6AU6 screen fed by voltage divider?

[jeff]

I had a little amp that I need to transplant to a bigger chassis because I wanted to add a tube and had no room. It is one 6AU6 one 6AQ5 and I'm adding a 12AX7 for more gain. I made a schematic based on what I saw before I removed all the parts. I've never used a 6AU6 before but noticed in most schematics I've seen using a 6AU6, usually the screen has a dropping resistor to get the correct voltage and a cap to ground. According to the schematic I drew the screen is feed buy a voltage divider and has no cap. I'm almost 100% sure that I drew my schematic corectlly before I disasembled the amp but I've never seen a 6AU6 curcuit like this.

Is this an alternate way to provide screen voltage? Do the values seem reasonable or does something look fishy? Is this a valid way to run a 6AU6 or is it more likley I drew my schematic wrong?
 The only other cap in the amp was the .033, but it seems to me if the screen had a cap it'd be bigger than .033.

I'm pretty sure I drew it correctally but I've never seen a 6AU6 set up like this.

(Program had no 6AU6 grafic so I had to use EL84so ignore pin #s, the 270K and 150K aren't in parallel the line looks like it's connected but only crosses over with no connection)

[DummyLoad]

either way works.

150K over 27K is approx. 1/6 of B+ to screen.

--DL

[HotBluePlates]

Well your drawing seems to indicate the screen is connected to the plate, but I'm getting that's not what you intended. The dot where a line from the plate crosses the screen divider indicates a connection.

It is possible to use a divider for the screen, but there is almost universally a cap across the resistor to ground. That keeps the screen voltage from varying under signal conditions. Gain with your divider, and no bypass, will probably be close to operating the 6AU6 as a triode.

And usually, the screen is fed by a series resistor bigger than the plate load resistor (when there is only a cap to ground). I can't think of a reason the designer would use a divider, unless those parts were also connected to something else (meaning, there was a 2nd function for those parts).

If it were me, I'd use a screen resistor about double the plate load, and a 0.1uF cap to ground.

[jeff]

Thanks

I was very careful when I drew my schematic and checked and check before I disassembled it so I was pretty sure that's how the original was wired. I just had to ask because I've never seen this setup for a 6AU6 before.

Thanks again
  Jeff

[jeff]

HBP

I didn't notice but you must have slipped in your reply as I was also replying to Dummyloads post.

So do you see this as a red flag that I made an error when I reverse engineered the schematic? Or are you saying I'd get more out of it by using a resistor/cap over the divider?

[jeff]

I just came across this

It was in a post and he mentioned that he changed the .033 cap to a 20 on his schematic and added the input jack 33k and 1M but it looks like this is the exact ciruit I drew down to the trim pot for the cathode resistor. I measured the value of mine and got a total resistance of 1K5 vs his value of 576 though.

So I'm thinking I probally did get it right.......or else we both reverse engineered it wrong...aghh!

Does 204V on the plate and 18V on the screen seem way off to you?

[sluckey]

18v on the screen seems way off. Ohm's law calculates 42 volts with 278v across a 150K plus 27K voltage divider.
 

[tubeswell]

You can get good screen current regulation with the voltage divider method, but the result is more hi-fi cleans and a harder overdriven sound (than if you used just a simple g2 resistor and g2 bypass cap).  As to where you set the screen voltage, that depends on how much gain you want out of the stage and where you've biased the stage. Lower screen voltage brings down the knee of the Vg0 grid curve, and this can be useful to get the chosen load line passing through the knee of the curve for the desired tone.

[HotBluePlates]

18v on the screen seems way off. Ohm's law calculates 42 volts with 278v across a 150K plus 27K voltage divider.

That's just for divider current, but the screen is drawing current as well. So the 150k is passing more current and dropping more voltage than we'd think.

Now the screen current should be a small fraction of total divider current, but who knows in this case.

Looks like Jeff reverse-engineered it right. Try it both ways and see what works for you.

[jeff]

I stumbled across some other Zenith schematics, not the same model, and it seems that's the way they did it-with the voltage divider.

Thanks for the help

[DummyLoad]

A bogen K10 that came across my path recently met my hacker dark-side; i recall that it had a similar arrangement for g2 power. attached is the schematic for reference.

why is your Vg2 low? if R1 is the upper resistor connected to B+ to g2; R2 is the lower 1/2 of the divider connected from g2 to ground; since current flowing through the Kathode is = to Anode current + g2 current, then g2 is in parallel with R2 (kind of, but for now it's easier to simplify with that view) so, assuming the engineers had picked our tube's parameters beforehand, we would assume that R1/R2 divider network would need to flow at least 10x the current needed by g2 in order to not grossly upset the ratio of the R1/R2 divider network.

you CAN configure the divider network to flow enough current to make just about any bias parameters of your pentode work IF you select values (read: lowER Rs) that flow enough current; the down side with that approach is that you excessively load the PS.

back to your low Vg2, this seems to not be the case if you're truly reading 18V on g2 - if your measurements are indeed correct, then g2 is sucking a lot of current in that bias configuration, or could this just simply be how this stage was engineered, in that this combination of R1/R2 gave the engineers the Vg2 they were looking for.  

to summarize: if you have plagiarized an ENTIRE stage and/or you know what your B+, Ia, & Ig2 currents will be by design, then you can make the R1/R2 divider work for you. otherwise, take the high road and use a singe series resistor bypassed to ground + your B+ will love you more.
 
the RCA Resistance Coupled Amplifier Charts provide a wealth of knowledge - attached are excerpts relevant to the 6AU6.

--DL

EDIT:

IR1 = 278V-18V = 260V so then IR1 = 260V / 150KΩ = 1.733mA

IR2 = 18V / 27KΩ = I of R2 = 667uA

so by KCL then IR1 = Ig2 + IR2.  rearranged; Ig2 = IR1 - IR2 = 1.733mA - 667uA  = 1.066mA

VRa = 278V-204V = 74V, so then IRa = 74V / 270KΩ = 274uA, yes, that's microamps or read as 0.274mA. you show .65V Vg, so then Ik = .65V / 576Ω = 1.13mA; if current through the cathode resistor = sum of screen current + anode current, then Ik = Ia + Ig2 = 1.066mA + 274uA = 1.34mA which is slightly more than what you show with Vg/Rk (.65V/576Ω). voltmeter loads the grid, measure from cathode to ground, that is very close the -V bias of the control grid. IOW, if you measure 0.65V from cathode (pin 7) to ground then the grid voltage is  approximately -0.65V.

would you measure Vdc from pin 7 to ground please? ....and lastly, why the 16 ohm trim pot? the trim pot does seem moot.

[jeff]

Just woke up so I'm gonna give that a thorough read and digest before I can respond
Thanks for the help

would you measure Vdc from pin 7 to ground please?

This was not my schematic, so I don't have voltages. I'll rebuild according to the schematic and post voltages I get. Like I siad, I was worried that I reverse engineered my schematic incorrectly and this schematic, to me, comfirmed that I didn't.

I'm drilling holes right now and building circuit, when I juice it up I'll post data.

....and lastly, why the 16 ohm trim pot? the trim pot does seem moot.

I think he meant the trim pot was set to 16 ohms and not that it is a 16 ohm trim pot. When I measured the one in mine it read ~1K.

He mentioned he removed the NFB which is why it does not appear in his schematic. Originally a 39K was used attached from OT secondary to the trim pot/cathode resistor so It seems to me the pot both adjusts the bias of the 6AU6 and changes the divider(made up of a 39K and the pot/560 cathode resistor) adjusting the NFB.

[PRR]

> possible to use a divider for the screen, but there is almost universally a cap across the resistor to ground.

If the audio impedance at the G2 is not down around 20K, gain is lost.

The 150K||27K divider makes 23K, low-enough to get 90% of max gain.

The downside is excess load on the B+. A 1Meg dropper and 0.1u cap would draw much less B+ current.

The up-side is cost. 0.1u 400V caps are expensive. 0.1u 200V caps eventually fail from the 300V-400V at startup before G2 starts sucking current. An extra mA of B+ current may be cheaper and more robust than an R-C dropper.

> 18v on the screen seems way off. Ohm's law calculates 42 volts

The 42V neglects G2 current, yes? So it will be lower.

18V on G2 is not wrong for a SMALL-signal pentode. Total voltage gain goes up as G2 goes down. If gain goes up, while output overload hardly changes, then input overload must go down. But what was the original input? A phono needle. The output is weak and WELL-controlled (there's limits how hard the disk-cutter can whack the wax). Input was under 1V. Input overload is very-roughly Vg2/Mu(G1G2). 6AU6 Mu (G1 to G2) is about 20. So 18V/20 suggests input overload near 0.9V. Which is entirely fine for phono duty (and possibly kewl for guitar).

[jeff]

Thanks

 It's gonna take a little longer than expected. I used treminal strips and back mounted sockets so I could build the ciruit outside the chassis and slip in. Well, I had it all layed out, this socket goes here that cap goes there transformer goes in the bottom left, and cut the hole for the transformer then it dawned on me... I layed it out looking at it from the inside and I just cut the Xformer hole in the bottom left... from the outside! Which means it's now in the upper left from the inside! I have to go back and mirror image everything.
D'oh!


 Jeff

[John]

I have to go back and mirror image everything.

I did the exact same thing when I built the HoSo. It's so much easier to follow the layout when everything points the right direction!

[jeff]

Sorry it took so long. Here's schematic and voltages. Some values drifted and I can't measure exact values of 27K and 150K while in curcuit(meter goes up and up) so assume +/- 10%

[sluckey]

32v screen sounds much more reasonable than 18v.

[jeff]

Thanks again for the help. It sounds great! Right now I'm using it for a speaker driven reverb/MP3 amp and haven't run straight guitar through it.

I was wondering how to add a dwell pot to the tank. I e-mailed Mojotone because they sell the mojoverb which is a speaker driven reverb unit and they replied that their dwell pot was a 1K linear pot, but I'm not sure how they wire it or what it's rated for. Would it just be a pot before the resistor or a varible resistor to increase the resistance? I don't really want to be able to turn down to 0 so would wiring it as a variable resistor allow me to adjust the drive to the tank?
What should the 1K pot be rated for?