Effects of resistor values on a v1 cathode/cap and controlling gain

[jukelemon]

Hi all.

Can someone simply explain the effects on the pre amp tube when you raise or lower the resistor value on the cathode/cap combo?

Second question, is it possible to control the gain of a v1 tube by altering XXX.  So, for example, is it possible to vary a tubes gain/effect on an input and how would one do that?

Possible?

So like instead of only having to choose from a 12at or 12ax, you could do XXX, put in a 12ax but decrease the gain to get down to a 12at if you so chose to do.

Thanks.

[jjasilli]

From Dave Funk's Tube Amp Workbook:  The K resistor is the main gain setting component: may range fro 820R to 50K;  lower values and the tube won't operate.  Above 50K is unity gain.  Try replacing a the K resistor with a 50K pot, in series with a 820R fixed resistor (to maintain bias) and you'll have a practical answer to your question.

Another step is to delve into tube specs, charts and loadlines.  Overall the gain of the stage is set by:  input circuit (voltage divider present?); grid leak resistor; supply voltage; plate resistor; K resistor; plate impedance (depending on which tube type is used); and grid leak resistor of the next stage. 

On a graph: the supply voltage & plate resistor form a straight line, called the loadline, at a certain angle & height above the base line.  The operating point is set by the cathode resistor and appears at a particular point on the loadline.  A signal wave too high above the operating point will saturate; too far below and it will clip.  It can do both.  So there are definite results to the values of components in the circuit.  The values chosen by manufacturers are pretty standard because they work well.

Also, too much gain in the first stage tends to send excessive hiss & noise down the signal path.

[jukelemon]

Hi.

Can you please explain Unity gain?

In practical terms, if one lowers the resistor (K resistor) does gain increase or decrease?

Thanks for the explanation.

[Willabe]

Here read this it might be helpful;

http://www.valvewizard.co.uk/gainstage.html


             Brad     

[Willabe]

Can you please explain Unity gain?

What goes in same goes out. 1v in 1v out. So really no gain.


            Brad     

[tubenit]

From ARCHIVES  FAQ (posted 8 yrs ago):

22)  Is there  some guideline for the values of a cathode resistor with plate resistor in
       regards to gain?

Generally speaking you’ll find ratios such as 100k (plate) to 1.5k (cathode), or 220k/2.2k or
330k/3.3k, etc……    There are exceptions to this of course.  In other words, they tend to be
a ratio of 100:1 units

            Examples if you change nothing else:
Increase plate resistor value = more gain   Decrease plate resistor value = more headroom
Increase cathode resistor value = more head room   Decrease cathode resistor value = more gain
           
             For Example:  a 100k/ 1.5k has less gain than a 100k/820 ohm

A browner more tweed sound may have 140 volts on a 12A_7 plate and 160 -200volts may have a cleaner sound and more headroom.

[jukelemon]

Thank you. It is getting clearer.  Thank you

These answers make sense for one tube BUT is it possible to vary the gain (literally with some sort of pot or xxx) of a tube in parallel with v1?

Unity gain was mentioned i.e. a cathode resistor above 50k would lead to Unity Gain and then someone else explained that as 0 Gain.

If that is the case, could you sub in a 50k pot across the cathode cap and control the gain structure of the tube?

I am not saying it would sound good or bad (at some point it would sound good/ok).  I am not talking tone at this point.  I am just talking controlling the tube's characteristics to provide gain to a stage.

I hope I am making sense.

Simply put, I am wondering how one can get in between a 12ax and a 12at7 in real time and without having to swap tubes.  It may be impossible..I don't know.

[Willabe]

If that is the case, could you sub in a 50k pot across the cathode cap and control the gain structure of the tube?

jj already answered you on this in reply #1;

Try replacing a the K resistor with a 50K pot, in series with a 820R fixed resistor (to maintain bias) and you'll have a practical answer to your question.

              Brad     

[jukelemon]

If that is the case, could you sub in a 50k pot across the cathode cap and control the gain structure of the tube?

jj already answered you on this in reply #1;

Try replacing a the K resistor with a 50K pot, in series with a 820R fixed resistor (to maintain bias) and you'll have a practical answer to your question.

              Brad     

Yes, I know that this was mentioned.  But there were also many other things mentioned and I wanted to make sure I was nto simplifying i.e. I could change JUST this and control the gain structure.

If it is that simple, then cool 

[sluckey]

Unity gain was mentioned i.e. a cathode resistor above 50k would lead to Unity Gain and then someone else explained that as 0 Gain.

Gain is defined as output voltage (V_out) divided by input voltage (V_in).

    Gain = V_out / V_in

So, the only way to have a gain of zero would be to have zero output voltage. Unity gain is the same as a gain of 1. This occurs only when the output voltage is the same as the input voltage.

Some real world examples of unity gain tube circuits as seen in guitar amps would be a cathode follower or cathodyne phase inverter. Actually the gain of these circuits is always slightly less than unity, typically 0.8 to 0.9, but it's often useful to think of this as "unity gain" when doing a quick analysis of a circuit.

[Willabe]

So, the only way to have a gain of zero would be to have zero output voltage. Unity gain is the same as a gain of 1. This occurs only when the output voltage is the same as the input voltage.

Thank you for correcting that.


           Brad     

[Willabe]

Simply put, I am wondering how one can get in between a 12ax and a 12at7 in real time and without having to swap tubes.  It may be impossible..

That's what a volume control does.

You could add an extra volume control on the stage(s) you want to have extra control over. But how many volume pots can you fit on the chassis face plate, or do you want?

Or you could put a trim pot for the grid leak R inside the amp on each stage to do this. Tweak them where you want them then close up the amp. Unless you have at least 3 true gain stages do you really need them? Even with 3 gain stages a volume and master volume are plenty.

Or you can temp in a pot, for the grid leak R, adjust it to where you want it/sounds good then without changing it's setting take it out measure the resistance it's set at and put in a R that's that value.  

When you start wanting to get between the gain of a 12AX7 and a 12AT7 your starting to split hairs.


             Brad    

[sluckey]

So, the only way to have a gain of zero would be to have zero output voltage. Unity gain is the same as a gain of 1. This occurs only when the output voltage is the same as the input voltage.

Thank you for correcting that.


           Brad     

It wasn't meant as a correction. You said "What goes in same goes out. 1v in 1v out. So really no gain.". And that's perfectly true. It's not the same as saying the gain is zero. Confusing, ain't it!   

And to add even more confusion,,,, "gain = 0" does mean "input = output", logarithmically speaking, log(1)=0.   

[Willabe]

Well how bout, thank you for giving a clearer definition.

And yes it is a little confusing.


             Brad    

[HotBluePlates]

Second question, is it possible to control the gain of a v1 tube by altering XXX. 

Simply put, I am wondering how one can get in between a 12ax and a 12at7 in real time and without having to swap tubes.

I added the emphasis on "real time".

Is this a requirement for you? Is it also a requirement that nothing else changes when gain changes? If so, it changes the whole approach.

The "organic" way is a variable cathode resistor, as folks have said. But tube voltage amplification isn't only determined by the cathode resistor, but also the supply voltage, plate load resistor and (as a result of the other 3 things) internal plate resistance of the tube.

All examples below assume a fully bypassed cathode resistor.

Below is a 12AX7 loadline with 5 different cathode resistors indicated from 820Ω to 10kΩ. Where each cathode line crosses the loadline is where the tube will idle with the resulting bias point. The short stubs which seem to mirror the gridlines is the internal plate resistance at each operating point. We could calculate what each of those numbers is, but just know that the more-vertical these lines are, the lower the internal plate resistance.

My shorthand calculation won't be fully accurate, but the internal plate resistance with the 820Ω is around 66kΩ while it's ~133kΩ with the 10kΩ cathode resistor. The equation for stage gain shows the output due to a tube's mu being divided between the internal plate resistance and the external plate load resistor, so with a given plate load you get a higher gain with a smaller internal plate resistance.

With our 12AX7 and 100kΩ resistor, gain drops from about 60 for the 820Ω cathode resistor to about 42 with the 10kΩ resistor.

But did you notice how the point the cathode lines cross the plate loadline at lower and lower current? The gridlines get crowded down there and with a big enough cathode resistor (which you think will give lower gain) you will get more distortion from a gvien input signal because the gridlines are equally spaced. You also run into tube cutoff.

[HotBluePlates]

My last post assumed the cathode resistor was bypassed. What happens if it isn't?

You get local negative feedback across the cathode resistor. You could probably slide from a normal plate-loaded stage towards a split-load stage (like the inverter) whose gain to either output is 1 or less.

The problem is the changing cathode resistor changes everything about the stage (not just its gain). It would be nice to just have a control that varied stage gain and nothing else.

You could do local negative feedback, but from the plate to the grid. See the example below, which is from the tweed Super. Pay attention to the series resistor going into the grid of 270kΩ and the feedback resistor 5MΩ. The gain of the loop will be:
Gain = 1+(Rf/Rg) = 1+ (5M/270k) = ~19.5 (for the entire loop).

Now the Super has a 12AY7 gain stage and a cathode follower in that loop, but you could just as easily have the 5MΩ come directly from the plate output of the 1st 12AY7 stage.

The cathode resistor for a gain stage inside such a local feedback loop should be fully bypassed, as you want it to have full open-loop gain before feedback (this helps it clean up effectively as you apply more feedback). You have to juggle the value of series input resistor to use and feedback resistor value with some thought given to available pot values and what range of gain you need.

Do you have a gain range in mind? We could probably work up a simple circuit to knock a 12AX7 down to 12AT7 or 12AU7 gain. The possible problem is what value pot would be needed.

EDIT: Merlin has a page dedicated to this idea. Local Feedback

[tubenit]

is it possible to vary the gain (literally with some sort of pot or xxx) of a tube in parallel with v1?

Geezer's original Tweed Overdrive Special

[jojokeo]

working on another patent idea jukelemon?

[jukelemon]

working on another patent idea jukelemon?

Not sure what you mean.

[jukelemon]

is it possible to vary the gain (literally with some sort of pot or xxx) of a tube in parallel with v1?

Geezer's original Tweed Overdrive Special

Interesting. He REALLY just uses a Volume pot to control the effect of the parallel tube. Cool.

[jjasilli]

1.  Nomenclature clarification.  Gain is always a multiplier.  So, Zero gain = No output, because any number X 0 = 0.  If signal strength input is equal to signal strength output, that's Unity Gain.  The multiplier for Unity Gain is "1"; hence the name "unity" (a unit = 1).  Any number X 1 = that same number.

There is another way to speak in terms of signal strength: Decibels (dB).  So, another way of saying "Unity Gain" is "Zero dB of Gain".  That's the way to talk about "0" , when signal-in = signal-out.  

Unfortunately the only way to really get into the theory of gain is to learn loadlines.  I found it to be a painful learning curve, just to get the jist of it.  The patient help of Dummyload, Sluckey PRR & Hotblue was invaluable!

2.  Simplicity.  A 12ax7 has a nominal gain factor of 100; A 12at7, 60. That's a difference of 40%.  A cathode bypass cap approx. doubles the gain of a tube, a difference of 50% which is in the ballpark.  So switching out the bypass cap on a 12ax7 stage will sorta make it perform like a lower mu tube. 

You can do more radical things to mess with the gain of a 12ax7 tube.  But again, if you force it to perform at only 60% of its gain potential, then it will will simulate a 12at7, but only in terms of the gain factor.  The loadlines & operating points will probably differ significantly for a 12ax7 vs. a 12at7 at such operating points. So tone, in terms of saturation, clipping and harmonic distortion generally, may be significantly different. 

Though gain is important, it is not the only, nor the most important factor.  Any tube type produces far more gain than we need.  The most important factor is Tone, which derives from the operating point. 

[LHPcope]

There are also other differences than just the gain, or mu, between a 12ax7 and the rest of the 12a_7 family of tubes, like plate resistance.  A circuit designed for a 12AX7 does not provide the ideal operating parameters for a 12AU7 even though the 12AU7 will work in the 12AX7 circuit.  There's a reason we see 12AT7s used for reverb drivers and PIs many times. 

For gain control take a look at the final stage of a MKII tonebender circuit; While I have not tried this with a valve I see no reason why it wouldn't work to vary the gain.

[jojokeo]

For gain control take a look at the final stage of a MKII tonebender circuit; While I have not tried this with a valve I see no reason why it wouldn't work to vary the gain.

it does work

[printer2]

I do not see a problem with putting a 10k pot in series with a 470R. So what if the operating point changes and you get different distortion characteristics. I would think that would give you a greater palette to work with. I experimented with a pentode using pots in the cathode, plate and screen positions to see the effects of changing each. Triode is one order simpler, give it a shot and see how you like it.

[jjasilli]

Below is a 12AX7 loadline with 5 different cathode resistors indicated from 820Ω to 10kΩ. . . But did you notice how the point the cathode lines cross the plate loadline at lower and lower current? The gridlines get crowded down there and with a big enough cathode resistor (which you think will give lower gain) you will get more distortion from a gvien input signal because the gridlines are equally spaced. You also run into tube cutoff.

So, picture the signal as a sine wave with its upper half above the loadline, and its lower half below the loadline.  For normal signal voltage swings, the 820R cathode resistor has plenty of room above and below the loadline.  Lots of clean headroom; little chance of distortion.  At 1.5K there's a chance of clipping for the bottom half of the wave. At 5K almost the entire lower half of the wave is cutoff.  That's true for the voltage (300VDC supply voltage) and plate resistor (100K) specified.

Basically, if you swap a 12AT7 into that circuit -- 300VDC supply; 100K plate resistor; 1.5K cathode resistor -- then the 12AT7 will behave something like a 12AX7 with an 820R cathode resistor: Lots of clean headroom; little chance of distortion.  Swapping tubes on the fly is a hassle.  Other alternatives a substantial change in supply voltage; hence VVR; or a SW to alter the dropping resistor value in the B+ supply.  Cascading gain stages & Master Volume. More than one way to skin this cat.

[HotBluePlates]

I do not see a problem with putting a 10k pot in series with a 470R. So what if the operating point changes and you get different distortion characteristics. I would think that would give you a greater palette to work with. ...

I agree, as long as you know what you're getting.

In other words, the rule presented was "Low cathode resistor = Higher gain" and most will associate gain with distortion. But I showed how "Low cathode resistor = Less distortion in that stage" while a larger cathode resistor seems to reduce the gain and output voltage of the stage while increasing distortion within that one stage.

The local feedback approach reduces gain and distortion at the same time within the one stage. And you can do it by either leaving your adjustable cathode resistor unbypassed or adding plate to grid feedback.

The only thing left is to taste-test and see what works for the amp and the desired result.

[jjasilli]

In other words, the rule presented was "Low cathode resistor = Higher gain" . . .  Yes, that'w why it's necessary to look at the load line, so your chart was very helpful.

. . .and most will associate gain with distortion. Yes, especially with modern, hi-gain amps - which tends to distort the understanding (no pun intended) of the basic or vintage circuit. The modern hi-gain circuit makes a very tall soundwave out of a rounded sine wave input.  As the wave gets taller ("excessive" amplification), its walls become nearly perpendicular; the top will clip, giving it square shoulders and a flat-top.  Now the sine wave input has been distorted into a square wave output with harsh harmonics.  Voila, a metal/shred amp.  So hi-gain has become associated with distortion.

[Willabe]

1.  Nomenclature clarification.  Gain is always a multiplier.  So, Zero gain = No output, because any number X 0 = 0.  If signal strength input is equal to signal strength output, that's Unity Gain.  The multiplier for Unity Gain is "1"; hence the name "unity" (a unit = 1).  Any number X 1 = that same number.

There is another way to speak in terms of signal strength: Decibels (dB).  So, another way of saying "Unity Gain" is "Zero dB of Gain".  That's the way to talk about "0" , when signal-in = signal-out.

Thank you for the Nomenclature clarification. Very helpful.


             Brad     

[tubeswell]

Try replacing a the K resistor with a 50K pot, in series with a 820R fixed resistor (to maintain bias) and you'll have a practical answer to your question.

              Brad     

Actually, try replacing the cathode-bypass cap with a 5k pot wired as a variable resistor in series with a 25uF cap, and use a cathode resistor between 820R and 1k8. That way you will have a choice of unbypassed Rk for lower gain and fully bypassed Rk for maximum gain boost.