Triode based Allpass Filter

[5v333]

If you search for allpass filters, youll prob get a opamp based circuit.
i have been wondering if its possible to make a solidstate APfilter and more so with a triode.

here is a cathode follower where both grid and cathode serves as inputs.
if there was equal signal on both grid and cathode, there would be no alternating current action through the triode.
if we RC couple the grid and direct couple the carthode there will be phase and amplitude difference where the filter has its pole.
so now we have difference and the triode will deliver AC current over the freq selection.
the gain will be no more then about 0.95.

the series input resistor in the cathode can be scaled so it forms a 0.95 factor voltage divider with the load/cathode resistor. i think by doing so the signal injected at the cathode will match the gain of the grid injected signal and CMRR will improve. if this is of any importance i dont know...

triode could be 12au7, B+ 400V, 22K load resistor, 620R cathode resistor.

havnt tried or simulated it.

why would anybody wanna do a SS APfilter?
it could be used for making a discrete CAPS network for ex.

your thoughts!?

 

[abbey road d enfer]

I don't see how that would work, since an APF needs an inverter, and your circuit uses a cath-follower, which is essentially a non-inverting circuit.
Triode-based APF's have existed since forever; you may want to check Vox's AC30 vibrato circuit.

 

[PRR]

Magnatone. Split-load cathodyne. For a fixed frequency the unobtanium varistors will be resistors. I think they doubled-up on varistors just to keep impedance down. Without vibrato a whole string can be direct-coupled.
http://magnatoneamps.com/FMvibrato.html

 

[5v333]

thanks PRR
my second suggestion wouldve been a splitter with inverting output.
direct signal on the grid and RC coupled signal at the cathode.
the gain is also about 0.95 so thats good.

i actualy thought i was looking for an noninverting  circuit but now i can see how thoose opamps flip the signal.

if you got more schems or pics on triode APF, dont hold on to them!

 

[PRR]

> InvAllPassFilter.pdf

You still aren't getting it.

What is C32 doing??

Opamps have two inputs.

That's not convenient in tubes, but two OUTputs is real easy. Your two all-pass legs connect to the two outputs.

 

[5v333]

i was  trying to mimic an opamp based APF. but i guess i got it all wrong.

if the input signal was appearing at both grid and cathode it would cancel any or a lot of ac operation in the tube right?
so in a way there are two inputs...

the cap forms a RC pole with R61.

the two signals would modulate over frequency. but maybe more like an LPF.

im glad for this last pic. i can see how it is set up but perhaps a little read up on APF is in order on my side.

 

[abbey road d enfer]

i was  trying to mimic an opamp based APF. but i guess i got it all wrong.

Actually, your circuit can be made to do about what you intend, i.e. shifting phase from 0° to 180° continuouly, whilst maintaining a reasonably stable amplitude. This is a first draft, you may tweak it to improve linearity, although there are serious drawbacks, such as the need for a very low source impedance.

 

[5v333]

Thanks for trying out a sim.
Looks like smooth curves.

I guess a proper Apf should be unitygain across all freq...

Low source z. Is it because the source sees the cathode as the dominant load? 1/gm.

The cathodyne apf that prr posted looks like there wount be any drive difficulties.
Its also a single tube operation which reduces costs and heater dissipation.

Care to sim that aswell?
Is it spice youre using?
I havnt got that far with computers.

 

[abbey road d enfer]

I guess a proper Apf should be unitygain across all freq...

As I wrote, a little tweaking is necessary. I could easily get the linearity within 0.005dB.

Low source z. Is it because the source sees the cathode as the dominant load? 1/gm.

Yes

The cathodyne apf that prr posted looks like there wount be any drive difficulties.

That's correct.

Its also a single tube operation which reduces costs and heater dissipation.

Well, the output needs to be buffered.

Care to sim that aswell?

See attached.

Is it spice youre using?

LTspice.