Hello!
Always connect G3 to cathode: however in small signal pentodes it won't make big difference.
See also here:
http://geek.scorpiorising.ca/GeeK_ZonE/index.php?topic=2345.0
Always connect G3 to cathode: however in small signal pentodes it won't make big difference.
See also here:
http://geek.scorpiorising.ca/GeeK_ZonE/index.php?topic=2345.0
Hi Alex,
Have you ever measured what you are talking about?
I just tried it with EL34. The current from g3 to plate connected through a 100 ohms resistor is in the lowish two-digit microAmpere range, so the voltage drop across a 100 ohms resistor producing the
Even a 1000-fold larger 100k resistor would drop only about a single volt or two.
You still claim that it
Welcome to Voodoo-Land
Tom
Have you ever measured what you are talking about?
I just tried it with EL34. The current from g3 to plate connected through a 100 ohms resistor is in the lowish two-digit microAmpere range, so the voltage drop across a 100 ohms resistor producing the
you claim is about a single milliVolt compared to direct g3-to-plate connection.
Even a 1000-fold larger 100k resistor would drop only about a single volt or two.
You still claim that it
?
Welcome to Voodoo-Land
Tom
This thread is going down a road once travelled. There was a similar thread a while back that was dealing with this subject and the EL-34. It was obvious then that there are some STRONG opinions on both sides. They are not always backed up by facts. In that thread I ran curves on a single sample of an EL-34. I came to the conclusion that the tube was more linear with G3 connected to G2, at least at low plate voltages. The results were disputed by some, and I had planned to repeat them at a later date, but never did.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67422&highlight=
I have EL-34 triode mode SE amplifier currently running (with JJ EL-34's) on my bench where I can easilly make the switch. I tried the G3 connection both ways. I can tell you that there is an audible difference, but it is not profound. I can't even tell you which I like best. I haven't ran any measurements yet. I am running about 420 volts across the tube, which may mask the effects seen in the curves.
The original question referred to the 6AU6. It is stated in the RCA tube manual to connect the plate, G2, G3, and the internal shield ALL together when triode connecting this tube. Many other tubes state that G3, should be connected to the cathode. So there is even some difference of opinion within RCA.
It would seem that there are two opinions on what to do with G3, and neither is invalid. So if you have the opportunity, try it both ways, and form your own opinion.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67422&highlight=
I have EL-34 triode mode SE amplifier currently running (with JJ EL-34's) on my bench where I can easilly make the switch. I tried the G3 connection both ways. I can tell you that there is an audible difference, but it is not profound. I can't even tell you which I like best. I haven't ran any measurements yet. I am running about 420 volts across the tube, which may mask the effects seen in the curves.
The original question referred to the 6AU6. It is stated in the RCA tube manual to connect the plate, G2, G3, and the internal shield ALL together when triode connecting this tube. Many other tubes state that G3, should be connected to the cathode. So there is even some difference of opinion within RCA.
It would seem that there are two opinions on what to do with G3, and neither is invalid. So if you have the opportunity, try it both ways, and form your own opinion.
Hi tubelab,
just to clarify, I was not talking about the difference between strapping g3 to cathode (this is quite evident for power pentodes and almost moot for small signal ones, and was mostly covered in the previous thread you mentioned, indeed), but the difference a small value resistor would make in the g3-to-plate connection compared to a simple wire.
Cheers,
Tom
just to clarify, I was not talking about the difference between strapping g3 to cathode (this is quite evident for power pentodes and almost moot for small signal ones, and was mostly covered in the previous thread you mentioned, indeed), but the difference a small value resistor would make in the g3-to-plate connection compared to a simple wire.
Cheers,
Tom
Tubes4e4 said:
Tom
One amp by someone not an expert hardly qualifies as proof of anything but I'll toss in my experience anyway. Take it for what it is, one person's experience with one amp.
I built a stereo EL84 amp running (IIRC) 350V B+. I put in a switch so that I could switch between pentode and triode. I initially just ran a wire from the switch to the plate.
Pentode mode worked just fine (but sounded like a pentode EL84
) but triode mode caused it to "warble". Not exactly like an ocillation but maybe it was. This occured at virtually any volume setting. Anyway it could not be run in triode mode.I put a 100 ohm resistor between G3 and the plate and all was well. It doesn't oscillate at any volume and seems perfectly stable.
So in my experience a 100R resistor between G3 and the plate does in fact make the amp more stable in triode mode.
(Maybe it was a poor amp design, maybe it had other faults, whatever, but the resistor made a difference.)
Maybe I was trying to answer everyone's question at once. I was trying to answer the original question (G3 to plate, or cathode) by saying that both are possible, and there are strong opinions on both sides.
As to the resistor, I used to use a direct wire from plate (or the mode switch) to the screen. I ran into a wierd oscillation problem with some EL-34's in UL mode where the tube would briefly oscillate during a very short period at the peak of the voltage swing (max plate voltage) when driven hard by a low frequency (100Hz) sine or triangle wave. I have since used a small (100 ohm) value resistor in series with the screen grid installed right at the socket pin.
In some Mullard literature a 1K resistor was recommended for UL use in push pull applications. I tried this with JJ tubes in SE UL and it just lowered the power a little bit, without making much difference in sound or distortion measurements. I did find out that the small metal film resistors do not live long in this application. They can't handle the peak currents that occur when the amp is driven to clipping.
The other side of my brain is RF-oriented. For those keeping track that leaves me with 50% of a brain for audio. Oh well…
Anyway, we sometimes used parasitic suppressors in the design of RF power amps. These are typically low value resistors in PARALLEL with a small air core inductor. These twin components are often placed in series with the plate/drain of the power device, before the tuned output circuits. At RF the choke becomes a high reactance and you have only the resistance in line, which acts to damp any unplanned resonances (parasitics). At low frequencies (DC in the case of RF designs) the resistor is shorted so that it doesn’t drop any DC voltage which would waste power.
Now for audio, you could place such a parasitic suppressor in the G2 or G3 lead to suppress oscillation. No, I haven’t needed to try this, since my power amp projects usually don’t have screens or suppressors to begin with. Use a carbon comp resistor, at least 1/4 watt and maybe 1/2 watt. Pick a value as discussed above. Solder an enameled magnet wire to one resistor lead close to the body. Using the resistor body as a form (essentially an air-core magnetically), wrap turns around it in a spiral and then solder the other side to the other lead close to the body. The exact values needed for the R and the small L are subject to experimentation, but should work over wide ranges for either part.
So what would we have achieved? At all audio frequencies and even well above, the inductor is nothing more than a short circuit. So, for audio, there is no resistor in line at all, just a wire. At VHF and UHF frequencies, the inductor approaches an open circuit since the inductive reactance becomes high. So, RF sees just a resistor, which damps any tendency to sing. It is possible under the right (wrong) circumstances for this little inductance to create a new resonance that could encourage oscillation, so experimentation is in order. There are too many variables among the parts, circuits and layouts to predict exactly. But if it can be made to work, it might be the best of both options rolled into one.
Anyway, we sometimes used parasitic suppressors in the design of RF power amps. These are typically low value resistors in PARALLEL with a small air core inductor. These twin components are often placed in series with the plate/drain of the power device, before the tuned output circuits. At RF the choke becomes a high reactance and you have only the resistance in line, which acts to damp any unplanned resonances (parasitics). At low frequencies (DC in the case of RF designs) the resistor is shorted so that it doesn’t drop any DC voltage which would waste power.
Now for audio, you could place such a parasitic suppressor in the G2 or G3 lead to suppress oscillation. No, I haven’t needed to try this, since my power amp projects usually don’t have screens or suppressors to begin with. Use a carbon comp resistor, at least 1/4 watt and maybe 1/2 watt. Pick a value as discussed above. Solder an enameled magnet wire to one resistor lead close to the body. Using the resistor body as a form (essentially an air-core magnetically), wrap turns around it in a spiral and then solder the other side to the other lead close to the body. The exact values needed for the R and the small L are subject to experimentation, but should work over wide ranges for either part.
So what would we have achieved? At all audio frequencies and even well above, the inductor is nothing more than a short circuit. So, for audio, there is no resistor in line at all, just a wire. At VHF and UHF frequencies, the inductor approaches an open circuit since the inductive reactance becomes high. So, RF sees just a resistor, which damps any tendency to sing. It is possible under the right (wrong) circumstances for this little inductance to create a new resonance that could encourage oscillation, so experimentation is in order. There are too many variables among the parts, circuits and layouts to predict exactly. But if it can be made to work, it might be the best of both options rolled into one.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.