Hello. I.m at the third project using el34 triode connected. I connected both g2 and g3 directly to the plate and get stable operation and enjoy the sound. Still seeing a lot of projects using small values as 100-220 ohm resistors into the grids. I just wondering and asking why ? Did you encounter oscillations connected the grids to the plate directly at the socket ? Usually we using grid stoppers and determine its values when oscillation problems appear, Do you mount grid stoppers "just in case of" and "did not hurt" or you encounter real problems connecting directly the screens to the plates (never happen to me) please ? Thanks
This is considered good practice to control ultrasonic behavior, and anticipates the possibility of RF pickup, tube variation, etc.
catalin gramada,
Normally, an oscilloscope or sound card is used to see if oscillations occur.
Consider how you will know when oscillations occur?
1. With an 8 Ohm non-inductive load resistor?
2. With brand X loudspeaker?
3. With brand Y loudspeaker?
4. With the fanciest and most expensive whiz-bang 25 foot speaker cables?
5. On the Left channel, but not on the Right channel . . . all because of how the wires are positioned and dressed differently?
6. Etc.?
Ultrasonic and RF oscillations can be very tricky, come and go, do transient bursting, etc.
Use grid stoppers
Do not use grid stoppers
Your Mileage May Vary
The only EL34 amplifier schematic I saw that tied EL34 G3 to the plate, was the Randall Amplifier group build project in the San Francisco Bay Area.
(Having nothing to do with Randall solid state guitar amplifiers).
Have Fun!
Normally, an oscilloscope or sound card is used to see if oscillations occur.
Consider how you will know when oscillations occur?
1. With an 8 Ohm non-inductive load resistor?
2. With brand X loudspeaker?
3. With brand Y loudspeaker?
4. With the fanciest and most expensive whiz-bang 25 foot speaker cables?
5. On the Left channel, but not on the Right channel . . . all because of how the wires are positioned and dressed differently?
6. Etc.?
Ultrasonic and RF oscillations can be very tricky, come and go, do transient bursting, etc.
Use grid stoppers
Do not use grid stoppers
Your Mileage May Vary
The only EL34 amplifier schematic I saw that tied EL34 G3 to the plate, was the Randall Amplifier group build project in the San Francisco Bay Area.
(Having nothing to do with Randall solid state guitar amplifiers).
Have Fun!
Last edited:
G2 resistors (triode connected pentode) can be a good precaution against overly high current into G2 - when Ua is low and Ia is high. G2 is often a fragile element of a pentode, and should be protected.
Self-oscillation is possible under some circumstances, but this is liable to be at VHF or UHF, so scope dispay traces may need experience to interpret.
Anyway, It's much easier to take a portable AM radio, tuned to a weak station, and hold it near the amp. Apply a large low-frequency signal to the amp input to test thoroughly - because oscillation is more likely at high gm (transcondctance), which usually occurs at peak anode current.
If there is any oscillation problem, the radio will squeal or make other stage noises.
Then, you can trace the problem or use cut & try methods: different stoppers, shorten the wiring, add decouling capacitors, etc...
Self-oscillation is possible under some circumstances, but this is liable to be at VHF or UHF, so scope dispay traces may need experience to interpret.
Anyway, It's much easier to take a portable AM radio, tuned to a weak station, and hold it near the amp. Apply a large low-frequency signal to the amp input to test thoroughly - because oscillation is more likely at high gm (transcondctance), which usually occurs at peak anode current.
If there is any oscillation problem, the radio will squeal or make other stage noises.
Then, you can trace the problem or use cut & try methods: different stoppers, shorten the wiring, add decouling capacitors, etc...
I do mostly for "a problem could come". In my recent experiments I had amplifiers that "worked" with dummy loads, but oscillated with speakers. My last experiment https://www.diyaudio.com/community/threads/beam-power-otl-shifting-bias-levels.403061/ oscillated so badly that my Fluke multimeters were going crazy (showing all the signals on the display at once, freezing measurements) until I place the 250R in series with each G2.
So yes, what 6A3 wrote: the amp can work under some conditions, but misbehave under other conditions.
So yes, what 6A3 wrote: the amp can work under some conditions, but misbehave under other conditions.
Thank you.
Hey Rod you last sentence made the whole point of my question. We solve a problem if we have a problem. But see in most design values from 100-220 ohm to even 1k used by default even for same tube in similar circuits. That meant some are prone to oscillation than other? How we apreciate if it really need it without symptoms? Grid stoppers are easy to install to solve problems if any. (The last used into a DC coupled CF replacing a 12bh7 with an ecc99 who squeal like hell just because the tube replacement need it).I didn't even thinking to install series grid resistors by default- just in case of. I much prefer to take other preemptive measures as shielding, screening, twisting and the best possible layout I can and then, and only then if not stable operation compensations.
There are circuits using stopper resistors even in the plates. I never seeing making a trend with this.
This is not a statement but my conclusions based on my circuits. Of course testing with nfb networks with different speakers, neon lamps, radios or cell phones in proximity of the tubes may show the opposite please.
Hey Rod you last sentence made the whole point of my question. We solve a problem if we have a problem. But see in most design values from 100-220 ohm to even 1k used by default even for same tube in similar circuits. That meant some are prone to oscillation than other? How we apreciate if it really need it without symptoms? Grid stoppers are easy to install to solve problems if any. (The last used into a DC coupled CF replacing a 12bh7 with an ecc99 who squeal like hell just because the tube replacement need it).I didn't even thinking to install series grid resistors by default- just in case of. I much prefer to take other preemptive measures as shielding, screening, twisting and the best possible layout I can and then, and only then if not stable operation compensations.
There are circuits using stopper resistors even in the plates. I never seeing making a trend with this.
This is not a statement but my conclusions based on my circuits. Of course testing with nfb networks with different speakers, neon lamps, radios or cell phones in proximity of the tubes may show the opposite please.
Last edited:
G2 oscillations can also be well suppressed by using a ferrite bead on the G2 cable. The beads work well, because G2 current is large, and EMC ferrite materials are designed to exert good control over wideband currents.
This may be a good solution if you do not want low frequency resistance in the G2 circuit (it can affect the audio response).
Or you can combine a bead and 100 ohms for a fairly non intrusive default precaution.
Either way, the AM radio tests are always worth running, since some transient bursts may not be immediately apparent.
This may be a good solution if you do not want low frequency resistance in the G2 circuit (it can affect the audio response).
Or you can combine a bead and 100 ohms for a fairly non intrusive default precaution.
Either way, the AM radio tests are always worth running, since some transient bursts may not be immediately apparent.
This is likely another case where higher peak current means higher gm (transconductance), and so higher RF gain - a prime cause of instability.
If it was indeed self-oscillation of the pentode, It's the speaker's reactive load and uneven impedance with frequency that often drives higher peak current, and in turn oscillation.
Thanks. It is my experiment to play with this connection and, to be honest I feel the best image I get with g3......unconnected at all. But that.s a very subjective appreciation unacceptable from any point.
When using a triode connection (only), G3 may be connected to the anode, but it should never float in any case.
Indeed higher gm playing a role: with the tubes biased very cold it "worked". As soon as it was increased, oscillation. It wereld indeed pentodes, with feedback to the drivers cathodes. Some pF of capacitance in parallel with the FB resistors solved the problem!
I accidentally left G3 floating, in both pentode and triode. In both cases power was reduced considerably, certainly sounded different. https://www.diyaudio.com/community/threads/problems-with-hybrid-circuit-partial-feedback.335036/
In triode mode G3 does not have its original purpose. When connected to cathode, it is practically transparent. When connected to anode, it increases the gm of the tube.
I come jn back with some considerations relative to my G3 connection to anode. First, at high volume level the amount of audible distortions is more substantial than classic connection of g3 to cathode. I installed 100 ohm grid stoppers in both g2 and g3 grids. The power dissipated onto suppressor g3 grid is 1.2-1.3W and 2.5W to g2. As the grid is not supposed to dissipate power I don't know if 1.2W put in danger the thin g3 wire or not. The measurements was done at max onset clipping level at 25W per tube. I think I.ll come in back with classical triode wiring.
Member
Joined 2009
Paid Member
I seem to remember a post somewhere and sometime ago that this resistor can be optimized to reduce distortion - I think it was by (you) artosalo but not sure. It may have been on an E180 rather than an EL34 though.
As Rod C has explained, a speaker is a reactive load. That is sometimes a problem with NFB amps.
But not an amp with no FB, that should run OK with most any speaker. But the CC resisters on
control grids may still be required with some very high G tubes. Two different kinds of oscillation problems.