I recently learned that a 6CA7 has a "user assignable" G3, which in most other power output tubes is internally connected to the cathode. The Tubecad page I recently visited shows a 6CA7 triode connection circuit where both G2 and G3 are connected to the plate via different value resistors. Is this optimal?
When browsing for full range speaker cabinet designs, I stumbled on a triode connection that used an amply bypassed string of 6.3V zener diodes (adding up to ~30V) for the G2 connection to plate - with the corresponding claim that this was the only way to do it properly. Most use resistors....
I tried that idea using a TL431 with some capacitive bypass, set to 27V. It works, with nothing on fire, and the tube pairs seem to DC bias more closely using the shared control as designed by the OEM. Sounds OK. Was using 500 Ohm resistors, but those didnt have anywhere near a 27V drop across them as I recall.
Let's say this constant voltage drop from the plate idea for G2 is actually really good. What would be as appropriate for G3 - since, as a 6CA7 user you're free to connect it as you wish? JB on Tubecad connects it through a 1K resistor to plate, while he connects G2 through 100 ohms. Thanks for any input / discussion!
When browsing for full range speaker cabinet designs, I stumbled on a triode connection that used an amply bypassed string of 6.3V zener diodes (adding up to ~30V) for the G2 connection to plate - with the corresponding claim that this was the only way to do it properly. Most use resistors....
I tried that idea using a TL431 with some capacitive bypass, set to 27V. It works, with nothing on fire, and the tube pairs seem to DC bias more closely using the shared control as designed by the OEM. Sounds OK. Was using 500 Ohm resistors, but those didnt have anywhere near a 27V drop across them as I recall.
Let's say this constant voltage drop from the plate idea for G2 is actually really good. What would be as appropriate for G3 - since, as a 6CA7 user you're free to connect it as you wish? JB on Tubecad connects it through a 1K resistor to plate, while he connects G2 through 100 ohms. Thanks for any input / discussion!
Last edited:
You properly already knew this. If someone down the line put in a 6L6GC or a KTxx just to see how it sound, there will be a short (if diode are used or the resistant of the G3-Plate resistor) between plate to cathode. Lets say plate voltage is 400V and that G3 to plate resistance is 1K, and Vk = 40V, that will be (400-40)/1000 or 360mA consuming 130W.
Some smart manufacturers of 6L6GC, 5881, KT66, and KT88 amps tie pins 1 and pin 8 together . . . just for the tube rolling enthusiast who puts in a 6CA7 or EL34 (without checking first).
#1. Watch out for a metal 6L6, whatever the cathode voltage is (pin8), the metal cover (pin 1) will be at that voltage. And some KT88 and KT66 manufacturers insist on connecting the metal shield to pin 1 (so you also get the cathode voltage on the metal shield. If it is the top tube in an SRPP output stage, watch out!
#2. Build a 6CA7 or an EL34 amp that ties the suppressor grid (pin 1) to the plate. Watch what happens when a tube roller puts in a metal 6L6, KT66, or KT88, and touches the metal shield.
I call #1 and #2 above the "surviving spouse syndrome".
#1. Watch out for a metal 6L6, whatever the cathode voltage is (pin8), the metal cover (pin 1) will be at that voltage. And some KT88 and KT66 manufacturers insist on connecting the metal shield to pin 1 (so you also get the cathode voltage on the metal shield. If it is the top tube in an SRPP output stage, watch out!
#2. Build a 6CA7 or an EL34 amp that ties the suppressor grid (pin 1) to the plate. Watch what happens when a tube roller puts in a metal 6L6, KT66, or KT88, and touches the metal shield.
I call #1 and #2 above the "surviving spouse syndrome".
> both G2 and G3 are connected to the plate
That would be Triode mode. Is that what you want??
The effect of each grid is less the further out from the cathode. G3 has very small effect but it deters secondary electrons and violent Tetrode action at low plate voltage.
For Pentode, G2 goes to a high voltage and G3 goes to a low voltage. All the small power pentodes expect you to tie G3 to cathode. Large deviation seems to have little effect until Vg3 is a large fraction of Vg2.
Pete Traynor (guitar amp pioneer) tried G3 to a significant negative voltage, under the theory that this would reduce the damage when his hot-rodded amplifiers were pushed LOUD. 40+ years later my impression is that the effect is insignificant.
And there is the Other Problem, different similar tubes using the same pin for different things, and other tubes potentially blowing-up the amplifier. I have tried 7027(6CA7) G3 via 470k to cathode, so if it got strapped to G2 by a different tube nothing died. It worked in quick bench-tests with 'right and wrong' tubes. I could not measure any significant voltage on the 470K (so G3 current is small, as we might expect for G3 low). But perhaps extended abuse would find a problem with that theory.
There are Transmitter tubes where G3 is optionally tied to +60V for some unclear reason. But these tubes and their uses are different from our little bottles.
That would be Triode mode. Is that what you want??
The effect of each grid is less the further out from the cathode. G3 has very small effect but it deters secondary electrons and violent Tetrode action at low plate voltage.
For Pentode, G2 goes to a high voltage and G3 goes to a low voltage. All the small power pentodes expect you to tie G3 to cathode. Large deviation seems to have little effect until Vg3 is a large fraction of Vg2.
Pete Traynor (guitar amp pioneer) tried G3 to a significant negative voltage, under the theory that this would reduce the damage when his hot-rodded amplifiers were pushed LOUD. 40+ years later my impression is that the effect is insignificant.
And there is the Other Problem, different similar tubes using the same pin for different things, and other tubes potentially blowing-up the amplifier. I have tried 7027(6CA7) G3 via 470k to cathode, so if it got strapped to G2 by a different tube nothing died. It worked in quick bench-tests with 'right and wrong' tubes. I could not measure any significant voltage on the 470K (so G3 current is small, as we might expect for G3 low). But perhaps extended abuse would find a problem with that theory.
There are Transmitter tubes where G3 is optionally tied to +60V for some unclear reason. But these tubes and their uses are different from our little bottles.
Yeah, Triode is what I'm doing with the 6CA7 based Sophia amp. Based on the discussion / information here, I'll leave the G3 as-designed by Sophia. I'll also remember to check for that pin 1 to 8 tie, next time I have the bottom cover off. Thanks!
Any comment on the ~30V solid state drop from the plate for the G2, for triode operation? I expect that's much different than a resistor and it would be interesting to know in what ways. (yeah, like how much worse...) The amp seems to sound OK this way, but, being a 62 yr old neophyte at this, these old ears can still learn a thing or two about what sounds "right".
Any comment on the ~30V solid state drop from the plate for the G2, for triode operation? I expect that's much different than a resistor and it would be interesting to know in what ways. (yeah, like how much worse...) The amp seems to sound OK this way, but, being a 62 yr old neophyte at this, these old ears can still learn a thing or two about what sounds "right".
After few tests with suppressor tied to anode, cathode and not tied at all, with no nfb compensation applied, found the only way it works with decent results from an audio test point, is with g3 tied to cathode. Diferences are effective and not something subtile. With g3 to anode found the bass freq out of control, at least with this tube and the speaker I used. Tied g3 to cathode in a classic triode mode way and get decent results.
So, if wanna get more tight low freq result, tie the suppressor to cathode as many other did (despite the fact is no reason to do it so...or maybe is...)
So, if wanna get more tight low freq result, tie the suppressor to cathode as many other did (despite the fact is no reason to do it so...or maybe is...)
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.