G2 mod gigs well

[Andy Le Blanc]

I finally had chance to gig an amp with bleeders between screen grids and
ground at the tube socket, after the typical 470r screen grid resisters

The mod moves the screen slightly negative with respect to the plate
instead of raising the screen grid resistor value, a divider instead of a series resistor

It was great, the transition to soft clip was smooth and glossy, no fizz...
Iconic blues tones... played very well and kept a definition when you dig in
for solos, really responsive to attack, sweet natural distortion.

After gigging it it seems worthwhile, It took a typical fender circuit, with
nothing special for tubes, transformers, speaker and components
and sweetened it up to exactly where you hope it should be.

a couple of 220k resistors, the g2 ended up a couple volts negative to the plate
and the bias was set to 50% of PT current rating, not really concerned
with the "tube bias" or type as long as its within the published max for
whatever you end up using. Nice result for such a simple approach.

[Cliff Schecht]

I do this in really big bottle SE builds already, especially with tubes that have a max plate voltage much above the screen. Agree that it's a simple add-on and a great sounding addition for such simplicity (one resistor for SE, 2 for PP). Lower screen voltage equates to a higher transconductance which makes the output stage easier to drive into saturation.

[Andy Le Blanc]

I've had fun with this one, most simple push pull circuits with common tube types
benefit audibly its gotten me more focused on the screens and off the
screen grid resistors, off tube bias and on to the power supply and a reliable
set of measurements to utilize at set up, and repeat the set up

The amp really was a pleasure to use, everything was in the right spot

Ive got another on the bench that runs 6l6, 6550, EL34 and KT88
set the current draw on the PS (bias), and set the g2 V at your know spot
The results are remarkably similar for all tube types.

just mind the relationship of tube bias to the PS, its more important to know
that tube bias is also the set up for the power supply, the PS is really what your manipulating, NOT the tube....

the G2 V can be set to taste, like tube bias, but its really the position of
a virtual grid in these beam tubes, thats what your fiddling with when you
swap around screen grid resistors, you CAN measure the V difference between
the screen grids and the plate, and use it as a setting the same as "bias".

[Firestorm]

You know, if Randall Smith had done this, he'd already have filed a patent application for his new Screen Bias Adjustment Technology (SBAT).

[Colossal]

You know, if Randall Smith had done this, he'd already have filed a patent application for his new Screen Bias Adjustment Technology (SBAT).

That's a great point and Andy you'd better jump on this if only to confound Smith. You could set up a PayPal fund; we could all contribute to it to defray the costs of the paperwork on your behalf. Thanks for reporting (and giving freely of) your efforts. Sounds like a very useful mod and goes beyond the usual Fender-Marshall-Vox thinking.

[Tone Lover]

Andy could you break this down a little further for us newbies this sounds like a good area to really break it down and get deeper into.
Sounds like you got very faverable results with a simple yet effective voltage lowering trick, if Im not mistaken between the plates and screens of the power tubes. Sad part is I still dont quite understand what you did. I see 2 people posted while I was writing, it sounds important and if thats the case I understand were you wouldnt want to.
Thanks Bill

[David Root]

I'll have to try that on my EH150 build. It has 470R screen resistors, where the original 1941 circuit of course had none. The screens at quiescent are about 1V or so above the plates. Using 6L6GAs.

What are the effects of going up or down from 220K? I realise it's probably not a good idea to find out by putting in a pot, because of the high DC across it.

Should probably be 2W resistors too, I guess, to be safe, yes?

[Cliff Schecht]

Agree on overrating because a constant DC current is going across that resistor is harder on parts than an AC signal. I'd go with 1W minimum in the power section for some safety margin. It's always good to calculate the power rating of a new addition like this and furthermore going with a higher wattage part means it will have some overhead on the often ignored voltage rating of the resistor.

[overtone]

Andy could you break this down a little further for us newbies this sounds like a good area to really break it down and get deeper into.
Sounds like you got very faverable results with a simple yet effective voltage lowering trick, if Im not mistaken between the plates and screens of the power tubes. Sad part is I still dont quite understand what you did. I see 2 people posted while I was writing, it sounds important and if thats the case I understand were you wouldnt want to.
Thanks Bill

Bill, here's some back ground reading in the meantime:
http://ampgarage.com/forum/viewtopic.ph ... highlight=
http://ampgarage.com/forum/viewtopic.ph ... ht=#163028
best, tony

[David Root]

Thanx Cliff, I figured 400 V to ground thru 220K is 0.73 W, therefore 2W. 3W even better, I think.

[Andy Le Blanc]

Its a very old approach, Crowhurst was a proponent.

What gets me is the assumptions made in design approaches.

There was a time when that Pittman GT book was the only easy to get resource.

O scope bias isn't a good method, but its in the book so a lot used it

It could be the same with screen grid resistor values, is it a valid approach?
because it recommended in someones book?

That post with vox's popping rectifiers is case in point, someone made a assumption...

Give the approach with G2 a try, repetition and confirmation is needed
I'd like to hear what you think.

I'm really jazzed that the measured V difference between plate and G2
looks like a usable number, there's been a good bit of discussion about
the difficulty of measuring over a series screen grid resistor, looks like
a usable DIY approach, like measuring bias over a 1r at the cathode

Those bleeders in old fender PS where small...
You can get away with a 1/2 w 220k, but not a 100k, the smaller the value
the higher the rating, there'll be a practical limit here, 3-5w certainly

thanx, overtone, for posting the links, I've 4-5 gigs a week and cant cover
all the loose ends

[Tone Lover]

Thanks Overtone I will definitly read up, thanks Andy for all your work. I seem to always be one thread late or 2 years worth of reading behind but Im catching up.
Bill

[David Root]

Andy, how do you figure you can get away with 1/2W in a 220K?

If you have a 400V screen to ground thru 220K, that is 0.73W at idle. Yes it will drop voltage as you crank the amp, but the screen's not going to drop to 200V where a 1/2W resistor might be OK.

One thing I nearly forgot, what kind of resistor? Metal Film, metal oxide, wirewound? Most resistors will tolerate about a 350V drop. I might use a 1W PRP MF, rated for 500V.

[Andy Le Blanc]

I scabbed together what was on the bench, just to see...

Its a bleeder... the plate V was around 413 at the bias point and
the G2 was -2.4 from there, higher values of bleeders just get the G2 v
out of positive territory can be low wattage, but you'll flame a 1/2 w 100k

the bleeder doesn't follow the screens current changes that result in V swings, its referenced to ground

If you go for a larger V drop at the screens for better regulation its a different story
you'll need to up the wattage, I've put the bleeder before and after the
screen grid resistors
Before , you need a set resistor, the bleeder can be low watt there too
the set resistor can be 100r too 1k, set the v drop, even an extra filter cap
then a 100r to 470r screen grid resistor

I like the voltage divider before a minimal screen grid resistor
you can uses a low wattage bleeder, the "set" resistor gives you the control

the typical 6v6/6l6 amp has the g2 slightly positive to optimize the
virtual cathode that makes the beam pentode a pentode
but you end up with fizz, that little swing as the g2 relation ship changes
relative to the plate when you apply signal, that little modulation draws
current which ends up as larger voltage swing at the end of series resistor
the bleeder just needs to back off the g2 v by a couple three volts
so you can get away with a low watt 220k, but you start really dropping the v
with a 100k there and it don't fly, up the wattage.....

[David Root]

OK, I can see that the bleeder resistor isn't following the screen current changes now, because it's ground referenced.

But I do not understand how a half watt 220K can stand up at 400V screen to ground. Please explain that, I am missing something. I am looking at it like a cathode resistor, V/R=I.