of the wireing. There's no benefit to doing it at the main cap, unless there's some physical constraint like lack of space anywhere else or the lure of lugs in plain sight to attach to, IMO not the best reason to do it that way.

It is not just an inductance of wireing. It is combination of spread inductance, capacitance, and non-linear semiconductor layer in aluminium-oxide semiconductor electrolytic capacitors. When (especially class AB) amp draws power from PS with such capacitors phase intermodulation happens. I desctibed what I found, back in the parallel thread about bypassing capacitors. Bypassing did not change neither THD, nor IMD, nor frequency response. But stereo image improved, also cymbals sounded much more natural. I suspected phase intermodulation as the only thing that can affect the sound such a way and indeed found strips of sidebands of slightly different heights.

[Wavebourn]

Originally Posted by Nounours18200

I understand that the bypass should positioned at the rectifiers levels, but I do not see what you mean by "The correct return point for the bypass is also important, and it may not be where you think" ??

Think about AC current that flows through the cap and so called "Common Wire". Common Wire is what some people call "Ground", but actually it is not a zero-ohm ground. It is a wire that has resistance and inductance. It is called "Common" because it is common for many currents that flow through it.

Take for example short 10 A peaks of current through diodes that charge a PS filter cap. If say resistance of the ground wire that passes this current is 0.001 Ohm that means already 0.01V peaks on this wire, from minus of a bridge, or a center tap of the transformer, to the filter cap. In relation to 1 V sensitivity of the power amp it is -40 dB of loud buzz. Solder there negative leg of the cap that filters B+ for preamp, and you have this loud buzz as if coming from nowhere.

People often claim that vacuum rectifiers "sound better". Indeed they would cause more forgiving to the wrong layout result because of higher internal resistance and lower level of charging peaks.

[disco]

And for the perceived "sound" of vacuum rectifiers internal resistance and damping of the tank is of importance.

[Wavebourn]

Originally Posted by disco

And for the perceived "sound" of vacuum rectifiers internal resistance and damping of the tank is of importance.

Still, carefully designed layout is more important in any case. With proper layout advantages of vacuum rectifiers play no role anymore, while lesser sag under the load leads to preference of SS rectifiers.