Technical question on resistor functions..

[diddymix]

Hello folks,

I was hoping anyone with techinical knowledge and experience could enlighten me on the function of some parts of a circuit, namely resistors and what theyre doing.. and how changing the values does what.. I really appreciate input as Im trying to grasp how to fine tune part of my amp so its balanced how I want it.

The first is the resistor that is in the signal path after the plate and the coupling cap.. and before the grid of the next tube (preamp or PI). For example in a dumble there is a 220k with a 500pf cap across it.. Or for example in a JTM45 you would see the volume pot followed by a 270K ..
Am I correct in saying this is to reduce level/gain before the next tube?? What would be the effect of increasing or lowering this value.. again would it be less/more gain into the next tube?? Is there another reason for it?..

Lastly with regard to grid leak resistors, specifically on overdrive stages... usually one would see a 1M grid leak, am I correct in saying this changing this has an effect on the clipping behaviour of the tube? Again, decreasing the value for example.. what would this do with regard to gain/tone..

Any advice most welcome and please let me know what youve found in your builds/experience! Cheers

[R.G.]

Good on you for trying to learn!

One answer that doesn't have any obvious relation to your questions, but is really crucial to your understanding: a resistor is a device that causes a voltage to appear across it linearly related to the current through it. The ratio of voltage that appears across it to the current through it is the resistance. I bring this up because you always, always need to keep this in your mind when thinking about resistors. A 1K resistor is the same as saying "one volt per milliampere". A huge thing about resistors, more related to your questions is that frequency doesn't appear in that volts-per-ampere definition is that resistors are ideally completely frequency insensitive. A 1K is a 1K at DC or 1MHz. Well, mostly. At RF, the signals start leaking out into space and the resistor leads look like inductors, but in general, think of a resistor as being frequency insensitive.

Directly related to your question is that there is no way to tell what a resistor is doing (except for that volts per amp thing above) without knowing the rest of the circuit. The whole circuit matters, not just the resistance. So to know what changing a resistance in a circuit does, you have to understand the whole circuit around it. Sorry for that being another non-answer, but it's true. I'll hit your questions directly below.

The first is the resistor that is in the signal path after the plate and the coupling cap.. and before the grid of the next tube (preamp or PI). For example in a dumble there is a 220k with a 500pf cap across it.. Or for example in a JTM45 you would see the volume pot followed by a 270K ..
Am I correct in saying this is to reduce level/gain before the next tube?? What would be the effect of increasing or lowering this value.. again would it be less/more gain into the next tube?? Is there another reason for it?..

Starting with the tube plate => cap => resistor => grid:
It's a mostly tone control in this case. However, to understand it, you have to realize that there are "hidden parts" in that circuit. The plate of the driving tube has an internal resistance of about 60K (for a 12AX7); the receiving grid almost certainly has a 500K-1M grid leak resistor to its bias voltage. Worse, the receiving tube has a grid-to-cathode capacitance, and a grid-to-plate capacitance and this last acts as though it's bigger by the voltage gain of the receiving tube. ACK!
So the signal at the plate of the driving tube has to travel through the 60K plate resistance, the coupling cap, and the series resistance to get to the next grid. At the next grid, it also has to drive the grid leak resistance, the grid-cathode cap, and the grid-plate-multiplied capacitance (also called the Miller capacitance). The thing that makes the series resistance into a tone control is those parasitic, hidden grid capactances. They are frequency sensitive, and suck more current linearly as frequency goes up. By putting that series resistor between the driving stage and the grid, the series resistor "eats" some of the available signal voltage depending on the current flowing through it, and the grid capacitors allow more current to flow as frequency increases. So less highs come through than lows. The series resistance makes a frequency sensitive voltage divider, a low pass (or high cut, same thing) filter with the grid capacitances.

The resistor in series with the pot wipe is a little tougher to suss out. It might be doing a bit of the same, knocking off little high frequencies, depending on what comes after it. Or it might be evening out the levels given by the pot. A pot is a voltage divider, and the voltage at the wiper is equal to the resistance "below" the wiper divided by the whole resistance times the incoming signal voltage. The equivalent resistance that the signal at the pot goes through, much like that plate resistance of the driving tube above, depends on the rotation of the pot. I won't bore you with the math, but the signal impedance of a pot is zero at each end, and a maximum of 1/4 the pot value when the pot is exactly in the middle So it varies. Some circuits would act a little funny with this variable resistance. Putting a single biggish resistor in series with the wiper smooths out this varied resistance, at the cost of some signal level. You have to know the preceding circuit and the successor circuit to figure out how much of what is affected.

Lastly with regard to grid leak resistors, specifically on overdrive stages... usually one would see a 1M grid leak, am I correct in saying this changing this has an effect on the clipping behaviour of the tube? Again, decreasing the value for example.. what would this do with regard to gain/tone.

.
You're correct, it has an effect, but what it changes mostly is in the driving circuit, which you haven't specified. Electrically, that 1M grid resistor appears to be in parallel with the grid and the grid capacitances to ground. Whatever is driving the grid has to drive the grid resistor and those capacitances, and the resulting voltage on the grid is what gets amplified. When the tube clips, it can clip in cutoff or grid conduction. The grid looks like an open circuit, all it's imperfections modeled by those grid resistances. Somewhere around -2V or so, the grid's negative voltage cuts off the conduction of the tube entirely, so no current flows in the plate. The driving signal can't "see" this, so all it knows is that it's driving a grid resistor and some caps. Changing the grid resistor makes the driving circuit put out more or less current to drive the resistor part. This is generally so small that it can be ignored UNLESS you have stuck in some additional series resistance, as above. In this case (negative signal clipping) any series resistance is more important, as it puts off negative grid cutoff.

For positive-going signals, the grid is biased generally at a volt or so negative. A positive going signal makes this less negative, until it drives the grid up 0V overall, where the grid suddenly changes from looking like an open circuit to looking something like a 5K -10K resistance as it's driven into conduction. That clips the positive side of the signal because a typically 60K internal impedance driving plate circuit is severely loaded by the sudden change from open circuit to a few K ohms. Again, changing the grid resistor doesn't do too much.

But a grid resistor has thermal noise. The bigger the resistor, the bigger the thermal noise. Lowering the grid resistor value reduces the thermal noise it contributes. This isn't too big a deal, but in critical and high gain circuits, you might get some noise reduction .... if your driving circuit has enough zoots to drive the additional load of the lowered grid resistor.

It's all about the circuits, much more than just the resistor. The water is a little deeper than it looked.

[martin manning]

Good stuff, RG, thanks for taking the time to post. Someone once said in analog circuits each part is doing about three things.

Recommended reading for those who want to understand triode stages: http://www.valvewizard.co.uk/gainstage.html

[diddymix]

Good on you for trying to learn!

One answer that doesn't have any obvious relation to your questions, but is really crucial to your understanding: a resistor is a device that causes a voltage to appear across it linearly related to the current through it. The ratio of voltage that appears across it to the current through it is the resistance. I bring this up because you always, always need to keep this in your mind when thinking about resistors. A 1K resistor is the same as saying "one volt per milliampere". A huge thing about resistors, more related to your questions is that frequency doesn't appear in that volts-per-ampere definition is that resistors are ideally completely frequency insensitive. A 1K is a 1K at DC or 1MHz. Well, mostly. At RF, the signals start leaking out into space and the resistor leads look like inductors, but in general, think of a resistor as being frequency insensitive.

Directly related to your question is that there is no way to tell what a resistor is doing (except for that volts per amp thing above) without knowing the rest of the circuit. The whole circuit matters, not just the resistance. So to know what changing a resistance in a circuit does, you have to understand the whole circuit around it. Sorry for that being another non-answer, but it's true. I'll hit your questions directly below.

The first is the resistor that is in the signal path after the plate and the coupling cap.. and before the grid of the next tube (preamp or PI). For example in a dumble there is a 220k with a 500pf cap across it.. Or for example in a JTM45 you would see the volume pot followed by a 270K ..
Am I correct in saying this is to reduce level/gain before the next tube?? What would be the effect of increasing or lowering this value.. again would it be less/more gain into the next tube?? Is there another reason for it?..

Starting with the tube plate => cap => resistor => grid:
It's a mostly tone control in this case. However, to understand it, you have to realize that there are "hidden parts" in that circuit. The plate of the driving tube has an internal resistance of about 60K (for a 12AX7); the receiving grid almost certainly has a 500K-1M grid leak resistor to its bias voltage. Worse, the receiving tube has a grid-to-cathode capacitance, and a grid-to-plate capacitance and this last acts as though it's bigger by the voltage gain of the receiving tube. ACK!
So the signal at the plate of the driving tube has to travel through the 60K plate resistance, the coupling cap, and the series resistance to get to the next grid. At the next grid, it also has to drive the grid leak resistance, the grid-cathode cap, and the grid-plate-multiplied capacitance (also called the Miller capacitance). The thing that makes the series resistance into a tone control is those parasitic, hidden grid capactances. They are frequency sensitive, and suck more current linearly as frequency goes up. By putting that series resistor between the driving stage and the grid, the series resistor "eats" some of the available signal voltage depending on the current flowing through it, and the grid capacitors allow more current to flow as frequency increases. So less highs come through than lows. The series resistance makes a frequency sensitive voltage divider, a low pass (or high cut, same thing) filter with the grid capacitances.

The resistor in series with the pot wipe is a little tougher to suss out. It might be doing a bit of the same, knocking off little high frequencies, depending on what comes after it. Or it might be evening out the levels given by the pot. A pot is a voltage divider, and the voltage at the wiper is equal to the resistance "below" the wiper divided by the whole resistance times the incoming signal voltage. The equivalent resistance that the signal at the pot goes through, much like that plate resistance of the driving tube above, depends on the rotation of the pot. I won't bore you with the math, but the signal impedance of a pot is zero at each end, and a maximum of 1/4 the pot value when the pot is exactly in the middle So it varies. Some circuits would act a little funny with this variable resistance. Putting a single biggish resistor in series with the wiper smooths out this varied resistance, at the cost of some signal level. You have to know the preceding circuit and the successor circuit to figure out how much of what is affected.

Lastly with regard to grid leak resistors, specifically on overdrive stages... usually one would see a 1M grid leak, am I correct in saying this changing this has an effect on the clipping behaviour of the tube? Again, decreasing the value for example.. what would this do with regard to gain/tone.

.
You're correct, it has an effect, but what it changes mostly is in the driving circuit, which you haven't specified. Electrically, that 1M grid resistor appears to be in parallel with the grid and the grid capacitances to ground. Whatever is driving the grid has to drive the grid resistor and those capacitances, and the resulting voltage on the grid is what gets amplified. When the tube clips, it can clip in cutoff or grid conduction. The grid looks like an open circuit, all it's imperfections modeled by those grid resistances. Somewhere around -2V or so, the grid's negative voltage cuts off the conduction of the tube entirely, so no current flows in the plate. The driving signal can't "see" this, so all it knows is that it's driving a grid resistor and some caps. Changing the grid resistor makes the driving circuit put out more or less current to drive the resistor part. This is generally so small that it can be ignored UNLESS you have stuck in some additional series resistance, as above. In this case (negative signal clipping) any series resistance is more important, as it puts off negative grid cutoff.

For positive-going signals, the grid is biased generally at a volt or so negative. A positive going signal makes this less negative, until it drives the grid up 0V overall, where the grid suddenly changes from looking like an open circuit to looking something like a 5K -10K resistance as it's driven into conduction. That clips the positive side of the signal because a typically 60K internal impedance driving plate circuit is severely loaded by the sudden change from open circuit to a few K ohms. Again, changing the grid resistor doesn't do too much.

But a grid resistor has thermal noise. The bigger the resistor, the bigger the thermal noise. Lowering the grid resistor value reduces the thermal noise it contributes. This isn't too big a deal, but in critical and high gain circuits, you might get some noise reduction .... if your driving circuit has enough zoots to drive the additional load of the lowered grid resistor.

It's all about the circuits, much more than just the resistor. The water is a little deeper than it looked.

Wow thanks so much! What a helpful and informative post. Aha I now see that its really more important like you say to look at what its connected to and from etc.. and this really determines its function! It also seems apparent that from what you say there can be many things to consider, but I think I get the jist as to the examples I posted.. Its interesting, because on my amp I hooked up a switch and made a small preamp circuit option which I can switch from.. and one channel goes direct to the PI.. however compared to the other channel it was quite loud comparatively on the same preamp setting.. so I used a 330K resistor from the new channel to the PI input cap. It actually did attenuate the signal but not by much.. however the tone did change a little and I was wondering why.. but now I know from your info! So its really more a low pass filter, which makes sense.
The grid leak is something I may leave alone, but might look at the grid stopper in conjuncton with that to change the overdrive beahviour more.. I would like to experiment in refining the OD tone more, its quite 'buzzy', and I feel that this could be tweaked to be a little smoother/complex in its overdrive..
Awesome thanks again, I am going to have a play with what youve outlined in mind, then I should be able to know abit more about what Im hearing and why. Cheers