Half-mu stage (or SRPP) w/6DJ8/6922/ECC88 tubes at Low B+ voltage?

[Mateo Jun]

I have a plan to build a low-voltage Aikido, however it bothers be that 6GM8 is obsolete and will be increasingly rare and rare over time. 6GM8 tubes are already too expensive, and with same amount of money, I can buy carefully selected 6DJ8 tubes.

There was a suspicion regarding possibility that 6GM8 and 6DJ8 could be identical tubes, because 6DJ8 works pretty good with low voltage. Some claimed it is common marketing behavior to mark same products with different labels with different purposes. But, 6DJ8 tubes are much more readily available.

A thing that bothers me is whether low voltage will be sufficient to give me the amount of voltage swing for the signal. Another issue is that some folks said the distortion was as twice with 6DJ8 than as with 6GM8 at low voltage.

For the latter, I found another article by Merlin Blencowe regarding ECCXX tubes working at low B+ voltage, and grid resistance could sink as low as ECC88 it is a mere 3.6kΩ at such low B+ voltage.

Triodes at low voltage

I suppose this is why people had more distortion with low-voltage working ECC88 than with 6GM8. Merlin Blencowe's suggestion is to employ 'pull-up gridleak biasing grid bias' than traditional 'grid-leak biasing' to maximize the input impedance.

My intention is to input 60 to 65Vdc B+ rail for a 6DJ8 Half-mu stage. Putting aside worries whether this voltage range will be sufficient for the following cathode follower in this discussion, do you think this makes sense to you?

All I need is capability to throw signal that swings RMS 3V (peak-to-peak -4.24 to +4.24) into very high input impedance. (Would be good, if I could see 7Vrms, tho)

[nazaroo]

Originally Posted by Mateo Jun

I have a plan to build a low-voltage Aikido, however it bothers be that 6GM8 is obsolete and will be increasingly rare and rare over time. 6GM8 tubes are already too expensive, and with same amount of money, I can buy carefully selected 6DJ8 tubes.

If your goal is to save money, and have reliable supplies,
switch to Russian 6N3P or 6N1P as others have.
The Chinese make cheap ones, and the Russians make good ones. ($2 each on the internet).

Don't bother with 6DJ8. Just choose a better power supply voltage.

There was a suspicion regarding possibility that 6GM8 and 6DJ8 could be identical tubes, because 6DJ8 works pretty good with low voltage. Some claimed it is common marketing behavior to mark same products with different labels with different purposes. But, 6DJ8 tubes are much more readily available.

Big mistake. These tubes aren't interchangeable.
The 6GM8 will short out and fry if you exceed its max Plate voltage (only 25 volts! - always use a 22 volt B+ supply to be safe).
And these tubes are expensive. Don't prove you can blow them up, unless you're a millionaire and don't care.

6DJ8 can operate with up to 130 volts B+. More than 5 times the voltage.
The reason is probably size and spacing of the internal elements.

The only thing kinda the same is the pinout, and the heater voltage/currents: 6.3 v (6DJ8 365 mA, 6GM8 330 mA).
The 6GM8 has an inter-section shield, thats a bit different than the RF shield in the 6DJ8.

Also the inter-capacitance (Miller etc.) is different:
Plate: GM - 1.8pf DJ - 1.4pf per section
Grid: GM - 3pf DJ - 3.3pf
Cath: GM 1.3pf DJ - 1.8pf

This Miller stuff interacts with the In/Out impedances:

Zout = 6DJ8 - 2k5 ohms typical
Zout = 6GM8 - ?

Which affects frequency response and oscillation issues.

A thing that bothers me is whether low voltage will be sufficient to give me the amount of voltage swing for the signal. Another issue is that some folks said the distortion was as twice with 6DJ8 than as with 6GM8 at low voltage.

You can't run a 6DJ8 with its plate much lower than 25v+
and hope for anything like a non-distorted result.
That would kill a 6GM8 tube.

On top of everything else, their Mu is different:

6GM8 Mu (Amp.Factor) = 14 (worse than a 12AU7 at 17).
6DJ8 Mu = 30 (at almost all normal operating conditions).

For the latter, I found another article by Merlin Blencowe regarding ECCXX tubes working at low B+ voltage, and grid resistance could sink as low as ECC88[?do u mean 86?] it is a mere 3.6kΩ at such low B+ voltage.

Triodes at low voltage

I suppose this is why people had more distortion with low-voltage working ECC88[6DJ8] than with 6GM8[ECC86]. Merlin Blencowe's suggestion is to employ 'pull-up gridleak biasing grid bias' than traditional 'grid-leak biasing' to maximize the input impedance.

The more complicated the powersupply,
the less money and effort you save,
and the less likely you'll get clean results...

My intention is to input 60 to 65Vdc B+ rail for a 6DJ8 Half-mu stage. Putting aside worries whether this voltage range will be sufficient for the following cathode follower in this discussion, do you think this makes sense to you?

No. It makes no sense.
Run any tube at near its ideal operating levels.
For a 6DJ8 that means providing at least 100 volts per tube.
If you really intend to stack them vertically in any unusual topology, double your voltage supply,
and make sure your power-up sequences don't put the tubes in jeopardy of too high a voltage or current while they warm up.

All I need is capability to throw signal that swings RMS 3V (peak-to-peak -4.24 to +4.24) into very high input impedance. (Would be good, if I could see 7Vrms, tho)

Make sure you always have plenty more swing headroom than you expect.
Music transients may be several times the average voltage/power you will measure.
This means that for brief moments your tube may be overvoltaged and in danger of internal arcing.

good luck.