F4 Beast Builders

Official Court Jester
Joined 2003
Paid Member
as it seems to me , that article is one big mishmash

few good pieces , rest fifi fog

in best case - poor transcribe of PL boy'z words
 

Attachments

  • the-fog.jpg
    the-fog.jpg
    18.9 KB · Views: 374
The one and only
Joined 2001
Paid Member
What puzzles me is the word "gain" in "stabilizing gain and bias". I do not understand what is being stabilized or possible mechanisms to perform the stabilization.

The circuitry is push-pull complementary. If you build such a thing with no
degeneration, you contend with high variable open loop gain and bias
figures. These should be stabilized.

You can imagine it as you like, but you can assume it's fairly simple.
 
The circuitry is push-pull complementary. If you build such a thing with no
degeneration, you contend with high variable open loop gain and bias
figures. These should be stabilized.

You can imagine it as you like, but you can assume it's fairly simple.

A conceptually simple approach might be 1 Farad cap across a 0.2R degeneration resistor, but that would be impractical. Or would it? AVX SCMR22C155MSBA0 is 1.5F, 5V, ESR=9mOhms and has a relatively small 16mm x 24mm radial package.
 
Member
Joined 2009
Paid Member
Because I still cannot really grab the phase margin problem I want to warn people seeing or even building F4 beasts similar to the circuits I posted that they could be unstable.
On the other side all builds I did showed never any oscillation tendency.

So I relax now a bit......:--))
 
Because I still cannot really grab the phase margin problem I want to warn people seeing or even building F4 beasts similar to the circuits I posted that they could be unstable.
On the other side all builds I did showed never any oscillation tendency.

So I relax now a bit......:--))

Yes, I do not know if low values phase and gain margin are a problem when the OLG is relatively low. Perhaps Nelson's setting of the -3dB frequency well below the frequencies where there might be an oscillation addresses the issue.