Fixed bias ltp question

[catalin gramada]

Considering a fixed bias source for a long tailed pair inverter I.m not pretty sure where to tie the return. I figured a sketch with possible points in tail node (A) or in stage ground (B). Can You help me please? Thanks.

[eschertron]

Why do you think there'd be any perceptible difference? How much does the voltage on top of that 100R resistor vary, and you'd see less than 1/3 of that at the grid? I remember reading that some of the design choices Fender made with LTP connections had to do with keeping existing eyelet boards and nothing to do with better/best sound.

[nickb]

Tie it to B.

I'm not seeing the merit of this approach. It uses two resistors and a big cap to do the job of one resistor and has less power supply hum rejection. What's the big idea?

[catalin gramada]

Tie it to B.

I'm not seeing the merit of this approach. It uses two resistors and a big cap to do the job of one resistor and has less power supply hum rejection. What's the big idea?

Thanks. Just to maximise the available gain of this stage in conditions of using unbalanced triode sections. Just a thought...

[nickb]

Thanks. Just to maximise the available gain of this stage in conditions of using unbalanced triode sections. Just a thought...

I don't think this approach makes any significant difference to the gain. It's the usual Gm. x Reff values that matter.

[Tony Bones]

Thanks. Just to maximise the available gain of this stage in conditions of using unbalanced triode sections. Just a thought...

A longer tail will make the circuit more tolerant of mismatched sections. One way to do that is to make a negative supply rail. If you're using a fullwave rectifier with CT (as opposed to a bridge rectifier) then another pair of diodes can be used to make a negative supply. Now you have roughly twice the total voltage to work with. Some of it can be used up with a large tail resistor to allow unmatched sections. The rest can be dropped over larger plate load resistors to get more voltage gain.

If you work everything out carefully you can put the grids at GND potential (as opposed to +50V as in the typical Fender LTP.) Thought there's only minor advantages to having the grids at 0V.

NFB would need to be applied only to the unused input in the LTP. But there's no need to bootstrap the tail anyway.

[J M Fahey]

Apply the KISS Principle

Just Keep It Simple

And IF you wanted to balance mismatched triode halves, you would need to add a little adjustment trimmer to each of them.
And even so, a 1k trimmer, with one end to each cathode and wiper to 22k resistor would do the job better, go figure.

[catalin gramada]

Thanks JMF, can You do a sketch, please? Thanks.

[J M Fahey]

Oh, too simple to need or merit a sketch.

Draw 2 triodes side by side ... separate cathodes instead of joined like in your drawing.

Join them with a horizontal 2k resistor.

Draw a wiper in the middle of that resistor so it becomes a pot.

Now connect the wiper to the top of the 22k tail resistor.

Presto.

Same as old shared 470 ohm self biasing resistor, but now you can adjust how much of it does each cathode get.

[catalin gramada]

Thank you