Fix the original P3A mistakes.
To much current for the differential
needs degen
Not enough current for second gain stage.
Probably a attempt to fix other issues.
Since all the currents are just wrong
to much DC offset.
After that gets fixed.
Should be marginally stable.
Its CFP cross your fingers.
Or cut your losses and go with something that's still active and supported in the forum. Mooly's MOSFET amp for instance.
Agree Ancient was being excessive.
Even with slower transistors the design is marginal.
To much diff current with no degen.
Not enough current for Vas and drivers.
Needs stabilization on the VBE multiplier
C6 is a band aid that might work with slower transistors.
actually causes problems with anything else.
probably try n fix the thing
and of course the usual alternative design that will get tossed
up.
All boring "lin" topology. Bad PSRR .
Oh yeah , the base circuit is too much "on the edge" to handle a good input stage. so lets "swamp" it with more Cdom...
Q1/2's beta would determine stability - use Re
Why not add the current mirror and the beta tranny ? $1.00 and 3 trannies = .003% @ 20K ??
Oh yeah , the base circuit is too much "on the edge" to handle a good input stage. so lets "swamp" it with more Cdom...
Q1/2's beta would determine stability - use Re
Not if the IPS is regulated. 12V Behind another cap multiplier.
That design I posted loses @ 4V per rail - but it's worth it.
Never actually clip the OP devices , LTP (diamond buffer) never sees anything - CCS's are isolated.
OS
Last edited:
original P3A
AM4 diff current too high 3.22 ma
feasible too high with no degen.
AM3 second Gain stage.
Marginal low current 5ma
Should be closer to 6 to 8ma
AM2 driver transistors
Marginal current 2.66ma could be closer to
3 or 4ma
VF1 DC offset 92 mV = bad
Look familiar? = Ringing Oscillator
Just removing C6 reduces the ringing, pretty much pointless part
causes more problems. Still not stable. So lets try to fix it
without redesigning the whole board and with just value changes
To fix DC offset raise R7 to 820 ohms
lowers current from 3ma to 2ma and helps to stabilize the amp
since degen cant be added.
Lower bootstrap resistors R9/10 to 2.7k
Raises current from 5ma to 6ma.
This changed DC offset from 92mV to 15.8 mV
vast improvement.
Improved stability greatly.
100n needs to be added across T9
and removing C6 contributed the most
to actually stabilizing the amp.
C4 can possible be lowered to 68p to improve
high frequency slew rate.
If the the amp is still unstable add 1.8p across R5
Lower R12 and R11 to 150 ohms to increase driver current.
from 2.66mA to 3.66mA
This will help keep bias from being sky high or way at the end
of the pot travel. If bias still to high at way end of the pot
you can even go lower to 120 ohms
No ringing, feasible DC offset
AM4 diff current too high 3.22 ma
feasible too high with no degen.
AM3 second Gain stage.
Marginal low current 5ma
Should be closer to 6 to 8ma
AM2 driver transistors
Marginal current 2.66ma could be closer to
3 or 4ma
VF1 DC offset 92 mV = bad
Look familiar? = Ringing Oscillator
Just removing C6 reduces the ringing, pretty much pointless part
causes more problems. Still not stable. So lets try to fix it
without redesigning the whole board and with just value changes
To fix DC offset raise R7 to 820 ohms
lowers current from 3ma to 2ma and helps to stabilize the amp
since degen cant be added.
Lower bootstrap resistors R9/10 to 2.7k
Raises current from 5ma to 6ma.
This changed DC offset from 92mV to 15.8 mV
vast improvement.
Improved stability greatly.
100n needs to be added across T9
and removing C6 contributed the most
to actually stabilizing the amp.
C4 can possible be lowered to 68p to improve
high frequency slew rate.
If the the amp is still unstable add 1.8p across R5
Lower R12 and R11 to 150 ohms to increase driver current.
from 2.66mA to 3.66mA
This will help keep bias from being sky high or way at the end
of the pot travel. If bias still to high at way end of the pot
you can even go lower to 120 ohms
No ringing, feasible DC offset
Many DIYA projects fall into this category. Just add a few more components and it becomes "magic".
I would gladly add a few more semi's or a cap to not have the shame of oscillation. "hang the oscillator"
I like oscillators - analog VCO synthesizers - yummy !
OS
Thank you , I'm too lazy for correcting a CFP. 😉
I'm a EF2/3 person.
Actually , I have a EF3 "fetish" - I run 3mA VAS's... (Ice cold)
OS
Just wanted to see if something was possible using the board
in hand. Got to download some new spice models.
All good.
3mA Vas is good times. Remember when I got 5mA to work
and thought how is this even possible. LOL
Not in this case obviously, always a fun thing.
There is a little magic with NPN diff and bootstrap Vas
for intentional high 2nd harmonic that people enjoy.
I get it why people like the sounds of this amp.
But CFP outputs can go and save a world of trouble to get
the same sound. Otherwise the board is what it is.
Was just working with what is there. Try find a solution.
Sim somewhat cold start devices temp 22 to 27c
Used stated 34 volt rails. Internal resistance added
to be less ideal in sim. Shown 8 ohm load.
More ringing with 4 ohm, forgot to show it.
Seems to help in sim land to expose stability issue
using less ideal voltage sources.
in hand. Got to download some new spice models.
All good.
3mA Vas is good times. Remember when I got 5mA to work
and thought how is this even possible. LOL
Not in this case obviously, always a fun thing.
There is a little magic with NPN diff and bootstrap Vas
for intentional high 2nd harmonic that people enjoy.
I get it why people like the sounds of this amp.
But CFP outputs can go and save a world of trouble to get
the same sound. Otherwise the board is what it is.
Was just working with what is there. Try find a solution.
Sim somewhat cold start devices temp 22 to 27c
Used stated 34 volt rails. Internal resistance added
to be less ideal in sim. Shown 8 ohm load.
More ringing with 4 ohm, forgot to show it.
Seems to help in sim land to expose stability issue
using less ideal voltage sources.
Last edited:
but you can get that "bootstrap sound" with EF2/3.
EF3 , you get the current gain of the CFP , without the instability.
Add the current mirror and the beta driver to the VAS , and you get the dom 2'nd and
.003% @ 20K.
Parts are cheap ... whats the issue ?
OS
EF3 , you get the current gain of the CFP , without the instability.
Add the current mirror and the beta driver to the VAS , and you get the dom 2'nd and
.003% @ 20K.
Parts are cheap ... whats the issue ?
OS
I completely agree.
just solution for the board in hand.
Otherwise yes
plenty of changes could be done.
CFP...snip
just solution for the board in hand.
Otherwise yes
plenty of changes could be done.
CFP...snip