DHT cathode follower idea

[wa2ise]

Had this idea for a cathode follower using directly heated triodes (the smaller ones, like those used in battery operated portable radios or hearing aids. Ones with specified + and - filament terminals). Fed by large resistors from high voltage + and - to approx constant current sources and to look like a 1K load. If I wanted 1K follower load, and the filament wants 50ma at 1.4V, I'd need a 2K resistor fed by 100VDC supplies on both sides (+100V to 2K, to filament, to another 2K resistor to -100V). Well, that could be done, but would burn a fair amount of power (10W). Well, if I did that, would I get any "DHT magic" from this cathode follower?

[billr]

hi

you don't need to do that, just wire up the heater as normal, then, if say you wanted a 10k cathode resistor, get two 20k ones, connect one to each heater connection, join the other ends together, and hey presto. they are now in parallel to make 20k.

The amount of heater current going through 40k worth of resistance is virtually zilch.

hope that this is of use.

bill

[kevinkr]

Originally posted by billr
hi

you don't need to do that, just wire up the heater as normal, then, if say you wanted a 10k cathode resistor, get two 20k ones, connect one to each heater connection, join the other ends together, and hey presto. they are now in parallel to make 20k.

The amount of heater current going through 40k worth of resistance is virtually zilch.

hope that this is of use.

bill

I agree, just remember you do need a floating filament supply which cannot be shared with anything else, and should not have a lot of capacitance to ground. Don't forget to ground the other end of those two resistors either.. Alternately you can use 3 resistors, two of which get connected in series across the filament, third gets connected to the center point of the first two and ground. Use 2 x 100 ohm and 1 x 10K, again filament supply most float. (I tend to prefer this as it is symmetrical from an ac signal perspective/cathode bias perspective, but either approach should work fine.)

[tubelab.com]

I think that the point wa2ise is trying to make is that he also wants to use the +/- 100 volt supplies to light the filament as well as provide the cathode return path. There are tubes out there that have low current filaments.

The 1U4 can be triode wired and has a 50 mA filament. It could be used in his example, but I have no idea how this would sound since I haven't tried it. I experimented with the 1U4 and the 1R5 since I have a bunch of them. I came to the conclusion that most any conventional tube could kick their butt and these were best used in a low powered portable device, which is their intended application.

[kevinkr]

Originally posted by tubelab.com
I think that the point wa2ise is trying to make is that he also wants to use the +/- 100 volt supplies to light the filament as well as provide the cathode return path. There are tubes out there that have low current filaments.

The 1U4 can be triode wired and has a 50 mA filament. It could be used in his example, but I have no idea how this would sound since I haven't tried it. I experimented with the 1U4 and the 1R5 since I have a bunch of them. I came to the conclusion that most any conventional tube could kick their butt and these were best used in a low powered portable device, which is their intended application.

I guess if nothing else it is a (really not so) good way to burn almost 10W of power to provide 70mW of heater power.. LOL Frankly I'd use a battery.. (AA)

[wa2ise]

Originally posted by kevinkr


I guess if nothing else it is a (really not so) good way to burn almost 10W of power to provide 70mW of heater power.. LOL

I'd agree, that this is rather impractical to actually build. If I did build this, I'd probably use the battery, or constant current sources to create the illusion of high impedance as seen by this cathode follower. But for this post, I wanted to keep the question as simple as possible, and to see if anyone might be able to say if this concept would merit further work , or is just a silly idea and don't waste the time...