DHTs and Cathode Bias

[6f6]

Sometime in the near future I will be attempting my first push-pull amp. I am considering using directly-heated triodes (#26 or #30) in traditional cascade-type r-c voltage amplification for the input and (maybe) driver stages. The filaments would be heated by a well-filtered and stiffly-regulated dc supply. I would like to use unbypassed cathode bias. My query:
Should connect the cathode bias resistor to the positive or negative end of the filament? Would it even make a difference either way?

[billr]

[i] the positive or negative end of the filament? Would it even make a difference either way? [/B]

hi, I would suggest neither, make your cathode bias resistor a pair, one connected to each side of the heater, and take the ground connection from the junction of the two resistors.

so for example, if your bias resistor is say 100ohms, nominally, use two 200 ohm resistors connected as shown above.

hope this helps

[Eusebius]

I've built about 6 DHT preamps as balanced differential pairs, and one as single ended. The diff pairs all had constant current sinks underneath - some taken from Morgan Jones book page 134, some off Allen Wright's website in his PP1 amp schematic. I never had a problem with microphonics in any of the six.

The SE amp was a nightmare, and I passed it to a friend to figure out because he has a SE amp. For a start, the unbypassed cathode hummed like crazy. Bypassing it with a decent cap was the first step to getting the thing remotely listenable. Then there were all sorts of other questions about making the whole enclosure totally dead and solid - any ringing just provoked the bloody thing - you'd touch the volume control and hear it. Blah.

So I went straight back to diff pairs, even for single ended inputs. In fact I bought a box of 100 3a5 tubes which are a double DHT and ridiculously cheap, like a couple of dollars. Otherwise I just use 4 DHTs in each line stage. OK - it's double the tubes, but it's half the trouble. Ground the second grid on the input and take the output from the first anode - at least you solve the microphonics question. I personally like to leave a choice of output - balanced or SE so people can use long cables if they like. My own sustem is all Push Pull balanced, because that solves the problems of both microphonics and power output. I use DHTs right through the 3 stages.

That's the way I see it a year down the line with DHT preamps.

[AndreasS]

Originally posted by 6f6
...My query:
Should connect the cathode bias resistor to the positive or negative end of the filament? Would it even make a difference either way?

For DC-heating the bias resistor should connected to the negative filament end.

Regards Andreas

[Dave Cigna]

Originally posted by 6f6
Should connect the cathode bias resistor to the positive or negative end of the filament? Would it even make a difference either way?

In a perfect world it shouldn't make a difference.

There are some people that will argue that balanced resistors (as described by billr) sound better. I think that enters into the black art of audio, and although I do not discount it in any way, it is something that only you can answer through experimentation.

From a more engineering perspective there is an advantage to grounding the the negative side of the filament. Consider the type 30 with its 2 volt filament. With the negative end grounded that end will (obviously) be at 0V while the other end will be at 2V. The average across the entire filament is 1V positive. That means you have 1V of bias for 'free' before adding a cathode resistor. The result is that the resistor you do end up with will be slightly smaller than what would be needed with AC heating. OTOH, if you connect the positive end to ground then the filament (on average) will be at -1V, so you'll need a larger cathode resistor to bias the tube properly. This might make a difference especially if you intend to leave the resistor unbypassed.

-- Dave