Fixed bias implementation

[vetmed]

I am working on a project using 46s and would like to try fixed bias. The power tranny  has a 55v tap so it would seem that half wave rectification would probably be simple to implement but have heard many pros and cons about its virtues. Another approach would be to use battery bias but again many people don't like batteries because of the risk of failure. My personal fave so far involves using a small 12v transformer and connecting its secondary to one of the 6 volt windings on the power transformer to get ~ 60 volts which can then be rectified to get the -30 volts or so I need. I'd like to know what people think of these various ideas, your opinions would be greatly appreciated.

Regards
     Robert Lees

[Paul Birkeland]

If you have a 55V bias tap, that's certainly the most convenient way to go.  How much bias voltage do you need?

[vetmed]

About -30 volts is what I need. My concern is that half wave rectification won't give a very smooth DC signal. You are after all using a rather coarse DC voltage on the signal grid. But maybe it does not have to be of the best quality?

Robert Lees

[Paul Birkeland]

The general practice is to use a large cap input filter (maybe 100uF at 160V), then a large resistor (10k?) into another cap.

The high resistance/high capacitance filter will obliterate the relatively large amount of ripple that you would have.

[vetmed]

OK I know V=IR and appreciate what you said about reducing ripple but how can anyone make anything but vague approximations if you don't know what current you are dealing with? I have been under the seemingly illusory impression that the grid is not supposed to draw any current in class A operation   Seriously, from the plate curves for a triode connected 46 you should be able to run it at Va=250, Vg=~-27 volts and Ia= ~35(ma) and stay within plate dissipation of 10 watts. Subject of course to one's interpretation of the available plate curves. What this might mean in terms of grid current in class A operation I have no idea, but I am willing to learn Thanks

Robert Lees

[Paul Birkeland]

Grid current is generally non-existent in many cases.  When you draw your loadline on the curves and you hit the 0V line, swinging beyond that will pull grid current. (Referred to as transfering from class A1 to class A2)

If you need to supply grid current with fixed bias, then grid chokes can come in handy, but I don't think you'll need this for your case!

-PB

[vetmed]

To make calculations using V=IR when I is close to zero; how to proceed? Would it be a good enough to assume that I~1ma? Thanks

Robert Lees

[Paul Birkeland]

I'd say 100 Microamps is a good place to start.

If the bias supply isn't regulated, then a pot will be used to allow control of the supply.  If we assume that you're using a 100K control pot across a 55V bias supply, that would be the predominant load (500uA).

-PB

[vetmed]

I have of course been looking at various circuits for this sort of bias supply, most seem to use a pot  and a fixed resistor in a voltage divider network. I'm not sure how the pot needs to be wired in this case, can someone elaborate? Thanks PB for your explanations, much appreciated.

Robert Lees

[Paul Birkeland]

If we assume that the 55V tap is used, and that the CT is grounded, your half-wave power supply will give about -77V unloaded, and maybe -76.7V with the 100K pot.


You said you want -30V, so it would be handy to add some resistance on either side of the 100K pot to give you better control of that voltage (so that if you turn the pot 1/8 of a rotation, it may change the bias voltage by one volt instead of 5).

I'd say use a 25K pot with a 20K resistor between the "top" of the pot and the negative supply, and a 30K resistors between the bottom of the pot and ground. 

This will give you -20V minimum bias voltage, -45V maximum bias voltage, with pretty fine control in between. Another decent option is to just use a 100K 10-turn bias pot.  These kind of cost a lot, but they work really well (I use them for hum pots in DHT amps).

[vetmed]

That's really great, thanks I have a 50K linear pot, how much difference will that make?

Robert Lees

[Paul Birkeland]

Not much, it will just give you more range of bias voltage.  If you double the other resistances, it will work out to be the same adjustment range.

-PB

[vetmed]

Very cool, thanks!

Robert Lees