Hi,
using low percentages of UL and flatter primary loads, the loadline is always well below the knee of the curves.
This means that the output stage could work with lower voltages on g2, so target the same power with less voltage swing from the phae inverter.
Is there any drawback in adding a zener diode in series with the UL connection, adding a resistor to ground in order to always ensure at least 2 mA through it (to avoid it switching on/off)? E.G. with 460V B+, 20% UL will swing approximately from 380 to 540 V. Adding in series a 47V 5W zener will make it swing from 330 to 510 V.
With a 220k 2W resistor to ground I'll be able to ensure 2 mA flowing through the zener.
Thanks
using low percentages of UL and flatter primary loads, the loadline is always well below the knee of the curves.
This means that the output stage could work with lower voltages on g2, so target the same power with less voltage swing from the phae inverter.
Is there any drawback in adding a zener diode in series with the UL connection, adding a resistor to ground in order to always ensure at least 2 mA through it (to avoid it switching on/off)? E.G. with 460V B+, 20% UL will swing approximately from 380 to 540 V. Adding in series a 47V 5W zener will make it swing from 330 to 510 V.
With a 220k 2W resistor to ground I'll be able to ensure 2 mA flowing through the zener.
Thanks
Not normally needed in UL mode.
I doubt it will increase or decrease anything except add heat from the zener; load resistor and G2 current draw.
I wouldn't.
I doubt it will increase or decrease anything except add heat from the zener; load resistor and G2 current draw.
I wouldn't.
Hi @JonSnell Electronic ,
thanks for your reply, but curves will follow new g2 voltages: they are obliged to change.
This is the way pentodes work.
I know that normally zeners are not used there, that's why I'm asking if there's anything I'm missing.
thanks for your reply, but curves will follow new g2 voltages: they are obliged to change.
This is the way pentodes work.
I know that normally zeners are not used there, that's why I'm asking if there's anything I'm missing.
It depends on the tube. For example, some sweep tubes do not like high G2 voltages, especially when the anode voltage goes below G2. In this case if you are set on UL then you need tricks like zeners.
Thanks @jcalvarez ,
I'm using standard audio tubes within accepted ranges for g2, but I would like to lower the voltages even more for the reasons explained above.
I'm using standard audio tubes within accepted ranges for g2, but I would like to lower the voltages even more for the reasons explained above.
I’m using resistors to bring the voltage down a bit. Usually 100 to 1000 Ohm will do. However for KT150’s in class A with 600 VDC plate voltage I use 1.2K for the screen grid (43% UL tap) to prevent orange glowing of the screen and still get good audio results.
I don't remember reading anything about that kind of usage.
I'm not the most informed on these things, but puzzling this out...
- Reducing the voltage to the screen grids will reduce the transconductance of the output pentode. That would raise rp a bit, which would 'flatten' the UL curves. Would you need to adjust the plate-plate primary impedance of your OPT to compensate? Or is that not the case and I'm not thinking about this correctly?
- UL curves look more like triode curves than pentode curves, so I'm not sure what 'below the knee' means in terms of UL operation of a 6L6 or whatever.
I'm not the most informed on these things, but puzzling this out...
- Reducing the voltage to the screen grids will reduce the transconductance of the output pentode. That would raise rp a bit, which would 'flatten' the UL curves. Would you need to adjust the plate-plate primary impedance of your OPT to compensate? Or is that not the case and I'm not thinking about this correctly?
- UL curves look more like triode curves than pentode curves, so I'm not sure what 'below the knee' means in terms of UL operation of a 6L6 or whatever.
@rongon
I don't know how the curves will react, because it is true that the transconductance will be lower, but the percentage of variation will be greater.
E.G. The UL tap will swing 400Vpp, this won't be affected by the zener in series, but if it swings from 300 to 700V or from 100 to 500V, in the former case the higher voltage is 2,3 times the lower, in the latter case the higher voltage is 5 times the lower. Will this make the curves steeper as if the UL had an higher percentage?
I talk about the knee because with 20% UL the curves have still a flattened knee.
I don't know how the curves will react, because it is true that the transconductance will be lower, but the percentage of variation will be greater.
E.G. The UL tap will swing 400Vpp, this won't be affected by the zener in series, but if it swings from 300 to 700V or from 100 to 500V, in the former case the higher voltage is 2,3 times the lower, in the latter case the higher voltage is 5 times the lower. Will this make the curves steeper as if the UL had an higher percentage?
I talk about the knee because with 20% UL the curves have still a flattened knee.
I will try the zeners in series without the resistors to keep current flowing through the zener even when the tube is off, then with the resistor.
I should not be necessary in a push-pull amp, because the high current side of the push-pull will dominate over the low current side, but I will check.
If anyone has tried it before, please share your experience.
I should not be necessary in a push-pull amp, because the high current side of the push-pull will dominate over the low current side, but I will check.
If anyone has tried it before, please share your experience.
I have used a zener, or a VR tube in series with the UL tap. I have also used a mosfet buffer fed by a resistive divider to provide for a variable UL tap to feed the screen. All experiments used a 6CD6GA tube because I had lots of them at the time and they did NOT like running in UL from the OPT tap. Works OK at full power output, but the tube fries at idle eventually leading to melted screen grid or red plate runaway due to gas and mobile ions from the glowing screen grid.
I came to the conclusion that pentode mode with cathode feedback with a fixed screen grid voltage gave similar power and distortion numbers, but didn't blow up. All experiments were done about 20 years ago so the details were lost. Resistor to ground to keep VR tube lit is needed due to difference in strike and run voltages.
I came to the conclusion that pentode mode with cathode feedback with a fixed screen grid voltage gave similar power and distortion numbers, but didn't blow up. All experiments were done about 20 years ago so the details were lost. Resistor to ground to keep VR tube lit is needed due to difference in strike and run voltages.
Attachments
Hello zintolo,
have a look here: https://rh-amps.blogspot.com/2013/02/rh84-amplifier-revision-2_26.html
have a look here: https://rh-amps.blogspot.com/2013/02/rh84-amplifier-revision-2_26.html
Thanks George, what’s the reason of this behaviour? Here I plan to lower screens around 50V.
@Dieter Geissel thanks for the li k, but that configuration is in single ended pentode mode with the zener used to lower the voltage on screen. Push-pull in UL is different.
I don’t guaranee that i have drawn the schematic quite right, went back &n forth with Bill a few times but the guys that know what they are doing can sort that, the zeners used may be NLA.
https://www.t-linespeakers.org/tubes/triode-trick.html
dave
https://www.t-linespeakers.org/tubes/triode-trick.html
dave
It took me about a year to figure this out. I was using 6W6GT vertical sweep tubes in an audio amp. The 6W6GT is rated for 10 watts of plate dissipation in "Class A amplifier service" with 300 volts on the plate and 150 on the screen. It is also specified for 7.5 watts of plate dissipation with 300 volts in triode connection or 7.0 watts on the plate with 300 / 150 in pentode connection for "Vertical Deflection Amplifier Service."
If I set the tube up with 300 volts on plate and 150 volts on the screen with the idle dissipation at about 5 watts you can beat it hard at 20 watts output per pair, crank the plate to 400+ volts for 30+ watts of output and it will live. Let it return to idle at 8 or even 9 watts and it will live forever as long as that screen grid stays below 150 volts. Best power VS THD comes with about 120 volts on the screen.
Let that same tube idle with even 200 volts in triode, and it will die a slow death, often taking several months or more. I got over a hundred 6DG6GT tubes cheap from Stan and built a rack to cook test about 6 tubes. It took a year, but all of them died. The 6DG6 IS a 6W6.
Previously I had built a hacked SSE board with 6LW6 sweep tubes wired as triodes. The first pair ran at 400 volts. It sounded great, made about 15 WPC, but one of the tubes went into red plate runaway within a month of daily use. I put in another tube and continued using the amp, but the other original tube died rather quickly. OK, the max screen voltage is 280, and I'm feeding it 400, but 50 volts are turned into heat in the cathode resistor. I swapped the power transformer and ran the next set of tubes at about 345 volts with about 300 volts between screen and cathode. They lasted longer but got gassy and died. My early conventional high powered SE Triode sweep tube amp experiments ended but eventually resulted in the UNSET design which runs at a fixed screen voltage that is usually about 150 to 175 volts with 30 to 50 volts on the cathode at idle. No failures have been seen.
It took a while for me to realize what was going on. Look at the specs for TV vertical sweep use with 300 volts on the plate. The tube eats 300 volts at 60 mA. That's 18 watts of power against a 7 watt dissipation spec. Why doesn't the tube melt or explode? Yes, 18 watts of DC power are being delivered to the tube but maybe 12 watts of 60 Hz (US NTSC) sawtooth waves are being delivered to the beam sweeping deflection yoke and only 6 watts are turned to heat in the plate. There is no volume knob on the vertical or horizontal sweep circuits. They run at full power output all of the time. If they didn't, your picture would shrink. If the vertical sweep had too much distortion, round things were not round on the TV screen (usually a failing cathode bypass cap). Too much 2H makes them egg shaped. Do we run our Hi Fi and guitar amps at 100% of max power all of the time? No, and that's what kills tubes in class A amps.
My dead 6W6 and 6LW6s were in SSE amps that ran them class A. The amp was on whenever my computer was on which was most of the time that I was home.
We know that most TV sweep tubes can be used for making audio amps, but the sweep related specs cannot be used. The plate dissipation spec IS useful, and often conservative depending on the amp design. The screen grid voltage rating MUST be respected, and I would venture to say that it is too generous. Keep on the low side of the limit by a considerable margin.
This leaves little margin for driving the screen grid, and rules out conventional UL designs. As previously stated simply adding a fixed offset with Zeners or VR tubes can impose too much voltage swing creating nonlinearity or even cutoff conditions in a class A SE amp. If I cranked the 6CD6 - VR tube thing in post 11 the VR tube would blink off and on resulting in severe distortion.
Thanks again George,
I will then try with the 6P6S I have available, running at 320V running at 33 mA on a 8 kOhm Raa.
I have a powerdrive on it, so it can go into AB2. I will implement also the other active feedback I developed for the single stage SE mosfet amp.
Anodes now swing from 50 to 590V approximately, while g2 from 250 to 390V.
g2 will then swing from 200 to 340V approximately, so I can use a 150 kOhm 2W resistor to keep 2 mA flowing through the zener when g2 are at 340V.
I will then try with the 6P6S I have available, running at 320V running at 33 mA on a 8 kOhm Raa.
I have a powerdrive on it, so it can go into AB2. I will implement also the other active feedback I developed for the single stage SE mosfet amp.
Anodes now swing from 50 to 590V approximately, while g2 from 250 to 390V.
g2 will then swing from 200 to 340V approximately, so I can use a 150 kOhm 2W resistor to keep 2 mA flowing through the zener when g2 are at 340V.
I was looking at those zener/shunt droppers for triode-mode pentodes as well. But what struck me was in the case of a pentode that can handle full plate voltage on the screen, it wasn't ever clear there was any actual benefit to it. And in the case of one that can't handle full plate voltage on the screen, your swing is limited by that lower screen voltage.
@Charles G
The main benefit in the case of a tube that can handle full anode voltage, with low UL percentage, is that you can use less negative bias to reach the same working point (so reduce the required swing to reach full power, so reduce the phase inverter distortion) without compromisimg the available power.
In my case, lowering g2 by 50V I can bias at around -17 instead of -23, so I need 34 Vpp to drive full power instead of 46.
It’s 26% less swing needed from the previous stage.
The main benefit in the case of a tube that can handle full anode voltage, with low UL percentage, is that you can use less negative bias to reach the same working point (so reduce the required swing to reach full power, so reduce the phase inverter distortion) without compromisimg the available power.
In my case, lowering g2 by 50V I can bias at around -17 instead of -23, so I need 34 Vpp to drive full power instead of 46.
It’s 26% less swing needed from the previous stage.