Quote:
Originally Posted by blippster /img/forum/go_quote.gif
Someone please correct me if I'm wrong, but I don't think the tubes would be too affected by running them on 12V heaters, since effectively that's just dissipating less power in the heaters. The extra 7V drop over the MOSFETs is balanced out by a drop in the bias current from 150mA to ~105mA, so the amp still runs much cooler, and I'm not a fan of huge bias currents anyway. Perhaps that may (or not) compromise sonically, but (imo) 105mA is still plenty, especially with comparatively high-Z Sennheisers.
|
Well, from a tube life perspective, running heaters too low is probably not ideal, but this circuit runs tubes so gently that it surely does not matter here.
There is some controversy regarding starving the heaters --
with directly heated tubes (2A3, 300B, etc) it seems to lower the distortion slightly. Some people think this is a good thing to do, others not so much. Other people claim this is the case with indirectly heated tubes, too, though I've not seen any research to back that up.
It probably does not matter here because the low voltage and current across the tube likely lead to a high 2nd harmonic component in this amp. However, too low of a heater voltage and the tube won't work properly -- probably can't drive the fet's gate capacitance which could lead to treble roll off.
Quote:
Originally Posted by blippster /img/forum/go_quote.gif
Just measured with the DMM. I get a reading of 13.7V across the heater pins with the 17EW8, and ~13V for the 12AU7. Looks like they were being run a bit hard.
|
The difference in current across the FET would account for the extra .7V on the lower current tube.
Quote:
Originally Posted by the_equalizer /img/forum/go_quote.gif
But what seems interesting to me from your post is that it's the MOSFET gate bias (the gate to source voltage) what determines the MOSFET source voltage
|
Yes,
Quote:
Originally Posted by the_equalizer /img/forum/go_quote.gif
You see, what I'd thought would happen was that the MOSFET source would sit at ~18 volts (the 17EW8 tube heater voltage) and, with R2, R8 = 390 K the MOSFET gate would sit at ~17 volts; so the MOSFET wouldn't turn on.
|
No
The Mosfet's source will be ~4.5V below the gate. There is a relationship between Vgs and Ids (current from drain to source). This is what allows a mosfet to be used as a CCS -- set that voltage and the fet will only allow the corresponding amount of current to flow across it.
Quote:
Originally Posted by the_equalizer /img/forum/go_quote.gif
So now I see that tube heaters do not behave like constant voltage sources, but rather like constant current sources!
|
No, I think something is wrong with your measurements -- heaters are resistors. Like all resistors, their resistance changes with heat -- happens that they change a lot more than a standard resistor. But, the current is simply determined by V=IR.
Quote:
Originally Posted by timmyGCSE /img/forum/go_quote.gif
ah ok well nothing a bit of active cooling wouldn't sort out
|
Quote:
Originally Posted by tomb /img/forum/go_quote.gif
Yep. If a 6.3V heater would work, we'd have gone with 6J6's, but it's an insurmountable problem with the present design, IMHO (Dsavitsk's too).
|
You can certainly run 500mA through one of these fets, and the more current generally the better. But, I do think that IRF610's perform very well for headphone use at ~100-120mA and any more is unnecessary. And, you will have not only the increased current to deal with, but a much larger voltage drop -- 40V at 350mA or so is not trivial -- ~15W per side. So, not only would you need a really beefed up PS, but you'd need a ton of heatsinking. It could be done, but you could build a better amp for less effort other ways.