Here is a quick test of a russian СГ2С (SG2S) 75V voltage valve regulator:
The CCS is a classic cascoded DN2540 MOSFET. A russian PIO capacitor is bypassing the regulator and the input impedance is the Pete Millet’s interface impedance. HT is a well regulated variable power supply.
When near maximum bias current is achieved (circa 30mA), the noise floor is surprisingly low: -85.5dB / 53uV.
Need to capture input of power supply at the CCS input 
Edit: after measuring the power supply noise level and finding that is near -90dB, came up to the conclusion that the above noise level is the VR noise. Quite good for a simple (but effective) voltage regulator. Gary Pimms used this configuration many times on his designs…
Update
Upon request from Vyacheslav, here is the noise analysis for different current points:
As you can see it’s very linear. There is a 4dB difference between 5mA and 40mA.
Also tested ripple attenuation using a poor-regulated HT power supply. Here are the interesting results:
- 62dB @50Hz
- 80dB @100Hz
+70
Hi Alejandro,
very interesting analysis СГ2С!
ripple -85.5dB / 53uV – if the current is 30 mA,
but what level ripple – if 5mA, 10mA, 20mA, and 40mA (when the maximum current)? you have the measurements?
I want to know when (level current) minimum ripple noise?
P.S. they also have nice illuminated
Vyacheslav.
Ukraine, Dnepropetrovsk
Hi Vyacheslav,
No, I don’t have. Did a quick test, so will try to measure noise level at different currents.
Cheers,
Ale
Hi Vyacheslav,
Had to rerun the tests, see updated post
Cheers,
Ale
Ooo, the linear response!
balloons glow (neon-argon mixture) 
Alejandro. Many thanks!
Photo – looks very good
Yes, a good stabilitron (VR) – and the price is low … for example, in Ukraine (in Dnepropetrovsk, the radio market), only 0.5$
Vyacheslav, Ukraine, Dnepropetrovsk
Hope this helps! Will drop you an email, is the email you registered a valid one?
Thanks
Ale
I already answered you
thank you
Vyacheslav