Slewmaster - CFA vs. VFA "Rumble"

It would be even more interesting to see how the circuit behave when followed by an output stage and FB loop closed. The compensation schemes and possibilities are also quite intriguing. Before having a chance putting it to simulator I personally have a bit of concern about the output impedance of the VAS being possibly higher than we would like it to be at high frequencies, which may not be a good news for 3EF OPS stability.

That is why I am using those 10-15pf shunts at the VAS collectors.
Seems to run real nice with the slew EF3.
With just Miib's 5pf lead comp.(from VAS) , it was insufficiently damped (ringing).
Tuned the bode response , tested square waves , and PSSR.
I now know where it excels (very high output) .... keeps 10ppm 20K 140V p-p.
The mirrors are "holding it together" at these obscene levels.
OS
 
The current-source can be made floating just by removing The GND connection on The VBE currentsource set-transistors, this will prevent any possibility of dirty ground sneaking into the circuit.
When you have a servo the GND connection is good to have in order to have a good reference.

For the input transistors find some good doubles, then performance will be stellar.

The VAS cascode can also be set with a LED reference to rails.

I like the added compensation to GND, It just may be the better option
 
The current-source can be made floating just by removing The GND connection on The VBE currentsource set-transistors, this will prevent any possibility of dirty ground sneaking into the circuit.
When you have a servo the GND connection is good to have in order to have a good reference.

For the input transistors find some good doubles, then performance will be stellar.

The VAS cascode can also be set with a LED reference to rails.

I like the added compensation to GND, It just may be the better option

I did that (below) , as well as make the CSS Re's globally adjustable to set
the current of the TIS. Injected my servo reference there , too.

I'll do both this one and the symetri ... one simple , one hard. AE is just short
of the spooks part count.
(below 2) AE self clamps on overload , as well. All parts conduct less (clamp).
.ASC below ....
OS
 

Attachments

  • AE V1.0.gif
    AE V1.0.gif
    30.7 KB · Views: 438
  • AEclamp.gif
    AEclamp.gif
    20.5 KB · Views: 398
  • MiB-AE.asc
    MiB-AE.asc
    33.3 KB · Views: 117