Fixed bias theory

[paggerman]

Hi all. I hope for those who have right to vote today exercise that right. Now onto amps.

Over the last few months I have read dozens of articles about fixed bias and studied heaps of circuit diagrams trying to understand what's going on in a fixed bias circuit. What I understand is:

An ac voltage is rectified to provide a negative voltage
The ac voltage goes through a pi filter to smooth the half wave negative voltage
There is a resistor to ground (56k in a Bassman)

What I have surmised is happening is that the the 56k resistor is behaving like a cathode resistor in a self biased power valve. But I have also read that there is next to no current through the grid, so that hypothesis is on shaky ground.

The other thing that puzzles me is the lack of information that is published about the voltage of the bias tap of a PT. The PT I am working on has a separate winding for bias (two wires) and the HT is not CT (the donor amp had a voltage doubler, as does my current project). The voltage on the bias winding is 24VAC. Is this lack of information because the voltage on the bias tap is not particularly relevant?

I'm sure you guys could tell me exactly what bias voltage range would be knowing that it is a 24v winding, a 2x10uF and 15k pi filter and a 50k Pot and 22k resistor in parallel for the lift. But how do you get to those figures? even RDH4 doesn't help me understand.

Thanks, as always.

[NateS]

There is a resistor to ground (56k in a Bassman)

What I have surmised is happening is that the the 56k resistor is behaving like a cathode resistor in a self biased power valve. But I have also read that there is next to no current through the grid, so that hypothesis is on shaky ground.

The other thing that puzzles me is the lack of information that is published about the voltage of the bias tap of a PT. The PT I am working on has a separate winding for bias (two wires) and the HT is not CT (the donor amp had a voltage doubler, as does my current project). The voltage on the bias winding is 24VAC. Is this lack of information because the voltage on the bias tap is not particularly relevant?

The resistor to ground in adjustable bias circuits sets the bottom of the bias adjustment range - otherwise you could go to 0v bias. The lack of information is probably because it's approximate (depends on your tubes), adjustable, and very low current without anything else in the circuit to perturb it.

[nickb]

You can draw the circuit and simulate it here https://www.circuitlab.com/editor/
Here is a sample (not sure if it is what you described,):



Set up a time domain simulation, probe the bias_out node.




..and here are the results.



Note the transformer is shown as an AC voltage source of 33.84 volts. That's because your transformer is 24V RMS and the peak voltage is 24 x 1.14 = 33.84V

[defaced]

What I have surmised is happening is that the the 56k resistor is behaving like a cathode resistor in a self biased power valve. But I have also read that there is next to no current through the grid, so that hypothesis is on shaky ground.

Take it back to basics. The cathode is an electron gun held at ground potential in a fixed bias setup. In front of it are a control grid, screen grid and plate. The control grid is negative with respect to the cathode, the plate and screen are positive. As an electron (a negatively charged particle) comes off of the cathode, it obeys the laws of physics and is pushed away from the negatively charged control grid and attracted to the positively charged screen grid and plate. Thus, no current flows from the cathode to the control grid. The 56k resistor is part of the bias voltage power supply and has no role of cathode resistor or anything similar. It does have a role in the grid to ground resistance of the circuit and if too large can cause the grid to be too easy to drive which can over dissipate the tube (in theory, in practice it's violated all the time, often by 2x).

[exclamationmark]

A very generic way of doing a bias supply is:
1). Rectify it.
2). Filter it.
3). Stick a voltage divider somewhere to achieve the proper bias voltage.

and optionally

4.) Filter it some more.

For some reason, guitar amp manufacturers like to use rather complicated voltage divider arrangements, and I suspect that's where you're having trouble understanding the circuit. For something relatively simple, there are more variations then you can poke a stick at.

[Chuck H]

It was eluded to in the OP that paggerman might think that the actual bias voltage was unimportant, as long as there was bias voltage. If this is not the case then I misunderstood. Otherwise...

The bias voltage needs to be a very specific amount in order to control the idle current of the power tubes. Since your supply is only 24V I suspect that the bias supply for the donor amp was also a voltage doubler circuit, UNLESS it was an el84 amp. So, count on designing a voltage doubler rectification for the bias supply. Other than that everything is as already explained. The bias supply is a filtered -VDC supply with a voltage divider to adjust it so it provides a specific voltage.

[paggerman]

It was eluded to in the OP that paggerman might think that the actual bias voltage was unimportant, as long as there was bias voltage. If this is not the case then I misunderstood. Otherwise...

The bias voltage needs to be a very specific amount in order to control the idle current of the power tubes. Since your supply is only 24V I suspect that the bias supply for the donor amp was also a voltage doubler circuit, UNLESS it was an el84 amp. So, count on designing a voltage doubler rectification for the bias supply. Other than that everything is as already explained. The bias supply is a filtered -VDC supply with a voltage divider to adjust it so it provides a specific voltage.

Not quite what I was trying to express. I was trying to express my thought that the voltage from the transformer wasn't particularly relevant as the 56k resistor provided the bias voltage - some sort of electrickery. I am very aware that the bias voltage is critical and can cause the amp to sound crap or melt the valves or give it the sweet sound depending on the current that flows through the cathode as a result of the grid bias. Before dismantling the donor I did sketch the circuit and the bias circuit does not have a doubler - the power valves were 7591As. My new build is using 6P3S-E power valves.

Is there any reason why the original circuits (e.g. 5F6A) had no means of adjusting bias?

I'm going to give the circuitlab gizmo a whirl and see what comes out.

Very many thanks

[paggerman]

Just a quick thanks - I've had a play, but gotta go now. I'll let you know how I get on.

[paggerman]

A very generic way of doing a bias supply is:
1). Rectify it.
2). Filter it.
3). Stick a voltage divider somewhere to achieve the proper bias voltage.

and optionally

4.) Filter it some more.

For some reason, guitar amp manufacturers like to use rather complicated voltage divider arrangements, and I suspect that's where you're having trouble understanding the circuit. For something relatively simple, there are more variations then you can poke a stick at.

Ah! That's what the 56k is for - a voltage divider. Got it. Thank you.

[paggerman]

You can draw the circuit and simulate it here https://www.circuitlab.com/editor/
Here is a sample (not sure if it is what you described,):

[ATTACH=CONFIG]20813[/ATTACH]

Set up a time domain simulation, probe the bias_out node.

[ATTACH=CONFIG]20814[/ATTACH]


..and here are the results.
[ATTACH=CONFIG]20815[/ATTACH]


Note the transformer is shown as an AC voltage source of 33.84 volts. That's because your transformer is 24V RMS and the peak voltage is 24 x 1.14 = 33.84V

Brilliant. Thank you very much. I love that circuit lab thing. I can now design a voltage doubler and voltage divider to give me the specified -48v.

[Amp Kat]

Brilliant. Thank you very much. I love that circuit lab thing. I can now design a voltage doubler and voltage divider to give me the specified -48v.

I think lots of people get fixed bias mixed up with adjustable bias because of the word "fixed". Even though the negative grid voltage is called fixed bias as opposed to cathode bias it can still be manipulated and changed in many ways. Some tubes bias up different because of the grid impedance so throwing different tubes to the bias circuit can have a different effect on voltage and current. If you used a tube like say the KT-88 you would find that you might not be able to bias it up because the negative grid voltage isn't high enough and you may run out of range. So you may need to make a voltage doubler to get enough voltage to bias it up to the dissipation wattage you want to use.

[J M Fahey]

Something important which nobody corrected:

I was trying to express my thought that the voltage from the transformer wasn't particularly relevant as the 56k resistor provided the bias voltage - some sort of electrickery.

Emphatically it is not so.
1) the bias winding provides the voltage. The 56K or whatever is a load resistor.
2) winding voltage *is* important.
You can only adjust "down" so the winding peak voltage must be at least the needed bias voltage (as in those ultra-crude ones where you have winding>diode>cap>R to ground) or substantially more, so you have room to move.
In typical Fender transformers the tap is around 50VAC, I have seen up to 60V in other amps.
3) of course, if you have only 24V available, you'll need to add a doubler.

[J M Fahey]

Something important which nobody corrected:

I was trying to express my thought that the voltage from the transformer wasn't particularly relevant as the 56k resistor provided the bias voltage - some sort of electrickery.

Emphatically it is not so.
1) the bias winding provides the voltage. The 56K or whatever is a load resistor.
2) winding voltage *is* important.
You can only adjust "down" so the winding peak voltage must be at least the needed bias voltage (as in those ultra-crude ones where you have winding>diode>cap>R to ground) or substantially more, so you have room to move.
In typical Fender transformers the tap is around 50VAC, I have seen up to 60V in other amps.
3) of course, if you have only 24V available, you'll need to add a doubler.

[paggerman]

Something important which nobody corrected:

Emphatically it is not so.
1) the bias winding provides the voltage. The 56K or whatever is a load resistor.
2) winding voltage *is* important.
You can only adjust "down" so the winding peak voltage must be at least the needed bias voltage (as in those ultra-crude ones where you have winding>diode>cap>R to ground) or substantially more, so you have room to move.
In typical Fender transformers the tap is around 50VAC, I have seen up to 60V in other amps.
3) of course, if you have only 24V available, you'll need to add a doubler.

hi JM. Yep, the a-ha moment came to me through exclamationmark's explanation. I already realised that my hypothesis was flawed as the the near absence of current through the resistor wouldn't produce a voltage. I couldn't figure out what the 56k resistor was doing - not knowing the source ac voltage (it doesn't get labelled on schematics for some reason) I wasn't able to understand what was going on. But thank you, as always, for your support.

[Chuck H]

Well, now... It was mentioned that the original amp used 7591 tubes. In that case it's entirely possible that the bias wasn't using a doubler rectification circuit. The 7591. to me anyway, seems like a sort of cross beween a big bottle type power tube and an EL84. They don't require much bias voltage as compared to 6v6, 6L6 or EL34 type tubes. I don't know the specs for the tubes your lanning on using. But that will be an important thing to know when setting up the circuit for this supply. It's entirely possible that you won't be able to achieve the needed voltage without a doubler circuit.

And, to put the 56K resistor issue in the simplest light. The 56k resistor is one half of the voltage divider circuit that determines how much of the supplied voltage will be on the power tube grid/s. Nothing more. The value could be 100k or 22k if a corresponding change was made to the other resistor in the divider circuit.