unbypassed cathode and frequency response

[catalin gramada]

Hi,
I just started a project. First stage analising 12ax7, 220k plate with 2.2k unbypassed in cathode. 2.2M in grid and 1M preceded by 10nF as load. 320v supply. Pretty large enough. What I don't understand is why for the same pp voltage in input e.g. 282mv get 9v at 400hz and just 4.8v at 8khz for instance.It was expected to start to roll off over 10khz in respect with internal capacitance by published graphs. How is explained that brutal depreciation in frequency response ,please?It was feed it from 600 ohm impedance source without any grid stop resistor at the input. Thanks

[loudthud]

What are you using to make this measurement? Does the generator measure flat all by itself?

[nickb]

^^^That and I'll add, to get a meaningful response you should feed from a source impedance representative of the final application, maybe 33k or so. Right now you are shunting the input and miller capacitances with the low generator source 600 ohm impedance.

[Dave H]

It should be flat to way beyond 10kHz with a 600R source. What's hanging on the output? x1 scope probe, DMM, coax cable? A couple of hundred pf capacitance would be enough to cause that roll off at the output but would measure flat when driven by a 600R source. Try measuring it with just a x10 scope probe connected to the output.

[catalin gramada]

Thanks for answers. I just monitoring on my dual scope Rigol 1102E with 1x scopes probe. Reversing channels get the same results.

[nickb]

Thanks for answers. I just monitoring on my dual scope Rigol 1102E with 1x scopes probe. Reversing channels get the same results.

A x1 probe will have significant capacitance about 50pF. Use a x10.

[pdf64]

Bear in mind that with an unbypassed cathode, the stage output impedance will be somewhat higher, ~= 220k//1M, ie maybe 4x higher than with a bypassed cathode. Hence any capacitive loading will have a proportionately greater curtailing of the HF response.
Input stages are generally much better to be fully bypassed; if unbypassed, noise at the cathode, eg from the heater and cathode resistor, are added to the input signal.

[Helmholtz]

I don't think any of the ideas above can explain almost 6dB attenuation at only 8kHz.

Please post a schematic.

[mozz]

This may help.
https://www.ampbooks.com/mobile/ampl...ode-capacitor/

[Helmholtz]

This may help.
https://www.ampbooks.com/mobile/ampl...ode-capacitor/

I just love the books of Richard Kuehnel, but as we don't have a cathode bypass cap here, the explanation of the early gain drop must be something different.

[J M Fahey]

And what is the actual generator used?

[catalin gramada]

And what is the actual generator used?

A cheap analog one: Instek gag 810

[bea]

I don't think any of the ideas above can explain almost 6dB attenuation at only 8kHz.

Please post a schematic.

Fully agreed. There must be a some significant stray capacitance somewhere in the circuit. Even a cheap generator should and will normally perfectly drive the input of a tube. And even the 80k output impedance of an ECC83 stage should be easily measurable. Quick check - the abovementioned 50 pF at a source impedance of 80 kOhms yields a -3dB cutoff at 40 kHz. Which means that the capacity of the probe needs to be taken into account, but should not explain the observation.

What's needed: a schematics of the stage PLUS a check if the actual parts correspond to the schematics. Maybe a check in the manual of the probe on its real input capacitance. Same for the scope - the two are in parallel and thus add.

The initial post does not tell us the size of the input grid stopper. If that is as large as it might be recommended by some for high gain stages, the loss at high frequencies might be expected (and also desired, high gain distortion with lots of treble does not sound well).

[catalin gramada]

There is nothing in between grid stop was bypassed and signal applied straight to the grid. It is just one triode connected we don.t talk how this triode works in some context, just by a measuring error generate by who knows reasons. The main suspect are the cheap 1× probes delivered with the scope. I ordered some decent 10x probes expected to come tomorrow. Thanks for Your answers.

[Helmholtz]

There is nothing in between grid stop was bypassed and signal applied straight to the grid. It is just one triode connected we don.t talk how this triode works in some context, just by a measuring error generate by who knows reasons. The main suspect are the cheap 1× probes delivered with the scope. I ordered some decent 10x probes expected to come tomorrow. Thanks for Your answers.

Is there a reason why you don't show a schematic?