MJ Optimized Valve Regulator

[korneluk]

Folks,
Any specific reason why Morgan Jones uses an EF91 pentode in the optimized valve voltage regulator (Fig. 5.38, 3rd ed., pg. 347) and an EF86 in the basic version (fig.5.37, pg.343)?
Any reason why the optimized version could not be implemented using two EF86s instead?
Thx,

 

[kevinkr]

I can't see any reason why you couldn't, but I can't see much reason for building such a complex regulator either.

Take a look at cascode error amplification, and a pentode connected pass element with a well filtered screen supply. This will get you reasonably low source impedance (a couple of ohms is possible), low noise and sufficient loop gain for good regulator performance. AC performance will be significantly better with a zener as its dynamic impedance is much lower than the gas reference, that said a gas reference like the 5651 or 85A2 often performs better in terms of stability and drift in my experience.

If you search carefully you'll find one of my designs lurking around on the forum and this could be a starting point.

 

[korneluk]

Take a look at cascode error amplification, and a pentode connected pass element with a well filtered screen supply. This will get you reasonably low source impedance (a couple of ohms is possible), low noise and sufficient loop gain for good regulator performance.

Steve Bench used this approach usig an EL509 with a floating screen supply. Output impedance of 0.7ohms across the audio band.

Tube Based Voltage Regulators - Part 3