Grid Leak

[trem]

This has always eluded me.
I have basically read there are two things that cause grid current:
1. Close proximity of grid to cathode causes grid to get very Warm/Hot. This causes electron flow just as it does with the cathode getting hot, but to a much smaller degree.
2. As the screen is controlling the current flow between Cat and Plate, some of the electrons strike the plate and "bounce off". This also causes grid current.

Whatever the reason for grid current, my question is...
Why does the current flow "out of" the grid, through the grid leak resistor, and back to the cat.? Why does not the grid current flow to the Very Positive plate, just like the Cathode Current does.?
Thank You

[es345]

some definitions I use in the text below

- grid = control grid or g1
- flow = flow of the electrons

About the flow:
On the way from the cathode to the plate some electrons don't pass the g1 but just land there. Without any connection path from G1 ground the G1 is getting more and more negative, at the end stopping the flow. If you now connect a resistor from the grid to ground these electrons are flowing from grid to ground.

Some additional infos:
How much of the electrons are flowing from the Cathode to G1 (and not to the plate) depends on the tube type, even the individual tube. The higher the value of the resistor G1 to ground the higher is the negative resulting voltage at G1 but also the variance of that voltage.

The maximum allowed G1 resistor value found in the datasheet of a tube type for fixed bias operation is an indicator for the amount of grid leak current.
The smaller the allowed resistor value is the higher is the grid leak current of that tube type. As long as you use a value equal or less the supplier assures a stable operation independant of the individual tube.

[trem]

Thank You -
You have explained things very well.
So I still have that One question.
Once electrons are "stuck" on G1, it is easier for them to follow a 1M resistor back to the cat, than it is to be drawn to the plate with (just a round number) 200 VDC.? If you understand my confusion..... it seems like getting back to the cat is a lot harder trip than being sucked off of G1 by the high voltage plate.?
Thanks Again

[R.G.]

Once electrons are "stuck" on G1, it is easier for them to follow a 1M resistor back to the cat, than it is to be drawn to the plate with (just a round number) 200 VDC.? If you understand my confusion..... it seems like getting back to the cat is a lot harder trip than being sucked off of G1 by the high voltage plate.?

A good question is how come the electrons can stick to the grid, anyway. The grid is not all that much cooler than the cathode, which is quite near it. The answer is that the surfaces are different. There is a certain amount of energy needed to boost an electron out of a surface. Different conductors have different requirements. Cathodes literally boil them out with heat, raising the average speed of the free electrons with raw heat.

For most metals, the necessary temperature to overcome the barrier to flying off the metal surface is white hot. Directly heated cathodes run nearly as hot as incandescent light bulbs. Special coatings are used to let oxide coated cathodes emit electrons when they're only dull orange.

And that's the answer - the cathode spits electrons easily, the grid is not coated like the cathode and is somewhat colder, so once an electron happens to bang into it and stick, it has to stay there until either the grid gets white hot, or until sheer electrical field rips the electron out in an arc. The physics are such that it take huge voltages to do that, and the grid, not being made of tungsten, would melt before it got hot enough to emit.

So the electrons stay there, causing more and more reverse bias and turning the tube more and more off.

We prevent this by putting in a resistor that lets them run out of the grid. The grid current is quite small, so even with large resistors, the voltage caused by the grid current coming out is small and does not leave much back-bias, and certainly no build up, because a resistor will conduct right down to zero voltage.

[Enzo]

The beauty of grid leak bias though.

Say that accumulation of electroncs on the grid wants to find ground, and sees that grid leak resistor as its path. It is not unusual to see 5 meg resistors and even 10 meg. Do the Ohm's Law and see that across a 5 meg resistor, even one tenth of a microamp still yields half a volt and 0.2uA brings that up to 1 full volt. Which is right in line with our usual needs. ground the cathode and that electron accumulation makes the grid naturally negative with respect to cathode.

[trem]

Once Again... Thanks
So the real issue is that electrons Do Not bounce off of G1, they stick to it. In order for the Plate Voltage to strip (ala cathode stripping) electrons from G1, the HT at the Plates would have to be REAL high. So they have an easier time exited from G1 via a 1M resistor.
I cannot tell you (enough) how much I appreciate your guys Time/Education.
Every time I learn one more of these "things", it REALLY adds to my ability to Repair/Build guitar amps.
Thanks Again To All

[loudthud]

You usually don't get grid current unless you also have a lowish plate voltage. In fact if you have grid current where you shouldn't, like at the guitar input jack, the triode connected there will usually have a low plate voltage, 100V or less. In some experiments I did, the grid leak resistor value didn't make much difference in plate voltage when changed from 4.7 meg to 22 meg. It sort of self regulates.

[Enzo]

Yeah, I think the important thing is the resistance just not be low enough to drain the charge off faster than it accumulates.

To use grid leak bias, you need to have a cap isolating it from the input, otherwise the low impedance of the instrument will drain off the charge and you lose the tube bias.

[trobbins]

You also can't normally directly 'test' the grid voltage (ie. voltage aross the grid leak resistor) because the test circuit resistance will be similar to the grid leak resistance and in parallel. The normal test method is to increase the input signal and check for saturation/cut-off clipping balance, or imbalance, depending on where you want the bias voltage to nominally sit.

[R.G.]

So the real issue is that electrons Do Not bounce off of G1, they stick to it. In order for the Plate Voltage to strip (ala cathode stripping) electrons from G1, the HT at the Plates would have to be REAL high. So they have an easier time exited from G1 via a 1M resistor.

You got it.

There's one problem with what you said, and that's purely terminology, I think. Cathode stripping refers to the removal of the oxide emitting surface of the cathode by bombardment by gas ions, or the chemical reactions with the same gas ions that change the emitting oxides to non-emitting chemicals. "Cathode stripping" sounds like the plate pulling stuff off the cathode, but isn't. The oxide surface does get "stripped" of the low-work-function oxides, and emission drops a lot.

You can even get the odd bit of emitting oxides stuck to the grid after being blasted off the cathode by heavy gas ions, and now the grid emits electrons. Even worse, the actual heater winding can get dosed with oxides, and wind up emitting its own electrons - it's even hotter than the cathode - and these all come off in synchrony with the AC power pulsing through the heater. You can get oxide-induced heater hum! These are just a few reasons that residual gas in tubes is bad.

But the idea is right - the plates don't have enough voltage to pull electrons out of the grids unless the grid surface is poisoned with cathode detritus somehow or way hotter than grids get. The electrons that don't get repelled by the few volts of negative grid voltage stick.

[trem]

Thanks R.G. -
10-4.
You guys are a well of info/knowledge (per usual).
You are right, I was referring to (perhaps the myth of) the electron emitting material that is stripped from a cathode if High Plate Voltage is present. My understanding is that guitar amps typically do not approach that type of B+, so cat stripping is not an issue for "us".
As you say, electrons that are "stuck" to G1 would be the same principle, but a different event.
Thanks

[Diablo]

So are there any advantages to grid leak bias vs cathode bias for the 1st stage of a guitar amp? Any magic mojo?

[Enzo]

SInce going from one to the other is a matter of a couple of 2 cent resistors and a cap, try them and see what you think.

[trem]

So are there any advantages to grid leak bias vs cathode bias for the 1st stage of a guitar amp? Any magic mojo?

SInce going from one to the other is a matter of a couple of 2 cent resistors and a cap, try them and see what you think.

On the same note (I guess).....
Do any of you guys use Grid Leak Bias when you build a guitar amp.?
Does not seem to be very common at all.
Thanks

[Enzo]

It was more common 50 years ago and more. It is MUCH more sensitive to the tube used, so the "normal" way is far more consistent and reliable.