6K6 SE - max grid resistance question

[mxrshiver]

just a simple sanity check. building a guitar amp with 3x 6K6 in parallel single ended, cathode biased. stated max grid circuit resistance is 500K, so i'm using a 150K grid leak, figuring that will give me enough wiggle room to use some decent sized grid stoppers to keep the blocking distortion down.

i'm trying to figure out the max value of grid stopper i can use before exceeding the max grid circuit resistance rating of 500K per tube... and i'm using individual grid stoppers for each tube, since i'm experimenting with what (if any) effect using dissimilar grid stoppers may have in this configuration. so, 3x 500K in parallel is 167K, so that's the max total grid circuit resistance i can have for the entire power amp. given my grid leak is 150K, that leaves 17K left over available for all the grid stoppers.

i just wanted to check - that means the maximum grid stopper i should put on any tube should be 51K, correct? because 3x 51K in parallel is 17K? i just wanted to check, because no matter the math, my tired brain keeps seeing the schematic drawn that way and going '150K + 51K is 201K, that's way higher than the 167K limit'...

[tubeswell]

Better to use separate 47k grid stoppers soldered directly at each 6K6 control grid pin - this will help reduce unwanted oscillation.
For that matter, you can have 3 separate 470k grid leak resistors (albeit these will be in parallel from the P.O.V. of impedance load)
For a driver stage, you'll want a fairly high-current preamp tube like a 12AU7 that is set up with a 30k (maximum) plate load - to get an optimal 1:5 impedance bridge with the three parallel 470k grid leak resistors.


The OT/speaker will want to reflect about 3k-4k at the primary winding for 3 parallel 6K6s. YMMV

[Merlin]

i just wanted to check - that means the maximum grid stopper i should put on any tube should be 51K, correct? because 3x 51K in parallel is 17K?

Your math is correct.

[mxrshiver]

thanks so much to both of you!! i just have one question:

For a driver stage, you'll want a fairly high-current preamp tube like a 12AU7 that is set up with a 30k (maximum) plate load - to get an optimal 1:5 impedance bridge with the three parallel 470k grid leak resistors.

this is something i'm unfamiliar with! i've been mostly thinking in terms of creating enough output voltage from the driver to push the 6K6's well into overdrive. currently i'm using a 12BZ7 with a 150K plate load with 240V supply. the AC loadline ends up around 70K, but i'm still getting around 140V maximum output.

i think i understand impedance bridging in theory - making the output impedance of the driver stage small in comparison with the impedance of the following stage results in maximum signal transfer. and i'm sure if i used a 30K plate resistor, the 150K grid leak of the 6K6's would hardly change the total AC load, so i can see the advantage in that respect.

but i'm worried with that small of a plate resistor, i'd be limiting the output swing, and thus limiting the max extent to which i can push the 6K6's into overdrive. i know it'll only take maybe 40 or 50V to overdrive them and that even with the 30K plate resistor i should still get at least 100V output from that 12BZ7, but this amp is centered around the power amp overdrive, and i really wanna be able to push those 6K6's until they're spitting out damn close to a square wave at maximum drive.

i'm also worried i'd be decreasing the small signal gain so much that i'd be losing touch sensitivity and increasing headroom more than i'd like, especially with a higher current (and thus lower gain) driver.

could i ask you to go into a little more detail on this one? where does this ideal impedance bridging ratio come from, and what aims does it serve? would it have any of the effects i'm concerned with?

[tubeswell]

A 12AU7 is designed to run with plate load at 10k to 20k and a colder bias (up to 8v ish with a 10k plate resistor*), which means you can (in theory) drive the 12AU7 grid with (say) 16v* peak to peak input signal. The output voltage swing will still be ‘high’ in this configuration - albeit the gain between the grid and the plate is lowish.

*numbers off the top of my head for illustrative purposes - run a load line and see