A tube can run well past Pdiss for a short time, even seconds.
In sweep or switch mode, or heavy class-B or C work, you "have to" exceed the DC limit on peaks to make the most of your precious tube dollar. In A or light AB audio, you would rarely/never cross that line.
If you are testing with low duty cycle pulses then you can exceed Pdiss by quite a bit. Use 50/60Hz half-sines and the worst heat is only 10% of the time. This was THE way before cheap CPUs and power semiconductors to force the tube in testing.
Destroy *U-50s? Do you know how many German and Russian soldiers worked with these tubes in careless ignorance? I'm sure it is possible to kill one. I have the impression that it is not easy to destroy these tubes. Maybe that's why they were used in such numbers that they are still fairly available at very low prices? $5-$10? Why do the Russians call it "poltinnik/half-a-buc" ?
That short fuse attitude won´t attract many answers.
Curves show everything the tube can do, you choose how to.
Curves in general show "static" values which apply literally to "static" working conditions, such as, say, regulated power supplies, etc. but in an Audio amp the concept of duty cycle applies.
In general, thermal time constants are way longer than individual audio signal waveforms, which combined with metallic parts thermal mass means they will integrate dissipation/temperature.
Even in Class A amps, as long as they are playing some Music, instant dissipation will vary along the cycle, how´s that?
Yes, in the long term temperature will reach the integrated value, that does not contradict what I wrote above.
If anything, tube datasheets "should" include dissipation vs. pulse duration , the way power transistor ones do.
Maybe some do!!!! and I´m not aware of that!!!! (Sweep/RF tubes?)
Curves show everything the tube can do, you choose how to.
Curves in general show "static" values which apply literally to "static" working conditions, such as, say, regulated power supplies, etc. but in an Audio amp the concept of duty cycle applies.
In general, thermal time constants are way longer than individual audio signal waveforms, which combined with metallic parts thermal mass means they will integrate dissipation/temperature.
Even in Class A amps, as long as they are playing some Music, instant dissipation will vary along the cycle, how´s that?
Yes, in the long term temperature will reach the integrated value, that does not contradict what I wrote above.
If anything, tube datasheets "should" include dissipation vs. pulse duration , the way power transistor ones do.
Maybe some do!!!! and I´m not aware of that!!!! (Sweep/RF tubes?)
That doesn’t answer my question, just another —I know it better— lecturing post. These blow my Fast Fuses…
I am looking for a save test setup for now. Not Prosa about tube amp operation from tube amp gods…
You guys are so far above GND/EARTH that you can’t even answer an —easy— question? If you know everything —better— than just why can’t you answer my question as asked? I asked it because I couldn’t find the answer here while searching through the forum. There’s skilled folks here building Pentode Spice Models and nobody can shed some light on some basic Pentode testing values? I have to wait 6-8 weeks to get replacement tubes if I fry them…
MasterQ, help me fill the table below (pre-filled based on my limited knowledge):
*)GU-50, Russian / Ukrainian made, load-curves please see above, I want to test against these curves (prove, re-build, model, see on my screen, htf have they drawn them in the first place, what values did they use, real or interpolated, whatever the right term is) with my own testing equipment, specification please see above). I am specifically interested about setting Ig2max (screen) and Iamax (Anode) for safe setup!
Ufil = 12.6VDC
Ifil = 1A
Ug2 = 250V
Ig2max = 20mA
Ug1 = -20V ... -50V
Ig1max = 5mA
Ua = 0 ... 400V
Iamax = 100mA
if you cannot or doesn’t want to answer this question you don’t need to reply.
I am looking for a save test setup for now. Not Prosa about tube amp operation from tube amp gods…
You guys are so far above GND/EARTH that you can’t even answer an —easy— question? If you know everything —better— than just why can’t you answer my question as asked? I asked it because I couldn’t find the answer here while searching through the forum. There’s skilled folks here building Pentode Spice Models and nobody can shed some light on some basic Pentode testing values? I have to wait 6-8 weeks to get replacement tubes if I fry them…
MasterQ, help me fill the table below (pre-filled based on my limited knowledge):
*)GU-50, Russian / Ukrainian made, load-curves please see above, I want to test against these curves (prove, re-build, model, see on my screen, htf have they drawn them in the first place, what values did they use, real or interpolated, whatever the right term is) with my own testing equipment, specification please see above). I am specifically interested about setting Ig2max (screen) and Iamax (Anode) for safe setup!
Ufil = 12.6VDC
Ifil = 1A
Ug2 = 250V
Ig2max = 20mA
Ug1 = -20V ... -50V
Ig1max = 5mA
Ua = 0 ... 400V
Iamax = 100mA
if you cannot or doesn’t want to answer this question you don’t need to reply.
Last edited:
p.s.
To shed some light I build a tube tester, specs above, and programmed a test procedure as follows:
If you want, yes, I duty-cycle through load values since a total test loop (one Ug set, one curve) takes about 30secs, less than 100ms per test point (Ua: 0-400V, 400 test points per curve)
To shed some light I build a tube tester, specs above, and programmed a test procedure as follows:
- set filament, manual enable to heat-up before testing
- set Ug2 fix and max allowed values for Ig1, Ig2, Ia
- set first Ug1 (first curve) and ramp through Ua
- set Ua back to zero next Ug1 (next curve) and ramp through Ua
- set Ua back to zero last Ug1 (last curve) and ramp through Ua
- all back to zero and manual switch off filament
If you want, yes, I duty-cycle through load values since a total test loop (one Ug set, one curve) takes about 30secs, less than 100ms per test point (Ua: 0-400V, 400 test points per curve)