Strange PX25 Circuit Diagram

[Paul Barker]

Very good of you to link that. It helped a little bit.

[Paul Barker]

OK after reading and absorbing some more rdh4 on the matter I remain with the view that the grid resistor is bypassed to the cathode of the output valve and the other capacitor is not bypassing anything to sufficient low frequerncy it serves a strange purpose. The power supply has to have it's own impedance reduction mechanism (usually the last cap), this small cap on the cathode resistor is not there to serve that purpose. I don't yet know what is it's purpose. RDH4 does not show it. Actully except int he OSRAM PX25 diagram posted on page one of this thread I haven't yet seen it elsewhere. It isn't a misprint because it is given a value.

What I did discover of great value from RDH4 about their thinking in those days, which we have perhaps lost from common working knowledge, are two interesting concepts.

1/ With RC decoupled stages, you can set the value of C so that it boosts the bass at a decided upon frequency to compensate for reduced bass elsewhere (like output transformer) and the phase effect is no different. So you have a phase stable bass boos tool to work with. The reaosn it works is that at the chosen frequency the "decoupling resistor" becomes adds to the load resistor, and therefore raises the voltage gain below the chosen frequency. Excellent tool!

Ref Bottom para of page 536 RDH4

2/ There is a very natty way of avoiding decoupling capacitors alltogether. this one I like even more. After your B+ you have a pair of decoupling resistors going to the two preceding stages. One of those could be adjustable. Because these two stages (we are talking standard RC coupled common cathode stages) are out of phase. Therefore the back emf between each of these former stages and the B+ output impedance can be balanced with adjustment of the adjustable one. (Hope you can follow the dyslexic way of explaining it, in my head it is clear as crystal but sometimes when transfered to the writen word it confuses others.)

2/ and 2/B covered in Fig. 12.54 page 537 RDH4

Now 2/ sub section B we can extend the meachnism to an input pentode. Of the two decoupling resistors spoken about in 2/ we take one to the pentode screen resistor and the other to the pentode anode resistor and we have a completely capacitorless decoupling beyond the main B+ without having to resort to shunt valves or VR tubes..

In 2/B we still have to decouple the screen, but that can go to the VR tube methode now popular.

[slowmotion]

Has anyone done more thinking on this subject?
Have dragged the RDH4 down from top of the bookshelf,
but in the meantime, by coincidence I stumbled upon this thread here:

http://www.audioasylum.com/cgi/t.mpl?f=tubediy&m=29549

where Lynn and Thorsten in a very roundabout way discuss strange circuits
related to this subject in a way, sort of ....

[IslandPink]

Not sure if you got to a conclusion on this, but I believe the feature in the 300B grid/cathode circuit is positive bass feedback , as described by Thorsten in his WE91B variant :

http://www.audioasylum.com/forums/tubed ... 71898.html

"The odd looking circuit between the cathode of the 300B and the 300B gridleak resistor is also original WE, it is a form of positive feedback that overcomes the limited LF frequency response cause by using fairly small value (but high quality) cathode bypass capacitors. In the old days large value capacitors where too large and expensive, today the same holds true for high quality types, hence this circuit still has use. Note it only works with penthode drivers or in specific configurations with interstage transformers, no point adding it to a triode driver amplifier."

I do remember this circuit ( I built in ~2008 ) had very good low bass.

[slowmotion]

Mark, no conclusion from my side what so ever, just a lot of curiosity
from years of looking at WE schematics and trying to figure out what and why. Thanks for the link, I'll read it over the weekend and see if it gets any clearer.

[IslandPink]

I'm still looking at the Thorsten circuit, as that mod could be very useful , and I will likely continue using pentode drivers mostly anyway eg. in the forthcoming 4P1L/GM70 amps.

I'm having trouble visualising how it works exactly, but so far, to me, it seems more like a negative feedback effect that becomes increasingly ineffective as you go below 20Hz . For instance, the 1.5u/10k ( to ground ) combination forms roughly a 10Hz corner.
Does anyone have a clearer or alternative explanation ?

I will have a go at modelling this on LTSpice in a couple of weeks when I get back from hols .