I'm starting a build a tube headphone amp from diyaudioprojects.com (the simple one: http://diyaudioprojects.com/Solid/12AU7-IRF510-LM317-Headamp/)

The projects uses a 12V supply for the tube plate voltage and states it gets decent result, although it should be run at higher voltages for better results. Considering I will build my own linear power supply for this project, I was thinking about using a higher plate voltage (say 24V, maybe more, depending on how big/expensive the input transformer will get).

Now, I assume some component values need changing, but I'm not clear on what and how to change those. I'm looking at P1, R2, R4...


A modest increase in R2 should suffice to re-bias the tube stage for 12V at the anode, P1 should be OK. Either recalculate using the load line method, or increase R2 until Va=12V, which will probably be around 200 ohms.

The output stage will still "work" as-is, but 125mA is quite a high current, so dissipating twice the power here will probably require heatsinking both FET and regulator for reliable operation.

Naturally you can't run the heater off 24V! So you still need to arrange a 12.6V or 6.3V supply for that.

One alternative would be to run both the heater and existing output stage off a regulated 12.6V supply (avoiding the increased power dissipation, though I would suggest heatsinking anyway), and a relatively low current 24V anode (B+) supply maybe using a voltage doubler then a regulator.

However this would mean you could no longer DC couple from V1 anode to Q1 grid cough sorry, gate. I would suggest AC coupling to Q1g, and providing the DC bias with a 1 Megohm "grid leak" resistor to a 6V supply (just a voltage divider with decoupling).


simulate this circuit – Schematic created using CircuitLab

  • \$\begingroup\$ Thanks for detailed answer! I missed the fact that I can't run the heater off 24V, which complicates things quite a bit (would either need to pull an extra regulated power rails from the supply or use a voltage doubler). At this point, it seems the extra complications are not worth it and I think I'll stick with the standard design for this. \$\endgroup\$
    – CatalinM
    Dec 6 '14 at 12:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.