Connecting G3 to a negative supply on a pentode

[bozole]

Hi

I'd like to have your opinions about this kind of connection.
Some people say that it gives more linearity, and it's a safier way to use some (fragile ?) pentodes than connecting G3 to the cathode (for an EL34 for example)

It was used on some guitar amps made by Traynor in the 60's or 70's, using EL34, model YBA3A for example : http://www.lynx.bc.ca/~jc/700116_YBA3_A.gif

 

[Eli Duttman]

Remember, g3's job is to force electrons knocked out of the plate back to the plate and keep them from being swept up by g2. If g3 is a FEW Volts lower than cathode potential, no big deal. OTOH, if g3 is significantly below cathode potential it will act as a 2nd control grid and choke the flow of electrons to the plate off.

 

[smoking-amp]

Here is a thread about g3 effects:
http://www.diyaudio.com/forums/tubes-valves/160240-suppresor-grid-used-feedback.html

Usually g3 gets positve voltage of 20 or 30 Volts for eliminating "snivets" on the sweep tube datasheets. And then there was some talk a while back about running EL34s without any connection to g3. From testing some other tubes, looks like that makes the g3 go negative from charge buildup. The effects on beam tubes are relatively small. Pentodes show a little more effect. The dual control tubes in the linked thread show large effects.

 

[bozole]

Hi

Many thanks for your answers, both of you

On the link you posted smoking-amp (and what a great "job" to show us all this curves !!!), we can see that the effect from varying g3 voltage is very tube relative, as you mentioned here, that's interesting, and as you told, it can have a big influence on screen current ... That's what I understood for the EL34, and I understood that for this tubes, connecting g3 to the negative bias supply had the effect to really make them more reliable !! (essentially because of the screen current I suppose ?)

If you have some time and EL34 available smoking-amp (but I don't want to disturb you, so let me know), could you please trace some curves with EL34 "playing" on g3 voltage ?

OTOH, if g3 is significantly below cathode potential it will act as a 2nd control grid and choke the flow of electrons to the plate off.

yes, and I suppose very different effect with different tubes as smoking-amp mentioned
On the Traynor amp which from I posted the schematic, the g3 is at a significant negative voltage, -45V ...

 

[smoking-amp]

The Traynor amp schematic linked above shows 6KG6A output tubes, which are beamers. So they would take plenty of voltage on g3 to have much of any effect.

Unfortunately I don't have any EL34 tubes here to try out. The GU-50 in the g3 thread should be somewhat similar in effects, being a pentode. Normally, I would expect a negative voltage on g3 to round the knees of the plate curves, with more current sent back to the g2 grid in that region (occuring at large signal input, near clipping). So I would be careful putting any neg V's on the EL34 g3. I suggest putting a milliamp-meter in series with the g2 grid while experimentally adjusting the g3 negative to see if g2 current goes up (with a sine wave signal, near max allowed, on the input). You will then have to calc. the max limit for g2 dissipation allowed, then back up some.

A somewhat similar effect to neg g3 knee rounding can be gotten by increasing the g2 stopper resistors. And since this drops the voltage on g2 near clipping, it is much safer for the tube.

Take a look at the 6HJ5 tube plots linked below for g2 stopper effects. Then compare that with the 6LE8 g3 plots in the other thread.
http://www.diyaudio.com/forums/tubes-valves/151206-posted-new-p-p-power-amp-design-94.html

http://www.diyaudio.com/forums/tubes-valves/160240-suppresor-grid-used-feedback-6.html#post2083661

 

[Miles Prower]

I'd like to have your opinions about this kind of connection.
Some people say that it gives more linearity, and it's a safier way to use some (fragile ?) pentodes than connecting G3 to the cathode (for an EL34 for example)

Not a good idea. Some pentodes specify some positive bias on suppressor grids/beam formers to reduce the tendency to oscillate. Negative bias on suppressor grids can easily lead to instability, and, indeed, back in the good ol' days was frequently done to produce a negative resistance screen characteristic that was the basis for UHF oscillators. This trick replaced using the negative resistance plate "kinks" of pure tetrodes for this purpose when tetrodes were largely replaced by pents.

Unless it's your intention to make an oscillator, I'd avoid it.

"It was used on some guitar amps made by Traynor in the 60's or 70's, using EL34, model YBA3A for example..."

Just because they got away with it doesn't necessarily mean they knew what they were doing.

 

[SpreadSpectrum]

Not a good idea. Some pentodes specify some positive bias on suppressor grids/beam formers to reduce the tendency to oscillate. Negative bias on suppressor grids can easily lead to instability, and, indeed, back in the good ol' days was frequently done to produce a negative resistance screen characteristic that was the basis for UHF oscillators. This trick replaced using the negative resistance plate "kinks" of pure tetrodes for this purpose when tetrodes were largely replaced by pents.

Interestingly, smoking-amp's GU50 curves seem to show the opposite behavior with kinky curves with positive suppressor voltages and less kinkiness with negative. The curves never seem to develop negative resistance.

I agree with smoking-amp that negative bias would be harder on the screen with increased screen current. Definitely not a safer way to operate, but maybe more linear depending on the application.

It would be interesting to see EL34 curves.

 

[TUBESMAN]

Interesting, all I wanted to do was propose a way to get UL operation in a RLD! I'd say if you control the voltages and watch your current in g2, and anode, it might work. Not sure how it can make an oscillator, but an EL34 seems like a lousy way to create one. I have a box load of 6ag7's for that!

So from smoking-amp's test, It looks like Vg3 can have a wide range of affects on a tubes behavior. Looks like we either need to live with pentode operation or find another way to go UL and LED bias.

In the past, UL was done as a separate winding on the output transformer. Is it possible to carry this one step further and have a separate UL winding for each tube? I asked Edcor to do something like this and they said they cannot do it. Is it just Edcor that can't or is it impossible for a reason I currently don't understand?

 

[DF96]

The main effect of pushing g3 negative is to push more of the cathode current to g2 instead of the anode. The total current (set mainly by g2 and g1) will be largely unchanged, so I would not expect any significant change in linearity.

 

[Osvaldo de Banfield]

The main effect of pushing g3 negative is to push more of the cathode current to g2 instead of the anode. The total current (set mainly by g2 and g1) will be largely unchanged, so I would not expect any significant change in linearity.

Exactly, but I add that if no means are provided to limit screen dissipation, it will be destroyed, screens aren't designed to carry all the cathode current.

Such a principle is used in Phantastron and Sanatron devices other times used to get linear voltage rundowns, and I test a EF184 (+ ECF8101) in such a connection at 100KHz some years ago.

 

[Wavebourn]

Positive G3 allows higher G2 voltage, for the same G2 current, hence higher anode current swing and lower saturation voltage. I used that in my amps with GU-50, 813 and GK-71 output tubes.

 

[TUBESMAN]

Acrosound made the transformer I want 60 years ago, the TO-350. Toroidy makes a cathode feedback transformer now, the TTG-CFB8000PP I'm sure they can adjust the turns ratio to the 19% that Arcosound used instead of 10% they are using. As far as I can tell, UL, RLD is solved. No g3 issues needed. Just lost technology from the 1950's.