I am playing around with some 12As in a preamp and i have noticed something odd:
I use a LM317T as current regulator for each tube (two separate LM317Ts plus separate floating PSUs). The CCS are adjustable so I set each one for 5V measured at the valve pins.
However current through both valves is 297mA?? I experienced the same with a 71A I plugged in...
Input to the CCS is 12V. The LM317Ts have ~4V across them so they are not dropping out. The voltage across the sense resistor (10R in parallel with 390R and a 1k pot) is 2.88V. Both filaments measure 5V
What gives? The spec for both 12A and 71A is 5V at 250mA....?
I use a LM317T as current regulator for each tube (two separate LM317Ts plus separate floating PSUs). The CCS are adjustable so I set each one for 5V measured at the valve pins.
However current through both valves is 297mA?? I experienced the same with a 71A I plugged in...
Input to the CCS is 12V. The LM317Ts have ~4V across them so they are not dropping out. The voltage across the sense resistor (10R in parallel with 390R and a 1k pot) is 2.88V. Both filaments measure 5V
What gives? The spec for both 12A and 71A is 5V at 250mA....?
Most DHT”s are specced for constant voltage operation. Wheras for indirectly heated tubes constant current examples exist as well.
Your measured current is within the tolerance normaly found in DHT”s
It is claimed by some that constant current sounds better, i cant say anything personally about that claim as i have no real world experience. The supposed disadvantage of using linear regulators is that it short-circuits the AC component on the filament.
The coleman regs are temperature compensated current sources wheras the modules by tent are voltage modulated current sources. I have build some modules using the latter topology, and they seem to operate as expected.
Your measured current is within the tolerance normaly found in DHT”s
It is claimed by some that constant current sounds better, i cant say anything personally about that claim as i have no real world experience. The supposed disadvantage of using linear regulators is that it short-circuits the AC component on the filament.
The coleman regs are temperature compensated current sources wheras the modules by tent are voltage modulated current sources. I have build some modules using the latter topology, and they seem to operate as expected.
+20% is a little high, but I do not consider it a fault.
The data sheet value for Filament Current is only a guide - it's not a guaranteed parameter. The value is very often ±10% away from the nominal. So, I don't think it's a worry.
Some DHTs, like the WE101D and other WEs, and the STC 3A/109 are specified for exact current. In this case, the current must be set carefully to the nominal value, and the voltage can fall wherever it likes (again, within 10%, usually).
Other check;
- have you added a hum pot, or a pair of resistors across the filament, for anode-current return? Remember to account for the current into them.
The data sheet value for Filament Current is only a guide - it's not a guaranteed parameter. The value is very often ±10% away from the nominal. So, I don't think it's a worry.
Some DHTs, like the WE101D and other WEs, and the STC 3A/109 are specified for exact current. In this case, the current must be set carefully to the nominal value, and the voltage can fall wherever it likes (again, within 10%, usually).
Other check;
- have you added a hum pot, or a pair of resistors across the filament, for anode-current return? Remember to account for the current into them.
Two 47Rs across filament pins, center of which connected to cathode resistor and bypass cap. Anode current is only 5mA. Both tubes drawing ~295mA. Exactly 5V across filament pins.
Always thought these old DHTs would be a bugger to match. But..0.05dB difference in gain and bias voltages 13.84V & 13.78 across 2k7.
Measured distortion is quite a bit worse than 6J5, 85 and 76s in same circuit - tons of intermods around each harmonic. I have a set of your regulators so perhaps they will improve that? Although filament feed should be pretty clean: 12Vac feeding 4700uF 》balanced 4R7 resistors 》 4700uF 》LM317T
Always thought these old DHTs would be a bugger to match. But..0.05dB difference in gain and bias voltages 13.84V & 13.78 across 2k7.
Measured distortion is quite a bit worse than 6J5, 85 and 76s in same circuit - tons of intermods around each harmonic. I have a set of your regulators so perhaps they will improve that? Although filament feed should be pretty clean: 12Vac feeding 4700uF 》balanced 4R7 resistors 》 4700uF 》LM317T
Two series-connected 47Ω Rs across the filament will draw about 53mA - you can subtract that from the 295mA, and everything looks normal!
Try the Regulators - the LM317 is not a very good current source.
Try the Regulators - the LM317 is not a very good current source.
Assuming the sense resistor is between the output and the ADJ terminal of the LM317, the voltage should be quite close to 2.5 V, between 2.4 V and 2.6 V if I remember well. So why is it 2.88 V, 15.2 % above nominal? Is the circuit different from what I assume it to be or is it oscillating?
That's the circuit. Is there a chance of oscillation...seems rather simple and 317s are supposed to be stable? Perhaps the IMD I am seeing is oscillation?
The circuit is simply not what I guessed; I thought there was 2.88 V between OUT and COM (a.k.a. ADJ), which would have been strange.
For the record, what I wrote in post 7 is all wrong, as the LM317's reference voltage is 1.25 V +/- 50 mV rather than 2.5 V +/- 100 mV.
One last question: so filament current is DC and regulated. So 5V across filaments is DC and stable. But my voltages across the 47R center tap resistors are not 100% equal. How does this affect my circuit? It's not an AC voltage - so no hum injection? Is the imbalance creating more anode current than necessary?
The imbalance is because a great portion of the anode current of the triode itself is circulating from one side of the filament to ground, one side of the filament has a greater potential than the other its like two triodes in paralell one with greater bias and one with lower bias