Help with HV adjustable power supplies for radiation detectors.

[Cai Arcos]

The problem is that I dont know the characteristic function of the CCFL-Inverter and bridge-filter themselves, so planning for compensation is, well, impossible.
The best I'm doing is inserting various capacitance values between the feedback and the output pin of the TL072 until the noise (which very clearly is caused by oscillations since by killing the bandwidht by putting 0.1uF at the output of the opamp eliminates it) goes down to acceptable levels and no oscillation is present.

[Joe Gayo]

Not impossible. It can be measured.

[Cai Arcos]

How can this be done?

[Joe Gayo]

This should get you started:

https://www.ti.com/lit/an/snva364a/snva ... 8718652691

[Cai Arcos]

If I understood correctly, my main objective using this technique is to observe the phase shift at 0db and make sure it is high enough (between 45 or 50 degrees).
If I make my function generator float, then the transformer is unnecessary, right?

[Joe Gayo]

Basically, yes.

I would use a cheap 120V to 12V 60Hz transformer.

[Cai Arcos]

Thank you very much!

[Cai Arcos]

Well, no mater what I do, I am not able to observe a sinewave on the scope. All I see is 50hz hum and the noise of the supply. I will have to try something else
EDIT: I also appear to have destroyed my last TL072, so until Monday, I'm not going to be able to do anything else
EDIT EDIT: Searching in my drawer, I found another TL072! Tomorrow will do more tests

[Richard Hull]

Those FET front-end op amps....Great in working circuitry.....Need a pile to really experiment with. Thank goodness they are cheaper by the dozen or thousands. I just bought a pack of 100, 2N7000's....Great little TO-92, low end sub-microamp input mosfet for turning on stuff in the .1 amp range using the joule energy from a mouse fart. When the turn-on energy just isn't there, the 2N7000 is.

I have used the OPA128 op amp to make single IC electrometers. It takes only 75 fempto amps to tickle the input. On a scope trace, on a winter's night, in DC follower mode with a 4-inch diameter aluminum ball on the input, my cat padding across the rug 5 feet away drove the trace off screen. I had to stay real still to not move the horizontal trace more than .5 volts one way or the other. Even manual shorting out was tough as moving away after grounding sent the trace off screen. I see why my vibrating reed Kiethley gold electrometer head has a small needle grounding solenoid to facilitate a remote grounding arrangement from the instrument box and a 6 foot cable going to the remote head.

Richard Hull