Pentode LED bias + screen grid divider

[blackcorvo]

So, I recently remembered about this topic https://el34world.com/Forum/index.php?topic=14096.0 , in which it's shown that you could use a voltage divider to feed the screen grid (g2) of a preamp Pentode, and it made me think of another possible use for it.
If you were to connect the lower resistor from the divider to the cathode instead of ground, couldn't you put an LED (or other diode) there to ground for biasing?
The divider would feed enough current to turn on the diode and keep it from being turned off when the tube is driven hard. Well, at least I think so?


I'd love to hear people's thoughts on this. I'll put a test together by next weekend, but if anyone feels like experimenting, have at it!

[PRR]

Most modern LEDs will stabilize with just the cathode current. I don't think turn-off is a major deal. (If you drive the stage THAT hard you are not expecting linearity.)

Adding a divider adds waste heat, but not enough to matter in context of a Whole Amplifier.

Do the test.

[blackcorvo]

The test I did, and weird the results were.


Seems like there's a much more direct effect on the Plate voltage by the Screen Grid's voltage when using a resistor divider, than with a traditional resistor-and-capacitor divider.
With the LED in place, this effect is even more pronounced, and makes it kinda difficult to find the sweet spot for highest output swing.


Since I usually have a lot of trouble to calculate pentode gain stages, I just do them by ear, so I'm still thinkering, and still haven't found the right values.


I'm using a 12AU6 going into a 12AQ5 SE stage @ 275v B+, around 250v on the preamp B+.


[EDIT] I attached a schematic of my circuit, and named the plate and screen grid resitors with generic labels R1, R2 and R3, to facilitate talking about them throught this discussion.

[blackcorvo]

After some tweaking, the best I was able to achieve was a gain of 40 with all 100k's. Not ideal for a single stage.

It seems the higher cathode voltage of an LED is too large to work with the voltage divider for G2. However, using a signal diode 1N4148 worked with the values from the thread I referenced in the original post (gain over 300 IIRC), but no combination of values for R1/R2/R3 from my schematic could ever match a similar gain.

LED + voltage divider might be interesting for a second stage you wanna overdrive, but not for a V1 IMHO.

[PRR]

Have you tried bypassing the screen with a cap?

[blackcorvo]

Have you tried bypassing the screen with a cap?

You mean, alongside the voltage divider? No. But now that you mention it, I'll try!

[PRR]

If there is significant impedance facing the screen grid, gain goes down. For small pentodes, pencil the screen as "20k". You are driving it with two 100k, so the screen sees effectively 50k. I suspect gain will go up appreciably with a 0.1u cap from screen to ground or cathode.

[blackcorvo]

If there is significant impedance facing the screen grid, gain goes down. For small pentodes, pencil the screen as "20k". You are driving it with two 100k, so the screen sees effectively 50k. I suspect gain will go up appreciably with a 0.1u cap from screen to ground or cathode.

After some tweaks, and adding your capacitor suggestion:

R1/R2/R3 = 220k/47k/33k (this should give a screen impedance of 19.39k, right?)

P = 47.8 V
G2 = 81.5 V
K = 1.87 V

Input @ 1Khz 0.2 Vp-p = 20 Vp-p Output
Input @ 1Khz 0.4 Vp-p = 38 Vp-p Output

This is with R3 going to the cathode. Grounding it doesn't seem to change much.

Neat! 100x Gain is definitely MUCH more useable, but if it were just a bit more... 

[shooter]

if it were just a bit more...

that's what the next stage is for, why is the better ?
20VAC will stroke out ANY pre-amp gain stage grid.
wouldn't it be simpler to just use a tube screamer pedal to generate square waves before the amp 

[blackcorvo]

if it were just a bit more...

that's what the next stage is for, why is the better ?
20VAC will stroke out ANY pre-amp gain stage grid.
wouldn't it be simpler to just use a tube screamer pedal to generate square waves before the amp 

It's a single gain stage going into a power tube. It's just fun trying to squeeze every ounce of gain from a single tube!

Pedals are fun too, but not very functional for me. No coordination to stomp on them while playing. Barely got any for actually PLAYING!

However, I might try 2 stages later, using 5725's. I've used them in the past, so I know they can sound good overdriven. Also I don't have another 12v filament Pentode handy, so 2x 6v ones it is.

[PRR]

> 100x Gain is definitely MUCH more useable, but if it were just a bit more...

Let's look in the manual. You probably have near 300V supply, and you picked 500k gridleak/volume on power tube. GE gives several conditions near here. "Best" seems like 240k plate load with specified Rk and Rg2, gain of 240 (very load sensitive) and 53Vrms at overload.
http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6AU6A.pdf

Note that 53V/240 means input overload is 200mV. Many guitars and players can exceed that, and distort before the volume control. OTOH some passages play best near 20mV and this one-stage amp can't boost that to full power. Single stage is a compromise. More stages of less gain each is more flexible.

Oldest Fender and cheap Kay use same/similar pentodes, but "gridleak bias" which overloaded too easily for most pickups and players.
https://el34world.com/charts/Schematics/files/Fender/Fender_champ_5c1_schem.pdf
https://el34world.com/charts/Schematics/files/Kay/Kay703.pdf

[shooter]

Input @ 1Khz 0.2 Vp-p = 20 Vp-p Output

I'm seeing G1 Max sig @ ~~~25Vac p-p (12.5 Peak) from the PA datasheet so you're there, re-bias til you have the best version of what you're looking for 

[DummyLoad]

If there is significant impedance facing the screen grid, gain goes down. For small pentodes, pencil the screen as "20k". You are driving it with two 100k, so the screen sees effectively 50k. I suspect gain will go up appreciably with a 0.1u cap from screen to ground or cathode.

use a higher impedance ckt. e.g., 3M3/1M ; 2M/470K, etc. for the divider network.


--pete

[blackcorvo]

Input @ 1Khz 0.2 Vp-p = 20 Vp-p Output

I'm seeing G1 Max sig @ ~~~25Vac p-p (12.5 Peak) from the PA datasheet so you're there, re-bias til you have the best version of what you're looking for 

Seems like that's it with this stage, really. Tried tweaking it more tonight, and all I've really found out is that R1 at 100K gives more symmetrical clipping when overdriven than 220k, at the cost of some gain.
I do have some vintage Russian red LEDs that show a voltage drop of just under 1.6v, but they don't get as bright as the one I've been unsing, and one of the reasons I wanted to use LEDs is to actually illuminate the tubes from under the socket.

[blackcorvo]

use a higher impedance ckt. e.g., 3M3/1M ; 2M/470K, etc. for the divider network.

--pete

I'll try that tonight.

[shooter]

I wanted to use LEDs is to actually illuminate the tubes

I switched from tube to SS rec, decided to glue a jumbo LED under a bad rec tube n use it as "power on light" saved ripping out the socket  WOW effect was more whimper

[blackcorvo]

I was looking up ways to get more gain outta pentodes, and stumbled upon this:


http://www.valveradio.net/audio/ultra-low-distortion-valve-hybrid-amplifier.html


Fig. 14C caught my attention. Very simple!


But these are ridiculously high gains (in the order of 1000's). Still interesting tho! Might try it for fun (I do have the parts for it so why not?).

[shooter]

that hurt my face just looking at Pic's N I like looking at pic's 

[blackcorvo]

Okay, G2 divider is not a good idea with LED biasing. No matter what values I use, R3 seems to suck gain. R-C is really the way to go instead. Might be the cathode voltage being too high.


But even using R-C, the values I got to for R1 and R2 with the most gain (50v p-p out, 0.4v in, gain of 125) had the tendency to be the same (220k in this case).


About to try the "ludicrous" gain option.
(Aaand it didn't work, and I'm NOT gonna fiddle with that).

[blackcorvo]

Doing some tests with 2 stages, the divider for G2 seems to help tame fizz when driving the pentode into distortion, so I'm using it for the second stage.


I'll however be using a regular R-C circuit on G2 for V1.
That allows me to adjust the voicing of the first stage by choosing a different value of capacitor for the G2 circuit.


Currently, I have something like a "top boost" circuit, using a 470pF cap for the G2 of V1, and to compensate the brightness, a 100nF coupling cap.


Second stage has a great output swing!
A bit too great, actually.
Keeps knocking the 12AQ5 into motorboating!
So I'll put a tone stack between there and the volume control. Might even use 250k for the volume if push comes to shove.

[DummyLoad]

rule of thumb i've come to live by from tinkering with pentodes:  if Vg2 is > than Vp when overdriven usually sounds really harsh. if Vg2 is kept < Vp, then overdrive is MUCH smoother sounding. < plate current sounds better overall with less noise and hiss. try to keep Av at around 100-125. i have pushed past Av 125 with B+ => 350V. if RF is an issue then shunt RF with a 22-47pF cap connected grid to cathode - keep the leads as short as possible and do that at the socket. 


--pete

[blackcorvo]

rule of thumb i've come to live by from tinkering with pentodes:  if Vg2 is > than Vp when overdriven usually sounds really harsh. if Vg2 is kept < Vp, then overdrive is MUCH smoother sounding. < plate current sounds better overall with less noise and hiss. try to keep Av at around 100-125. i have pushed past Av 125 with B+ => 350V. if RF is an issue then shunt RF with a 22-47pF cap connected grid to cathode - keep the leads as short as possible and do that at the socket. 


--pete

Didn't really get any RF interference, but I added 47pF caps anyway, on both tubes. It helped tame ice-picky highs a tad (barely noticeable, but you're glad it's there, kinda thing).
I also changed the 470pF for a 1nF. More "orangey" drive tones, still with that "top boost" vibe but a bit meaner.
Tone stack definitely helps, but will take a bit of fiddling to get just right.

[blackcorvo]

Here's the schematic and some sound samples:

https://www.youtube.com/watch?v=WsjAcTRAV98

I changed the 100nF coupling cap for 10nF because it was a bit too "flubby" when using the neck pickup for any drive sounds.