Abomination! or Painting a Mustache on the Mona Lisa meets the Island of Dr. Moreau

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
MJL21193 said:
I'm feeling postpartum depression here or some such. An anticlimax that has me wondering what I'll do next.

Hi John,

Ever consider trying Scott Wurcer's distortion reduction technique from the AD797 datasheet on this thing? I don't think anybody here has done that with a power amp. Might be interesting to just look at in the simulator. Not sure what effect it has on stability.
 

Attachments

  • wurcer.jpg
    wurcer.jpg
    29.1 KB · Views: 604
Disabled Account
Joined 2007
andy_c said:


Hi John,

Ever consider trying Scott Wurcer's distortion reduction technique from the AD797 datasheet on this thing? I don't think anybody here has done that with a power amp. Might be interesting to just look at in the simulator. Not sure what effect it has on stability.


Interesting. I hadn't known of this but I just gave it a try. The sim shows a THD drop - from 0.001 to 0.0007 so that's not hurtin'.
Stability seems to be unaffected in the sim.
Worth trying on the actual amp (not that I'm likely to hear the 0.0003 % THD drop) to see if it does alter stability.

Thanks Andy.
:)
 
MJL21193 said:
Interesting. I hadn't known of this but I just gave it a try. The sim shows a THD drop - from 0.001 to 0.0007 so that's not hurtin'.
Stability seems to be unaffected in the sim.
Worth trying on the actual amp (not that I'm likely to hear the 0.0003 % THD drop) to see if it does alter stability.

Not much of a drop - LOL! Just to check, you did use 30p and not 15p for the extra cap, right? (sum of compensation caps).
 
Disabled Account
Joined 2007
andy_c said:


Not much of a drop - LOL! Just to check, you did use 30p and not 15p for the extra cap, right? (sum of compensation caps).


Ah, no. I used 15p. I just tried 30p and the drop is an extra 0.0001% but there is a slight overshoot evident, especially on the bottom half of the squarewave (in simulation).
I'll need to try it on the actual amp to see the effect. Back in a bit...
 
MJL21193 said:



Working with both the simulation and the actual amp, i haven't had any luck removing C4. Stability seems fine, but the DC offset is a real problem. Simulator shows 3.5VDC and the actual amp shows 2.1VDC or so offset. My attempts to bring this down in simulation have ended with worse overall performance.
A servo circuit would be the answer, but I'm not ready for such a complication right now.
[/IMG]
if you remove C4 then the inverting input sees 22k//680r.
The non-inverting input sees 22k.
The input offset current generates a voltage difference due to the difference in source resistances, thus the output offset.
Remove C6 and the non-inverting input sees 22k//2k2. Not quite matching but much closer. Short out both DC blocking caps and measure the output offset. Adjust the offset by changing that 2k2 down towards 680r, or maybe lower.

I keep repeating myself;
Either build a DC coupled amplifier with all the protections required to prevent speaker damage
or
build an AC coupled amplifier.

Do not try to build a mixed AC & DC coupled amplifier.
And if you do build a DC coupled amplifier then the source must not have a DC blocking cap on it's output or the servo will have to offset hard to correct for the input resistance error.
 
Disabled Account
Joined 2007
AndrewT said:


I keep repeating myself;
Either build a DC coupled amplifier with all the protections required to prevent speaker damage
or
build an AC coupled amplifier.

Do not try to build a mixed AC & DC coupled amplifier.


Hi Andrew,
It was never my intention to build a DC coupled amp. I have no interest (personally) in such an animal. It was a request by KLe that I try the amp without C4. I gave it a little attention but in the end gave up. One needs to be properly motivated to make these changes and I am not.
The mixed idea you may have is from my first posted schematic where I hadn't attached the input blocking cap yet. The correct and fully up to date schematic is in post 141.
 
Disabled Account
Joined 2007
AndrewT said:
Hi,
yes, post141 schematic is the one where both C4 & C6 are shown and you discuss the offset problem if C4 is omitted. Mixed AC & DC coupling with the unmatched resistances explains all.

Well Andrew, you motivated me to try again. :(
My fatal flaw the first time was running all of the simulations with the input unconnected.
:eek:

Another rookie mistake. With the input grounded or connected to a signal, I have the offset down to 50mV range and with your suggested change to R15 (from 2.2K to 500R), 6-7mV.
Another lesson learned. :)
 
Disabled Account
Joined 2007
Abominable results for BAIN!

Happy New Year to you also KL!

I became restless last night and played around with this for a while:

BAIN-1.png


Many changes and if I'm not mistaken, slightly better performance than Abomination. More simulated testing is needed though.
Please look it over for obvious bone-headed mistakes. :)
 
It looks like "proper" CCS's are popping up everywhere..
(reproducing):D :D :D ...And hey ...a buffered widlar CM
(betcha it gets 70-80 db gain now)
and a CFP input . Whoaaa...
Happy New Year (NY) ...I used to live there:mad:
OS

Edit: U5 and U1 emitters need a 33R across them.. then you will
have a"leach" OP..:)
 
in the spirit of new year,

how about bob cordell thermal-trak bias spreader and T-output? u need to modify the upper and lower CCS for VAS connection.

you could add mosfet flavour by swapping bjt driver over with low Ciss MOSFET.


keep up the good art works..


cheers.
 

Attachments

  • thermaltrak_bias_2.png
    thermaltrak_bias_2.png
    81.2 KB · Views: 782
Disabled Account
Joined 2007
ostripper said:
It looks like "proper" CCS's are popping up everywhere..
(reproducing):D :D :D ...And hey ...a buffered widlar CM
(betcha it gets 70-80 db gain now)
and a CFP input . Whoaaa...


Hi OS,
In for a penny, in for a pound...may as well go whole hog and throw more transistors at it. The ones that stick improve the design! :D

Just ignorantly playing around last night, so I may have some more work to do.

The CFP input was in the last version of this, several posts ago. I haven't examined the gain yet, but here's the simulated PSR. Better, by a long shot, than Abomination:

An externally hosted image should be here but it was not working when we last tested it.



Happy New Year! :drink: :drink:
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.