Hello everyone...
So I'm wondering if there are any merits to heating filaments before applying (fixed bias)
I had an 813 PP amp, and the 813's made funny sounds if I didn't apply filament voltage , then wait for ~30 seconds, then apply the fixed bias, (then HT), so I've kept going with this order in my new DHT amp...but wondering if anyone has any info/experiences...
So I'm wondering if there are any merits to heating filaments before applying (fixed bias)
I had an 813 PP amp, and the 813's made funny sounds if I didn't apply filament voltage , then wait for ~30 seconds, then apply the fixed bias, (then HT), so I've kept going with this order in my new DHT amp...but wondering if anyone has any info/experiences...
Last edited:
Actually that's backwards. Since the plate current will start to flow as soon as the filament starts to heat up, you want the grid to control that current from the start. Unless you are intentionally withholding plate voltage until the grid bias is applied, the cathode bias should be applied at the same time you begin heating the filament.
I did that with one project I designed. That design uses DC couplings, and so you apply heater voltage first so the operating voltages will be correct when the HV is switched on (solid state diodes that hit max voltage within a second or two). Other than that, I don't know of any particular advantages there could be so far as preheating is concerned. I've done other projects where that wasn't necessary.
As for the "funny sounds", I don't know what's up with that. There are certain types (6AS7) that need preheating due to the high VHK rating, and the thicker than normal insulation between the filament wires and cathode sleeve that requires. Also, some of the TV damper diodes, but these are indirectly heated cathode types.
Push Pull:
Heat the filaments and apply negative grid bias before you apply B+.
With hot filaments, and correct grid bias, when you apply B+, the output transformer goes from 0 mA per side, to say 60mA per side.
If there are any unbalanced plate currents, you will get a click in the loudspeaker. Perhaps quite soft for push pull.
Single Ended: Do not do the above for Single Ended or it will be very loud, and might even blow out the tweeter.
0mA to 60mA creates a powerful super fast transient. The tweeter might live, but your ears will not like it.
Heat the filaments and apply negative grid bias before you apply B+.
With hot filaments, and correct grid bias, when you apply B+, the output transformer goes from 0 mA per side, to say 60mA per side.
If there are any unbalanced plate currents, you will get a click in the loudspeaker. Perhaps quite soft for push pull.
Single Ended: Do not do the above for Single Ended or it will be very loud, and might even blow out the tweeter.
0mA to 60mA creates a powerful super fast transient. The tweeter might live, but your ears will not like it.
DHTs can draw small anode currents when only the filament is powered. It's not harmful.
It's better to apply the grid bias at the same time as the filament supply, to be sure it is biased to the intended anode current before the anode supply is applied. Bias voltage of zero means a dangerously high DC current.
Sudden application of the anode supply (SS rectified supply) when the valve is already heated might be unkindly abrupt . I prefer a Cap multiplier for the anode supply - which is very effective for separating the capacitors used for the rectifiers from the DC Link capacitor for the power valve's local storage.
The capacitor multiplier can be configured for a 10 second rise time easily enough.
It's better to apply the grid bias at the same time as the filament supply, to be sure it is biased to the intended anode current before the anode supply is applied. Bias voltage of zero means a dangerously high DC current.
Sudden application of the anode supply (SS rectified supply) when the valve is already heated might be unkindly abrupt . I prefer a Cap multiplier for the anode supply - which is very effective for separating the capacitors used for the rectifiers from the DC Link capacitor for the power valve's local storage.
The capacitor multiplier can be configured for a 10 second rise time easily enough.
The original TSE and newer TSE-II amps both use solid state components for the filament and bias supplies. They arrive roughly simultaneously. The plate supply is rectified with a 5AR4 tube which provides a delay, then a slow ramp up of B+. The DHT's warm up faster than the 5AR4, so they are ready to eat before they are fed. This is the start up sequence that is the gentlest on the tubes. You must have a negative bias on the output tube's grid before its plate supply is applied, or it must ramp up faster than the plate supply does. In some pentode applications the screen grid voltage can be the last voltage to be ramped up, as most pentodes draw minimal plate current without screen voltage.
I will repeat my design of an intrinsic soft start B+, using a 5Y3 rectifier.
6.3V filament winding Must be a separate 6.3V winding, must Not be used to power the other tube filaments!
0.65 Ohm Series Power Resistor
5Y3 (it has a cold filament resistance of less than 0.3 Ohm).
So . . . about 4V is across the 0.65 resistor, and about 2V across the 5Y3 filament.
Later, when the 5Y3 filament warms up, there is 5V across the 5Y3 filament, and 1.3V across the 0.65 Ohm power resistor.
Simple. Inexpensive. Less parts than a complex soft start circuit (less parts to fail, take up real estate, require more connections, etc.)
Done!
6.3V filament winding Must be a separate 6.3V winding, must Not be used to power the other tube filaments!
0.65 Ohm Series Power Resistor
5Y3 (it has a cold filament resistance of less than 0.3 Ohm).
So . . . about 4V is across the 0.65 resistor, and about 2V across the 5Y3 filament.
Later, when the 5Y3 filament warms up, there is 5V across the 5Y3 filament, and 1.3V across the 0.65 Ohm power resistor.
Simple. Inexpensive. Less parts than a complex soft start circuit (less parts to fail, take up real estate, require more connections, etc.)
Done!
JMFahey,
There is nothing wrong with applying fixed bias from the beginning.
For years, I used 9V alkaline batteries in series to bias 300B and 2A3 tubes.
Always on, always working.
I do see potential problems with turning on B+ (a quick transient turn-on, not a soft start).
A well balanced push pull output stage should only have a minor output transient.
But if an input stage suddenly gets B+, then Bang! A large transient travels all the way to the output stage, and on to the speaker.
I no longer use battery bias; I use self bias for output stages.
My balanced amplifiers do not use bypass caps across the input stage self bias resistor, and no bypass cap across the output stage self bias resistor.
Although I use solid state rectifiers for B+, I most often use choke input filters, followed by a medium capacitance, followed by a series resistor, and then to a Very large capacitance to drive the output stage; another series resistor to another capacitor to power earlier stage(s).
That makes a medium slow rise of B+, so there is no transient to the loudspeaker.
There is nothing wrong with applying fixed bias from the beginning.
For years, I used 9V alkaline batteries in series to bias 300B and 2A3 tubes.
Always on, always working.
I do see potential problems with turning on B+ (a quick transient turn-on, not a soft start).
A well balanced push pull output stage should only have a minor output transient.
But if an input stage suddenly gets B+, then Bang! A large transient travels all the way to the output stage, and on to the speaker.
I no longer use battery bias; I use self bias for output stages.
My balanced amplifiers do not use bypass caps across the input stage self bias resistor, and no bypass cap across the output stage self bias resistor.
Although I use solid state rectifiers for B+, I most often use choke input filters, followed by a medium capacitance, followed by a series resistor, and then to a Very large capacitance to drive the output stage; another series resistor to another capacitor to power earlier stage(s).
That makes a medium slow rise of B+, so there is no transient to the loudspeaker.
Last edited: