12au7 load line

[labb]

I have a 18Watt TMB Superlite that I am trying to reduce the volume output. I put a 12AU7 in the 1st preamp position and this tames the amp down to where the player wants it to be..So I started looking at the load line to see where the tube was operating. I don't really know a lot about load lines so I am kinda in the dark..The amp has 240 volt dc to the preamp section . The tube has 100K plate resistors and 1.5K and 820 ohm cathode resistors. I am just look at the first part of the tube,,the part with the 1.5K cathode resistor for now. Ploting the load line, 240 volt for plate voltage on the curves. 240/101500=2.4 milliamps for the plate current...The cathode load lines cross the plate load line at about 2.3 milliamps and about 67 volts plate volts..This puts the tube operating way down on the curves and way to the left...So I looked at a plate current of 10 ma to see what the total resistance would be. I got 240/.010=24000. just does not seem right. When I look at schematic that use the 12AU7 I see numbers like 220K for the plate resistor and 5600 for the cathode resistor...What am I doing wrong..TKS

[shooter]

I've use 12AU7 as V1 in most builds, started with AX but found too much gain most of the time.
My "fix/understanding" is use the cheat sheet posted and know what gain I want (Vo/Vin)
so say the AX was 35, I cut in half and aim for mid teens by using the chart.
smarter folk will have the proper solution 

[jjasilli]

https://robrobinette.com/Drawing_Tube_Load_Lines.htm

[jjasilli]

With the 12au7 substituted-in, more current is drawn.  I'd verify that the supply voltage has remained at 240VDC; it may have dropped a bit.  Looks like the longest loadline for a 12au7 would be a Supply Voltage in excess of 300VDC; and a plate current draw around 10 - 12mA.  That said the "operating area of a 12au7 is deceiving: the grid voltage curves are only 0.5V apart for the ax7; 2V apart for the au7.


Subbing an au7 into a typical ax7 circuit definitely restricts the au7's operating area into the lower left hand portion of its Chart, especially with less than a full compliment of supply and resultant plate voltage.

[PRR]

> 12AU7 ...240 volt dc ... 100K plate resistors and 1.5K and 820 ohm cathode resistors....The cathode load lines cross the plate load line at about 2.3 milliamps and about 67 volts plate volts..

Correct. And gain is about 15 and input signals to 3Vpp in (45Vpp out) will pass OK.

You DO leave a lot of available supply on the table. This leads to a high-gain solution (though 1/3rd the gain of a 12AX7) with not a lot of output swing. This may run just fine.

More general crayon-marks for amplifier stages are to set your plate voltage "around half" of your supply. (If you have 240 peas, and two children V and R, give each child 100-140 peas.) And Rk can often pencil-in as Rp/Mu, giving for Rp=100k about 1k for 12Ax7 and 5k for 12AU7. (Not carved in stone: Fender likes 1.5K 12AX7; G.E. liked ~~4k for 12AU7.)

[shooter]

Just a FWIW

this is my current amp during tuning.  (I changed some switched things to fixed, but data out the pre should be close)

[PRR]

And you do NOT (generally) want to use a high current tube at currents like 10mA. That will sag your B+ chain, and force 10X larger preamp filter caps. In guitar-amps the signal loads are a small fraction of a mA, you don't have to run hot. And in resistor-loaded amplifiers, gain and max output goes up and distortion goes down as Rp goes up (until it becomes comparable to the load).

You can also put attenuators between 12AX7 stages, and get whatever gain you want, without the hassle of finding a tube with "the right gain", the re-biasing to make the most of it, and the larger stock of spare tubes to carry.

[labb]

'You can also put attenuators between 12AX7 stages, and get whatever gain you want,'


Thanks PRR and can you explain what you mean about the attenuators between stages?

[jjasilli]

You can also put attenuators between 12AX7 stages, and get whatever gain you want, without the hassle of finding a tube with "the right gain", the re-biasing to make the most of it, and the larger stock of spare tubes to carry.



Typical "attenuator" between gain stages is a voltage divider comprised of a 470K series resistor with 470K shunt resistor.  The series resistor might be bypassed by a 470pF cap to preserve hi's.

[shooter]

Look at the schematic I posted, between V1a and V1b.  there's a R junction of 470k / 280k  "attenuator" .  look at the VAC on plate of V1a 2.2vac  the voltage divider shaved that to 1.2vac (along with normal loss and mis-match impedance)

[labb]

This is the schematic for the amp. Looking at the 470K/470K voltage divider between V1a and V1b. I can lower the value of the 470K resistor that goes to ground and control the gain of this stage. I think Shooter is using 280K here. Is there a limit to what you can change here? And thanks guys for the help. I think that I am a pretty good builder. Been doing it for years but am really weak on the design side.

[shooter]

Is there a limit

There is, usually called out on the datasheet for G1 max R.  I typically just play with them.
I'm a steal-do-tweak-design/builder 

[jjasilli]

Shooter, that's not the voltage divider.  The voltage divider is between the 2nd 470K series resistor and the 200(?)K shunt (also the grid leak) resistor.  Note that output voltage is not being tapped between the two 470K EDIT: series resistors.  To divide voltage the shunt resistor must follow the series resistor, with output voltage tapped from between them.


However that's oversimplified for this circuit.  For AC (signal) operation, I believe that the preceding 470K shunt resistor is in parallel with the 200(?)K shunt (grid leak) resistor.  Caps can also form a voltage divider.  So the series caps, in conjunction with the EDIT:  shunt resistors, might also be doing some voltage division.  And that leg of the circuit is in parallel with the 100K plate resistor of the driving stage - which plate resistor for AC operation functions as a shunt resistor to ground.  The proper analysis of this whole shebang is above my pay grade.  Nevertheless, the result is measured and clear.


E.g., this is often seen with Fender style input circuitry: with a 68K series resistor and a 1M grid leak (shunt) resistor.  If the 68K R precedes the 1M R, they form a voltage divider.  But if the 68K R is wired to follow the 1M R, there is no voltage divider.

[Tony Bones]

I ended up reducing gain in a similar design by splitting the 100k plate load of V2 and tapping off the junction. (Essentially turning the plate load into a voltage divider.

See the Deluxe 6G3 for an example. https://el34world.com/charts/Schematics/files/Fender/Fender_deluxe_6g3_schem.pdf

[shooter]

this whole shebang is above my pay grade.

YUP, me too
In my "example" the original steal was 470k / 680k (I think)
anyway. kept dropping til I got ~ 1Vac and less un-happy sound. 
coulda been because impedance miss-match, or any other reason I don't know about
I just know, it was the direction I was listening for and scope agreed with  ears 

[jjasilli]

Shooter, you may also wish to state the frequency of the 130mVAC input signal.  Because frequency is unstated, people might feel free to "assume" that it is the "standard" 1000 H.  But because of the T or Pi or whatever filtering between stages, other frequencies might get more or less attenuation.

[shooter]

 
I used 1khz for so long, when I started doing tube stuff I switched to 880hz, it's more musical 

[PRR]

> Is there a limit to what you can change here?

Look at the variety of interstage networks Carvin used to get 11 stages cascaded without impossible total gain.

[labb]

Wow!  I will open the amp up Thursday and play with the 470K resistors and see what happens. With the layout used they are easy to change..Would love to get this amp running on AX7's at a reasonable level. Again, guys, thanks for all of the input.

[tubeswell]

Here's a thing you can play around with

[labb]

So I tacked a 470K resistor across (parallel)the 470k resistor that goes to ground between V1A and V1B. That should have given me a voltage divider of 235K/(470k+235K) or a factor of 0.33 times Vin. vrs. the factor of 0.5 the two 470k resistors in the schematic..I could tell no difference in the volume level of the amp. Should I make R2 even smaller, say 100K and see what happens. Any recommendations. TKS

[shooter]

try the suggestion in reply #13

[labb]

I'm kinda lost as to what Tony actually did there..Would I split the 100K plate resistor of V1A to say 50k and 50K, tap between and feed that to the volume pot?

[Tony Bones]

Exactly. That'll cut the signal in half. I think I went as far as 82K + 15K to attenuate even more. A bonus is that the tone stack sees a lower source impedance which tends to make them work a little better.

[labb]

Thanks Tony. I'll dig through my supply of resistors and see if I have the correct ones. If not I'll get several combinations on order and give it a try.

[labb]

Too follow up on trying to lower the volume of this amp:

Well I did some mods to the amp...I put a 10K resistor between the 100k load resistor of V1A and the tube and tapped off between them to go to the volume pot. I did the same with V1B. Tks Tony.  I put a 470K resistor in parallel with the 470k resistor that goes to ground between V1A and V1B. None of these had much effect on the volume of the amp. I put a 50K pot in series with the 820 ohm PI cathode resistor. This allows me to really drop the volume. TKS Steve. These changes along with using a 12AU7 tube for V1 really tames the amp. The 12AU7 tube makes a huge difference.

any other suggestions and I will give them a try while I have the amp on the bench.

[Tony Bones]

TI put a 10K resistor between the 100k load resistor of V1A and the tube and tapped off between them to go to the volume pot.

You need to swap the resistors. Connect the 10k to the PS and the 100k to the tube...

[labb]

TKS. Tony. Thought I might have that backwards.