Mpp

[Joachim Gerhard]

I am tired. I will continue tomorrow.

[Stefanoo]

Originally Posted by Joachim Gerhard

I have some concerns about the gain i am getting. If i have not wired something wrong it seems too low.

The gain strongly depends on the gain introduced by then CCS.
If I am not wrong that arrangement should not boost more than 50db or so
This is just my guess and I am going to run simulation tomorrow and post results.
The dc offset can't be evaluated until the circuit provides about 80dB.

Accordingly to my simulation you only need 4 k170 with no feedback resistor and a resistor of about 270Kohm to obtain proper gain.
However I have noticed that with BJT current sources the input impedance doesn't reach high enough value to have a reasonable current riaa.
However them CCS I drew allows proper functionality.

Today. Was too lazy..I had dentist and then I didn't get a chance to work on the CCS.
Beside I am also really engaged working on a new line stage with a very regarded designer of this community which name shall remain anonymous and I am ideally behind with that too!!! Ahhhh so many things I would like to do so little time I can and want to do them.
Please keep us informed on your findings with the CCS and e gain on the circuit.

Thanks for all your efforts.

[FdW]

After some experimentation I got this (see attachment) it has a very slight negative temperature coefficient and seems to be really stable.

Varying the load from 1 to 100 ohm gives only a delta of 2.7 uA on the 26mA output current, that is about 1ppm.

As it is running at a good current it will get warm and thus the negative temperature coefficient will help stabilize it.

The zener diode can be replace by a string (3 or 4) of led’s. And a capacitor could be added parallel to Q3(ec), this will remove some noise. I do not know what the overall noise figure of this circuit is.

[Edit] P.s. it runs (in the schema) at/from 200V, but that is not needed the collectors of Q1 and Q2 are (in the schema) at about 7 Volt.

[Edit] P.s. it will be needed to replace the 1Meg resistors with 100K resistors when running from a more realistic voltage.

[Joachim Gerhard]

So you are really using a big P-Channel one. I should have some.
Have you calculated the gain of this circuit ?
No success on my side so far. I get the DC drift super stable but i end up with too low gain. Of cause the low gain relaxes the demand on the CCS.

[FdW]

Originally Posted by Joachim Gerhard

So you are really using a big P-Channel one. I should have some.
Have you calculated the gain of this circuit ?
No success on my side so far. I get the DC drift super stable but i end up with too low gain. Of cause the low gain relaxes the demand on the CCS.

As the current in the 170's is very critical it is difficult to get a realistic figure for gain and noise (and this is not my field of expertise you know ). Maybe the easiest way is just to build it (it's not very complex).

See it as just one of Frans’s doodle’s

[Joachim Gerhard]

Ok, i do the doodledoodoo today.

[Joachim Gerhard]

I will use 3 x green Leds. Are they not running on very low current ?
It may work though. Can use a zener instead.

[FdW]

Originally Posted by Joachim Gerhard

I will use 3 x green Leds. Are they not running on very low current ?
It may work though. Can use a zener instead.

Yes, they are running on 'ultra'-low current, no light will be visible

[Joachim Gerhard]

I think the voltage divider needs to be high impedance for high output impedance of the CCS. I think that is the problem in my recent design.

[FdW]

Originally Posted by FdW

After some experimentation I got this (see attachment) it has a very slight negative temperature coefficient and seems to be really stable.

Varying the load from 1 to 100 ohm gives only a delta of 2.7 uA on the 26mA output current, that is about 1ppm.

As it is running at a good current it will get warm and thus the negative temperature coefficient will help stabilize it.

The zener diode can be replace by a string (3 or 4) of led’s. And a capacitor could be added parallel to Q3(ec), this will remove some noise. I do not know what the overall noise figure of this circuit is.

[Edit] P.s. it runs (in the schema) at/from 200V, but that is not needed the collectors of Q1 and Q2 are (in the schema) at about 7 Volt.

[Edit] P.s. it will be needed to replace the 1Meg resistors with 100K resistors when running from a more realistic voltage.

Some research gives the attached graph (for the IRF9610) it shows that Vgs is independent of temperature for about 100mA Id; for the IRFP240 this is at 4.8V at 300mA (about). Somewhere there should be a fit for 30mA (as is needed here).