Neg Bias Quest

[C Moore]

Is it bad or wrong to run the bias voltage straight to the power tube grids.?

A typical scenario has the Neg Bias Voltage split by a pair of 220k resistors...but these resistors are on the PI side of the 1.5k power tube grid resistors.

Does it have to be that way for some reason.?
Could the 220k drop the bias voltage on the power tube side of those 1.5k grid resistors.?
Thank You

[Phil_S]

I'm an amateur, so my explanation may be less than technical and probably incomplete in some respects.

What you propose won't work. The pair of 220K (or whatever) resistors are called "grid leak" resistors. They provide a reference for the power tube grids. Look at a typical grid leak biased preamp tube. The grid leak resistor is grounded. In the case of a fixed bias power tube, the negative voltage is, well, a sort of negative elevation of the grids (did I really say that?) to keep them operating in a "space" that establishes the proper flow in the power tube.

Without these reference resistors, the grids are floating. Simply running a negative charge to them might cause them to float at a different point. I'd expect unstable operation.

I'm sure someone else will jump in to fix whatever I may have misstated here, but I'm pretty sure I covered the basics.

[shoggoth]

It's only floating if there's no common ground reference between the negative bias supply and the positive supply that's powering the tube.

[potatofarmer]

Those grid leak resistors also set the input impedance of the power tubes, e.g. the load that the phase inverter has to work into.

Here, check out this calculator and see what happens to the gain of the LTP as the grid leaks get smaller: http://ampbooks.com/home/amplifier-calc ... iled-pair/

edit: Oops, thought you were talking about getting rid of the 220k grid leaks entirely and just applying the bias voltage directly to the grids. My bad

[shoggoth]

As the grid leaks get bigger, you mean - "removed entirely" would equate to infinite resistance.

That simulator is for the PI, not the power tubes, so it won't say much about the effects of removing the grid leak resistors on the power tubes.

The grid leak resistor does provide an alternate path to ground for your signal, that is in parallel with the resistance from the signal to the negative voltage through the negative bias supply. It will also act as a voltage divider with the impedance of the negative bias supply, and it will provide a load for the bias supply, so that its caps discharge after power turns off.

All this without looking at any schematics, so I could be missing something.

[Firestorm]

The 220Ks don't drop the bias voltages. They provide the AC ground reference the grids require. Look at a cathode biased amp: the 220Ks are in the same place, but go to ground. In fixed bias, the bias supply is an AC ground, so it's the same topology. The grid stoppers work best right on the tube, so you don't want to move then. If you put them on the other side of the grid leaks, they'll form a voltage divider that will drop a teensy bit of the signal.

[FUCHSAUDIO]

The series grid resistors are usually required to maintain stability in many amps. The RCA manual says they are required, and I wouldn't argue with those guys.

[C Moore]

The 220Ks don't drop the bias voltages. They provide the AC ground reference the grids require. Look at a cathode biased amp: the 220Ks are in the same place, but go to ground. In fixed bias, the bias supply is an AC ground, so it's the same topology. The grid stoppers work best right on the tube, so you don't want to move then. If you put them on the other side of the grid leaks, they'll form a voltage divider that will drop a teensy bit of the signal.

Yours is the only answer I am understanding.
Anyway.....I see what you are saying about the divider. So 1.5k compared to 220k...there will not be much loss... right.?
I am just wondering about placing the grid leaks right on the grid...not talking about moving the grid stops anywhere.
Thanks

[FUCHSAUDIO]

Fender using Pin-1 as a tie-point for pin 5 for the grid resistor is a perfect solution. I've seen amps get all spitty and shitty without them or in one amp, we wired to 5 instead of one (the resistor was there, it was a senior moment), and the amp motorboated....

The tube manual says basically "as close to the grid as possible". We've gotten away with it on PC board an inch or two away from the tubes in most models, but recently one new model showed problems that were only fixed by moving the grid stopper to the socket.

[C Moore]

Fender using Pin-1 as a tie-point for pin 5 for the grid resistor is a perfect solution. I've seen amps get all spitty and shitty without them or in one amp, we wired to 5 instead of one (the resistor was there, it was a senior moment), and the amp motorboated....

The tube manual says basically "as close to the grid as possible". We've gotten away with it on PC board an inch or two away from the tubes in most models, but recently one new model showed problems that were only fixed by moving the grid stopper to the socket.

Hey Andy - I hear what you are saying...thanks for the info.
Just to be clear, I am not talking about moving the stopper off of the grid...just moving the 220k on to the grid also.
What I am asking...isn't that how A Lot of amps are wired at the input. The typical Fender.....it has the 68k and the 1M both wired to the grid...in a divider scenario as Firestorm mentioned above.
Thanks

[Phil_S]

Putting the stopper and the grid leak resistors both on the pin repositions the grid leak and by passes the stopper, which I don't think you want to do. The stopper and the grid leak are wired like a voltage divider (though given the relative values it doesn't really function like a divider.) The amp, as designed, delivers the negative bias voltage to the grid and also requires the negative supply to pass through the stopper.

You could try this, but why?

[talbany]

Fender using Pin-1 as a tie-point for pin 5 for the grid resistor is a perfect solution. I've seen amps get all spitty and shitty without them or in one amp, we wired to 5 instead of one (the resistor was there, it was a senior moment), and the amp motorboated....

The tube manual says basically "as close to the grid as possible". We've gotten away with it on PC board an inch or two away from the tubes in most models, but recently one new model showed problems that were only fixed by moving the grid stopper to the socket.

Hey Andy - I hear what you are saying...thanks for the info.
Just to be clear, I am not talking about moving the stopper off of the grid...just moving the 220k on to the grid also.
What I am asking...isn't that how A Lot of amps are wired at the input. The typical Fender.....it has the 68k and the 1M both wired to the grid...in a divider scenario as Firestorm mentioned above.
Thanks

The 1M at the input does Basically 2 things
1. Provides a gnd reference for the grid
2 The resistor value sets the input impedance of the amp

The 68k Grid stopper (in Series) acts as a lowpass filter to help block stray RF from entering the grid

In the bias supply. The total resistance is the sum of the series grid resistor and the bias feed or "grid-to-ground" resistances, (usually around 220k)..The bias feed resistors also sets the RC time constant (along with the PI coupler's) that sets the Bias excursion/Recovery time,(which also has a direct effect on blocking distortion)..Along with the value of the grid resistor

Tony

[C Moore]

Putting the stopper and the grid leak resistors both on the pin repositions the grid leak and by passes the stopper, which I don't think you want to do. The stopper and the grid leak are wired like a voltage divider (though given the relative values it doesn't really function like a divider.) The amp, as designed, delivers the negative bias voltage to the grid and also requires the negative supply to pass through the stopper.

You could try this, but why?

I am glad you wrote this.
I guess that is exactly what I do not understand.
What is the advantage for the incoming AC to ride the Neg Bias Voltage on one side the the Grid Stop Resist compared to the other.?
Or what is the advantage of the AC ground reference being on one side of the grid stop or the other.?
Thank You

[talbany]

Or what is the advantage of the AC ground reference being on one side of the grid stop or the other.?
Thank You

Re-read my post

The bias feed resistors also sets the RC time constant (along with the PI coupler's)

So they should to be connected to the PI coupling caps! and the other side close to an AC ground source
Tony

[Firestorm]

As long as you're dealing with Fenderish values, it is probably not going to make an audible difference. If you move the grid leaks and the grid stops right to the tube, signal attenuation will be less than one percent and the grid stop will still be in series with the signal so it will still work with the tube's capacitance to block high frequency garbage. The differences come if you overdrive the tube to the point of grid conduction. The time it takes to charge the coupling cap (making it a clamp) will be different and the time it takes for the cap to bleed off its voltage (recover) will be different, as Tony said. But if we're talking about 1K5 and a stock Fender, I'll bet it isn't noticeable.

But again, why bother?