Express Bias Circuit Question(s)

[surfsup]

I'm looking at this bias circuit and wanting to figure it out better. I understand V=IR but never building an express...

How much current runs through this circuit at idle? The full screen current of the power tubes? I am having a hard time figuring out how to get V estimates without knowing how much current is flowing. TIA

[surfsup]

I should add if I ignore current and treat the bias junction as a simple voltage divider (220k+15k on one side, 47k+25k on the other) the numbers don't add up. I get 59V for bias voltage as the junction of the 15/47 with the trimmer set midway using half of 585 as voltage.

[surfsup]

I built the circuit in PSUD and it seems I am not doing something right.

PSUD values are confusing me more. Are these plots the voltage drop across each resistor? anyone use psud?

[img:962:642]http://www.chicagocadcam.com/ChrisHahn/ ... G/PSUD.jpg[/img]

[martin manning]

Yes. You want V R3 for the bias voltage.

[surfsup]

I think i just figured out the math on this too. My issue is i dont have any 15k resistors. I think im good now, especially since i believe i know how to calculate the current, which will change as the trimmer is adjusted. Thanks martin.

[surfsup]

Looking at this more something is still not right.

The PSUD circuit gives me a range of 19-28V for bias and the TW Express schemo shows -30 as the setting. So I would assume the "range" to have 30 in the middle. The SPUD results show 30 is not even achievable.

Using my math, I get 33-47V bias.

I am using 285 VAC rectified to Vdc = Vac*0.707

I believe this is due to the actual circuit having the resistor before the diode. Or, is it due to current leak into the bias circuit from the screens due to the grid stopper 220/220 voltage divider on the power tube screens?

[martin manning]

First, you are saying screen, but I think you mean control grid. Pentodes and tetrodes have control grids and screen grids, but most people use "grid" to refer to the control grid, and "screen" to refer to the screen grid. The bias voltage is applied to the control grid. You can assume that the grid current is zero, and it is virtually zero at idle where you are setting the bias.

From the schematic you can see that with bias voltage of 33V and ~60k Ohms to ground, the average current flowing through the bias circuit is about 0.55 mA.

The voltage dropped across the 220k resistor, however, won't be anything close to this average current times 220k, and the difference is not just a simple conversion factor. The peak voltage achieved after the resistor and diode depends upon the peak current, which can easily be three or more times the average current, and this peak current will be a function of the time constant of the series resistance and the reservoir capacitance, and the equivalent load resistance downstream.

With some work you could calculate an expected voltage, or run a simulation, but the chances of hitting it right are pretty small unless you have all the relevant series resistances acurately represented and you know what the loaded voltage of the PT secondary is going to be. It is much easier to wire it up and then adjust the value of the 220k resistor to get the bias voltage close to the desired value with the bias pot set at mid range. There is plenty of it available from the HT to get what you need.

[surfsup]

martin as everyone knows here already, I am a glutton for punishment. I'm going to press forward to form a mathematical basis. Thanks, as always, for your help on the math/tech side of things.

Incidentally I did create a spice sim in TINA and it shows very similar to my math. I will look at the power tube graphs and try and figure this thing out...

[img:614:326]http://www.chicagocadcam.com/ChrisHahn/ ... G/TINA.jpg[/img]

[martin manning]

The math is known and available. Do you have Merlin's power supply book? There is an explanation and hand calc method there that can get you pretty close, probably close enough that the range of the bias pot would cover the error.

What do you mean by "power tube graphs," and in your sim, is the 281.25 rms voltage, and what value is the pot set at?

[surfsup]

I have the book and read that page. He states to find the relationship btwn Rs and Rl and consult a graph. I didnt understand it nor did i think it was mathematical. Probably due to my lack of knowledge but like many things as i learn more, the book will make more sense later.

The pot is 11k there is a p11k down to the left. Mid setting.

Not sure about voltage i will check later. In car on the road today w.o pc.

I found a book showing exactly how to do this type of calc based on the tube graphs. I will explain it better once i read thru it. Prob is the references are old and so are the graphs and symbols so im struggling to reproduce the method with today's data that i can find.

Anyway i'm trying...

[surfsup]

Martin, wouldn't this whole thing be easier if I just moved the resistor after the diode, and calculated the rectified voltage drop across it?

Why does it have to be in front of the diode on the AC side?

I am trying to make sure I have enough range for the -33V the Express needs with EL34s as well as -13V for a pair of 6SN7s

[martin manning]

You could swap the order of the resistor and the diode and it won't make any difference. The range you are looking for is pretty wide, but you might just get it with a 22k resistor and a 50k pot after the bias voltage node (I just traded the resistance between the pot and the 47k), but you might need to go to a 100k pot to be sure. The large resistor can be adjusted to move the range up or down.

[VacuumVoodoo]

You need to reverse the diode to get bias voltage negative relative to ground in the real circuit. For the math and sim it doesn't matter in this case.
Also Vdc=1.4xVac not 0.7.
What are R4,R5 and C3 doing?

The numbers you are getting are wrong. TINA works correctly buy the circuit is wrong:
- Transformer model? primary to secondary coupling coefficient? (In PSU designer you don't need it)
- Remove C3, at mains frequency it is a dead short. It makes R5 disappear as far as AC is concerned thus loading transformer with ca 0.6A.
- BTW what frequency did you set for the VS1? Is 125V defined as RMS or Peak?

But: you don't need the transformer in simulation of bias voltage rectifier. Just put a sine source 280Vrms/60Hz at the input (left side of R1)

[img:614:326]http://www.chicagocadcam.com/ChrisHahn/ ... G/TINA.jpg[/img]

[surfsup]

Alexander thanks for looking. I did figure out the problem with the three components because I forgot to put in the rectifier. Here is my current bias circuit in TINA. You can also see the voltage source/generator icons above it. There is an icon to pic a voltage source with a sine wave and using 175V as peak input (=125VRMS) I get weird results with that or with 125V peak:

[img:635:455]http://www.chicagocadcam.com/ChrisHahn/ ... /TINA2.jpg[/img]

[VacuumVoodoo]

The results are not weird. You are seeing AC ripple not DC values, except the 399V as it's peak value of AC voltage on the secondary. You need to attach DC voltage meters/probes to points of interest. Personally - I don't like TINA, too cluttered for my taste.
Are you sure the source is a sine? /the symbol is confusing me, all my sim SW shows one period of a sine wave inside the circle/
How long simulation run time did you set? The circuit with half wave rectifier will probably need to run a 30 second simulation for everything to stabilize.
The first capacitor is charging up through the 220k during 2 to maybe 4 ms, after that it is discharging through the rest of the circuit while also charging up the second capacitor when diode isn't conducting. This is repeated every period of the mains frequency.

BTW the 1N1183 diode you've put in the HV rectifier is a high power diode rated 35 A, perhaps a slight over kill at ca 10$/pc? 1N4007 is fine.