Universal Audio Amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
OK, so my problem is that the original post is 6+ years ago and Mr. Bayes has abandoned it, hopefully have learned something.
And y'all have hijacked the thread talking about alternate solutions.
OK, besides the QSC floating supply, you can bootstrap the op-amp for up to 2x the chip voltage (~+/-30VDC). However this taxes the input CMR.
And there are rail driven op-amp circuits, but these use more added bits than just making a nice clean "blameless" amp.
 
Disabled Account
Joined 2015
The rule of the thread.

SSBayes who posted 3 months ago last time had the idea to drive the outputs directly by an op amp . I fully respected his rule .
There are 32 solutions to drive a high power amp by an op amp , the most elegant is to short the output of the op amp and use the current of it's power supplies. I post an example 500w/4ohm CFA topology for you bellow but I don't want to discuss anything about it, as it is against the rule of SSBayes.
 

Attachments

  • Demo.JPG
    Demo.JPG
    87.4 KB · Views: 227
Disabled Account
Joined 2015
The thermal stability can't be simulated. It should only be conceived on the real object. As I mentioned ,either you use a emitter resistances and R9 is NTC or voltage multiplier mounted on the heat sink , or the voltage multiplier(BD140) by 2 determined by 3Mohm, is installed upon the PNP transistor (screwed). I have a 50w class A without emitter resistors regulated by this method works perfectly since a decade . You can even by adjusting the two resistors of the multipler and the pull up resistor, program to have a variable bias current suitable for each temperature .
 
Last edited:
Disabled Account
Joined 2015
more temps

I also used several times LM334 temperature dependent current source. It can very precisely adjust the temperature drift . You need to make a piece of metal holder to apply upon the transistor .Calculation of the resistors is relatively simple arithmetic once you know how the bias voltage should decrease vs temperature.
 

Attachments

  • bias circuit.JPG
    bias circuit.JPG
    52.5 KB · Views: 201
OK, thanks. Although with LTspice (and I think also TINA) you can include temp in your simulation.
Jan
Yes and No.
Spice ( The original for LTSpice, Tina and other variants ) has the temp 'variable':).
But, it is more like a constant :mad:. Spice does not calculate temp. The user can set a temp value for each componant, but that value will not change during the run of a simulation.
There is no way to see transistors junction temperature evolve according to varying powers from current and voltage variations.
There are some clumsy workarounds, but all is to be made by the user, there is no true support from Spice.
Devices never burn in Spice, device matching is perfect, temperature drift is just a dream. No thermal runnaway, temperatures will stay set at their default or initial values.
 
I have received a huge number of posts, which, I am unable to read now. I hope, I would be able to review them in the future.

I expect most posts to be in relation to the push pull output stages. These have been made in another project, which, has been published in another post. They have been tried and they work well.

Other people have tested them with Spice.

There are two versions : direct ( just amplifier and transistors ) and indirect ( amplifier, circuit, transistors ). The two approaches work well. The direct approach makes something, which I call " Machine Gun Noise " : one to a few bangs per second. I do not have an oscilloscope and I do not know what the reason for this noise is, but, this is the only problem. The noise seem to be present only with the direct approach : when the transistors are unbiased. The problem is not present with an output capacitor, which is the way I have been able to make the direct approach work.

The indirect approach is OK.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.