Hybrid amp Aloia GY-50 "a step beyond of"

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
as you know the input sensitivity of the CFP (Sziklai connection)
is higher than normal EF output stage
because inside it there is a very big gain

as you know the quality of the scene produced by an OTL stadium-style (Futterman)
is higher than normal SRPP
because there are two active devices

this is an evolution of famous Bartolomeo Aloia GY-50 :bullseye:
 

Attachments

  • OTL_DIAMOND_CFP.jpg
    OTL_DIAMOND_CFP.jpg
    69.5 KB · Views: 760
  • gy-50%20modify.gif
    gy-50%20modify.gif
    25.4 KB · Views: 871
Post it ? Well, I was going to wait until it's in a more mature state before throwing it out to the wolves. I'm looking only at the output stage, looking for options to use very little gnf and accept higher distortion as a result. The CFP provides a lot of local feedback to keep the output more linear. The issue is turn-on and turn-off of the CFP which the emitter resistor topology your circled can help with. When my current project is done I'll revisit and let you know if I have something useful to share.


Are you actually going to build one of these ?
 

Attachments

  • Bartolomeo CFP.PNG
    Bartolomeo CFP.PNG
    30.3 KB · Views: 542
  • CFP ZERO GAIN.PNG
    CFP ZERO GAIN.PNG
    14.9 KB · Views: 347
Literally every circuit in this thread has some sort of error, some are catastrophic.
First, NONE except the last one are really CFPs. A CFP 'collector' is made up by connecting the first (input) transistor emitter and the second (output) transistor collector to a single point. Such a compound transistor still needs an emitter resistor to stabilize it's working point. When connected this way the bias network only has to have thermal contact with the first transistor. Note that none of the circuits above do this, they all must have the bias network in thermal contact with the output transistor heatsink, and in most cases the drivers need to share this heatsink too. Too bad in some circuits here the bias network has no thermal compensation at all, or even worse, no way of defining the standing current at all. Others have standing current trim that will immediately destroy the whole output stage if the trimmer wiper loses contact even momentairly. Yet others have gain, which seems to completely be missed by most posters, so they are not just followers/buffers but complete amplifiers.
And, that 'OTL' schematic with a concertina phase splitter is just plain wrong. The top output triode is a follower with gain 1, the bottom works as a grounded cathode amp so with a gain much larger than 1 - hence impossible to connect as a totem-pole push-pull and have it work with anything approaching symmetry.
Oh and that 'DC servo' is nothing of the sort, rather a bootstrap connection to increase the input impedance of the BJT buffer stage. Too bad the point of it has been missed because it cannot increase the impedance past that of the DC balance trimmer, which is not bootstrapped. Incidentally this is the only proper CFP schematic, too bad about the tube part....

All of this just goes to show you cannot trust everything you find on the net, and if you can trust something, you can't just go splicing bits and pieces together without understanding how they work.
 
Thanks ilimzn

wrong Capacitor on CFP
see now the correction:rolleyes:

the first problem of CFP is the self-oscillation
best solution that I have found
without difference of sound
it's adding a POWER RESISTOR on output;)

see OTL now, with ground input reference :cool:
 

Attachments

  • BARTOLOMEO ALOIA 2.GIF
    BARTOLOMEO ALOIA 2.GIF
    13.1 KB · Views: 398
  • rs pwr resistor.JPG
    rs pwr resistor.JPG
    176.8 KB · Views: 306
Last edited:
Finally CFP !

Literally every circuit in this thread has some sort of error, some are catastrophic.
First, NONE except the last one are really CFPs. A CFP 'collector' is made up by connecting the first (input) transistor emitter and the second (output) transistor collector to a single point. Such a compound transistor still needs an emitter resistor to stabilize it's working point. When connected this way the bias network only has to have thermal contact with the first transistor.
So let's say if you put the CFP emeitter resistor in the ESP project CFP, You may eliminate the CFP collector resistor (bridge it). At most you could put a small driver-only emitter resistor for stabilty. What do you think?
See schematic..
 

Attachments

  • SLA.pdf
    13.1 KB · Views: 201
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.