The contents of this post may make a few tube lovers cringe but I felt this was the best section to post for the information I seek.
I want to build a simple source follower headphone amp using a Sic JFET. This one to be exact https://www.mouser.com/ProductDetail/UnitedSiC/UJ3N065080K3S?qs=byeeYqUIh0OYOU7Cis0jMg== . Simulating this part with a 24V supply and a 42 ohm resistor as source load puts the source at 11.48V(roughly half the supply) for a bias current of 273ma. That resistor would have to dissipate a little over 3 watts. Now I could use a nice fat 10W resistor and be done. But what if I took a page from Pete Millett's Starving Student and used a tube's heaters as the source load. I have 15-20 used pull 12ax7 and -au7's lying around and no plans to use them for anything, so why not. I don't have to spend any money on 10W resistors and I'll have a little glow for late night listening 😛. Now if I use two tubes per channel in series with the individual tubes wired for 6.3V, I should end up with 12.6V / .6ma = 42 ohms. Which leads me to my question, will the FET see 12.6V on its source because that's what the tubes are designed to draw or will the source voltage be dictated by the resistance of the heater string(42 ohms)?
Side note:
1. To be clear the tubes will serve no other purpose than to provide a source load for the FET
2. I understand at 11.48V, I would be starving the heaters therefore reducing tube life. I plan on using 50+ year old used pulls of commonly available tubes so I don't think anybody should lose sleep over that fact. I won't.
I want to build a simple source follower headphone amp using a Sic JFET. This one to be exact https://www.mouser.com/ProductDetail/UnitedSiC/UJ3N065080K3S?qs=byeeYqUIh0OYOU7Cis0jMg== . Simulating this part with a 24V supply and a 42 ohm resistor as source load puts the source at 11.48V(roughly half the supply) for a bias current of 273ma. That resistor would have to dissipate a little over 3 watts. Now I could use a nice fat 10W resistor and be done. But what if I took a page from Pete Millett's Starving Student and used a tube's heaters as the source load. I have 15-20 used pull 12ax7 and -au7's lying around and no plans to use them for anything, so why not. I don't have to spend any money on 10W resistors and I'll have a little glow for late night listening 😛. Now if I use two tubes per channel in series with the individual tubes wired for 6.3V, I should end up with 12.6V / .6ma = 42 ohms. Which leads me to my question, will the FET see 12.6V on its source because that's what the tubes are designed to draw or will the source voltage be dictated by the resistance of the heater string(42 ohms)?
Side note:
1. To be clear the tubes will serve no other purpose than to provide a source load for the FET
2. I understand at 11.48V, I would be starving the heaters therefore reducing tube life. I plan on using 50+ year old used pulls of commonly available tubes so I don't think anybody should lose sleep over that fact. I won't.
It will be mainly the FET threshold voltage determining the source DC voltage, so it can vary all over the place.
By the way, for very low frequencies, the behaviour of heaters is somewhere in between a linear resistor and a current source. The higher the voltage, the hotter the heater and the higher its resistance, so the current increases less than proportionally with the voltage. Incandescent light bulbs have the same behaviour to a stronger degree, because they get even hotter.
By the way, for very low frequencies, the behaviour of heaters is somewhere in between a linear resistor and a current source. The higher the voltage, the hotter the heater and the higher its resistance, so the current increases less than proportionally with the voltage. Incandescent light bulbs have the same behaviour to a stronger degree, because they get even hotter.
Tube heater resistance has a nonlinear dependence on voltage across it. Not a good load for an audio device. Use a fixed resistor instead.
The thermal time constant is low enough that for 20Hz and above I think a heater (not a filament) is pretty much a resistor. But it will vary a bit with DC current.
But the Vgs(th) spec is crazy loose on that FET (-6 to -14V). If the gate is at 0V then the source could be anywhere from 6V to 14V. So the current could vary 2:1.
I would use a CCS...
Pete
But the Vgs(th) spec is crazy loose on that FET (-6 to -14V). If the gate is at 0V then the source could be anywhere from 6V to 14V. So the current could vary 2:1.
I would use a CCS...
Pete
Thanks for the input everybody. I thought it might be a fun idea. But it looks like I'd be asking for trouble with no benefit in return. Pete, I have a circuit drawn using this device with a ccs that performs very well... in simulation. In the real world, thermal runaway could be an issue. I'm planning on prototyping it but wanted to throw together something simpler first, so I can get some tunes going. I don't care much for the IC based headphone amps I have at hand. Ok, I will stop talking SS in the tube section now