That's actually a really good point. I hadn't thought of that. By making the end of the filament/cathode that's connected to the cathode resistor the "positive" end, it will be the end with higher emission as the |Vgk| is lower on this end. Interesting...
But why does this improve the sound (which I think is the fundamental assumption in this thread)?
Or is that topology chosen for different reasons? Is the chance of thermal run-away less in this configuration, for example?
Could you list one of those manuals you've found this info in? I'd like to dig into this further.
~Tom
Last edited:
Are you sure? I'm thinking about it and it doesn't seem that it would even anything out. The tube doesn't know if the filament is connected to ground or not. It just knows the relative voltages of its electrodes. Changing between connecting positive end vs negative end of filament supply to ground would only change the grid bias by Vh volts. Am I missing something here?
Probably because it changes the bias point of the tube. RDH4 covers the need to change biasing between cases of center-tapped filament transformers and DC.
Some designers like to return the anode-cathode current via the negative filament terminal only, for the simple reason that it sounds better that way.
My 300B heater adopts a resistor network that compensates for the filament voltage by offering a higher impedance to anode return supply on the filament +ve compared the filament -ve. See the resistors R8..R11 in the schematic (this is an outdated schematic BTW - for updated circuits & ideas see thread: http://www.diyaudio.com/forums/tubes-valves/151421-26-pre-amp-3.html
One thing Guido Tent (the Tentlabs DHT heat-module designer) correctly identifies: The principal benefit of current heating of DHTs is that you don't present a low ac-impedance across the filament. Perhaps the cathode current distribution across the filament is upset by low shunt impedances. But if anyone is in doubt about the truth of this - here is a test. Put a LT108x 3-term reg on a heatsink. Connect ADJ to GND, OUT to GND via 1R/5W, IN as the filament negative, +9V dc supply as F+ve - a rudimentary current heater for 300B. When you have enjoyed the sound for a while, connect a 1uF cap across the filament, be it as boutiquey as you have on hand. Now see how the sound is degraded - to ac-heat level and below.
Attachments
Hey Jeff,
Just browsing here. I have some SMP5 switch mode DC supplies at my house. You can run the 2A3 filaments in series, and then put outboard 5v regulators on the output. The SMP5 will deliver 6v @ 4A, so there is plenty of overhead. Just let me know if you are interested.
Blair
Just browsing here. I have some SMP5 switch mode DC supplies at my house. You can run the 2A3 filaments in series, and then put outboard 5v regulators on the output. The SMP5 will deliver 6v @ 4A, so there is plenty of overhead. Just let me know if you are interested.
Blair
my father used to work in electron-microscope design and one thing he would take from there into his dht amps was to make shore that the voltage was only the one you wished to use before the ccs.
they used to have some sort of nfb or something which I can't remember (I will try to contact one of his old work mates)
they would do this because any fluctuating trace would send the ccs out.
they used to have some sort of nfb or something which I can't remember (I will try to contact one of his old work mates)
they would do this because any fluctuating trace would send the ccs out.
Wouldn't this then connect the cathodes together via the path of the filament supply? How then would you account for the cathodes bring connected prior to the cathode resistor? I realize the cathodes are eventually "connected" at the ground, but this occurs after the cathode resistor and bypass caps.
Please 'splain more...
Jeff
- Status
- Not open for further replies.