I am pretty sure someone already did this before. For low power class A SE resistor loaded amplifiers - could one use the cathode heater winding instead of the resistor?
For example, EL34 needs 6.3 V for 1.5 A heater current. Two in series could be used with 24 V power supply and 1.5 A bias current. The main point would be that the tubes would dissipate the heat naturally and there would be no need for a high power resistor and a heatsink. The bonus would be the looks of glowing tubes.
EL34 is a common tube and if you have a friend guitarist with a Marshall amp, you could just use the tubes that would go to trash after re-tubing, because the heater winding is usually still fine even when the tubes are worn out.
Is this something that would work? The intended use of this amp would be for horn loaded compression drivers, maybe even just a MOSFET follower at the output of the DSP. No more than a few Watts is needed there.
Is that idea worth trying or is it a complete nonsense?
For example, EL34 needs 6.3 V for 1.5 A heater current. Two in series could be used with 24 V power supply and 1.5 A bias current. The main point would be that the tubes would dissipate the heat naturally and there would be no need for a high power resistor and a heatsink. The bonus would be the looks of glowing tubes.
EL34 is a common tube and if you have a friend guitarist with a Marshall amp, you could just use the tubes that would go to trash after re-tubing, because the heater winding is usually still fine even when the tubes are worn out.
Is this something that would work? The intended use of this amp would be for horn loaded compression drivers, maybe even just a MOSFET follower at the output of the DSP. No more than a few Watts is needed there.
Is that idea worth trying or is it a complete nonsense?
The heaters start cold with little resistance. So essentially when doing a cold start they would present a dead short load.
In theory, one heated filament is the same as another. However it would need some experimentation.
What happens before the current stabilises though. You might need some over current limiting to protect the single ended output stage. Cool idea though to get a valve look 😎
So you want to use a highly nonlinear resistor to set the bias for the output tube(s)? That could work, but probably not very well. I'm reminded of a buddy of mine who "repaired" a TV by using a clothes iron as a power resistor until he could source the replacement resistor. It worked, but it wasn't a production solution.
As others have pointed out, the cold resistance of the filament is pretty low. This means your output tube will draw lots of plate current until that current heats up the filament of the "bias tube". That'll be really hard on the output tube. It would also likely cause the output transformer to saturate, which further stresses the output tube.
Power resistors are not that expensive... Use one. Or if you don't like power resistors design something with fixed bias.
Tom
As others have pointed out, the cold resistance of the filament is pretty low. This means your output tube will draw lots of plate current until that current heats up the filament of the "bias tube". That'll be really hard on the output tube. It would also likely cause the output transformer to saturate, which further stresses the output tube.
Power resistors are not that expensive... Use one. Or if you don't like power resistors design something with fixed bias.
Tom
You could get that just by lighting the filaments though. No need to destroy the rest of the circuit during startup. 🙂
Tom
True... but think of the power wasted. I think Nelson has an amp using a bulb as load resistor for a single ended output stage and so using heaters kind of seemed a cool idea if the values work out.
No output tubes at all, he´s talking MosFets.
He uses dead/damaged tube filaments as resistive loads.
Using highly nonlinear red hot tungsten wire as Audio load is horrible.
Oh, I see. So a Class A, resistor biased, single-ended MOS output stage with a tube filament instead of the resistor? That MOSFET will go poof as soon as the power is turned on due to the cold resistance of the filaments. Even if by some miracle it doesn't the nonlinearity of the filament resistance will wreck performance.
I suppose one could design a circuit that pre-heats the filaments until they reach their nominal 'hot' resistance and then switches over to the MOSFET before the filaments have a chance to cool off. A simple relay with a delay circuit would probably work well enough. Or one could get fancy and measure the current through the filaments and switch over once it reaches 95% of the steady-state value. That starts to approach the infamous 10 lb of sh** in a 5-lb bag pretty quickly though and still leaves you with a highly nonlinear load resistor. But hey... Maybe the added nonlinearity adds something-something to the sound. It'd be quick to try in a prototype.
Another thing to consider is that the resistance of the filament/heater is not tightly controlled from tube to tube, so the bias of that MOS output stage may vary all over the place depending on which tubes are stuffed into the bias circuit.
Tom
I suppose one could design a circuit that pre-heats the filaments until they reach their nominal 'hot' resistance and then switches over to the MOSFET before the filaments have a chance to cool off. A simple relay with a delay circuit would probably work well enough. Or one could get fancy and measure the current through the filaments and switch over once it reaches 95% of the steady-state value. That starts to approach the infamous 10 lb of sh** in a 5-lb bag pretty quickly though and still leaves you with a highly nonlinear load resistor. But hey... Maybe the added nonlinearity adds something-something to the sound. It'd be quick to try in a prototype.
Another thing to consider is that the resistance of the filament/heater is not tightly controlled from tube to tube, so the bias of that MOS output stage may vary all over the place depending on which tubes are stuffed into the bias circuit.
Tom
I built a MOSFET SE amp loaded with 300W light bulbs... Works very nicely but I bias the MOSFET with bias voltage (it's a source follower so bias is easy even though I use depletion mode parts).
Tube heaters are different from light bulb filaments....one is almost immediately white hot, and the other warms up to red heat.
I agree, use a load resistor, 10 to 50 cent part here in India for up to 5W rating. And light the filaments separately if you want that retro look.
I agree, use a load resistor, 10 to 50 cent part here in India for up to 5W rating. And light the filaments separately if you want that retro look.
Why the MOSFET will.be damaged/destroyed?. There are MOSFET's capable of sustaining such currents and much more. For instance I remember STP80N05 for 80A and 50V.
There is at least one validated design that uses tubes' filaments as SE MOSFET source follower loads.
I knew I saw this idea somewhere! So in theory, it should work. I will put it on my list of things to try🙂
You forget Power
Current alone does not tell the picture. imagine same voltage (if supply holds) but 5X to 10X the current.
As pointed out further up. Tubes (ie cathode to anode) vary quite a bit in manufacture. I would suspect the distribution of heater resistances would be even wider. This is before we get to any in operation linearity or other discussions! The same is true for matching two light bulbs used in two oscillators.
Thinking about this more - I probably wouldn't. It may be better having some old 1920s steampunk lightbulbs for effect but leave the real work to resistors.
Thinking about this more - I probably wouldn't. It may be better having some old 1920s steampunk lightbulbs for effect but leave the real work to resistors.
I just measured the resistance of a couple of old Tesla EL34 at room temperature and both are about 1.3 to 1.4Ohms. At 6V it is about 4Amper and some coins. So the cold temperature isn't an issue for any suitably chosen MOSFET unless you are using a 2N7000.
Oh, sure. But only for a very short while. Measured in µs. Not on the time scale required to heat up a tube filament.
Definitely check the SOA curves of the MOSFET you choose for this project.
Tom