Adjustable bias in Roadster- the 22K resistor

Hi
I just registered to the board
Have previously read lots of stuff on the board.
I have a Roadster and seek help with the bias mod.

I located the 22K resistor, added in series 50K multiturn blue trimpot.
When measuring the leads of the resistors in series which go back into the PCB (exactly at the point they enter into PCB), I measure a large change in resistance value. From 17K min to around 35K and maybe higher.

Connected a probe which measures mA to one of the tubes.
The maximum change I get although I turn the trimpot many many turns is from around 27 mA max to 32mA.
Even removed the 22K resistor at a later stage, got same results.
High voltage is 450VDC.
Measured the amp with stock 22K resistor before the mod, it was 27 mA.

My son came over, di all the same things I did and got same results.
I am stumped

Any suggestions?
The amp is around 6 max 7 years old
Thanks in advance
Ephraim

I looked at the road king schematic, which is no doubt a little different.

Two things are at work: parallel resistances and potential division. The 3.3k with an 825 ohm in series will add up to about 4.1k. The resistors in parallel (22k, 82k, and maybe the EL34 resistor/pot) reduce their total values to be less than the lowest value resistor in the parallel circuit. The idea with the mod is to reduce that parallel resistance so the divider sends less negative voltage to that grid.

If the 82k and 22k are the only things in there, this is easy. What you could do it to replace the 22k resistor with a 10k pot and a 12k resistor. Solder the resistor to lug 1 of the pot and use it for safety and also to make the total max resistance 22k. When the pot's wiper shorts out the pot at the minimum setting, the maximum parallel resistance will be about 10k, which keeps it from going so low as to red plate the 6L6. You can try different values, but a stock 3 Channel Recto reduces down to 6.2k for EL34 mode. The Road King (and probably the Roadster) are quite similar to the 3 Channel for the bias circuit and I figure that's a good value to stay well above, since we can see what it does to a 6L6 in a stock amp.

In any case making the amp difficult to set above 70% max dissipation is probably a really good idea. I would still try to shoot for tubes in the middle range or higher. Mesa's range is the same as Ruby 33ma to 52ma PC. My experience has been that the colder tubes are harder to red plate in EL34 mode. Something above 42ma might keep accidents down.

Switching to EL34s is a matter of maxing out the pot before flipping the bias switch. You could probably even just set them with 6L6 mode or slightly increase the resistor they use on the bias switch (like that 22k you now have set aside) to make the EL34 capable of being used in Bold/Silicon with no issues and adjustable with the same pot.

Check my math if you need to, but I'm pretty sure it's good advice.

Thank you for your help.
Will try it out soon and update with the results.
CHEERS!!!!!!!!