High Power Class G amplifier

[digitoxin]

Hi!

I made a class G amp. Online here: http://audiolab.i-networx.de/

Comments are welcome!

Harald

[Sheriff]

Hi Harald,

Are you happy with the sound quality? What have you compared it to?, and what have you driven speaker wise with this brute...?

Best Regards

Sheriff

[Samuel Jayaraj]

What were the rail voltages you used? How does one manage at these kind of power levels without short-circuit or overload protection? Sound??

[fr0st]

That's exactly the kind of amp I've been looking for.
Have you made it yet?
Do you think that changing the PMBTA92 for a Bc *** type would have any affect to its operation? (BC556 for example)
and the output pairs to MJ15003/4? or maybe TIP3055/2955 for an easier to manage package.
Other than that I can get everything required to make it.
Well done

[digitoxin]

Hello!

About the sound quality: Spheric, precise, powerful. Perfect for PA use. Problems can arise with cheap drivers that tend to be sharp, because precise+powerful+sharp sound can be unpleasant (the amp itself does not sound sharp). The results with good drivers and bass speakers were excellent (not just my opinion (which is a little biased .

@Samuel: The amp does have a current limiter/overload protection.

@frost: The transistors were chosen because of their CE-voltage.

With the given rail voltages the PMBTA-transistors can be changed to others with at least 150V CE-Voltage, the inner power transistors can be changes to ones that can withstand 1xouter+1xinner voltage= 170V, the outer can be changes to ones with outer-inner voltage=80V. The power transistors should be able to handle 15A continous current. And a small SOA chech should be done.

The 3055/2955 can not handle these voltages and cannot be used therefore. The BC-types as diff-pair also can not be used because of their CE-Voltage.

Harald

[sajti]

Originally posted by fr0st
That's exactly the kind of amp I've been looking for.
Have you made it yet?
Do you think that changing the PMBTA92 for a Bc *** type would have any affect to its operation? (BC556 for example)
and the output pairs to MJ15003/4? or maybe TIP3055/2955 for an easier to manage package.
Other than that I can get everything required to make it.
Well done

BC556 is not useful, due it can handle only 65V. Try to use MPSA92 or 93, which can go up to 200V, of higher.
For output the MJE15003/4 is OK, but I think MJ15024/25, or MJ21193/94 (even MJL 21193/94) is much better.

sajti

[sajti]

Is there anybody who can explain, how I can count the voltages for class G amplifiers?
For this amplifier the total psu voltage is +/-125V. But why 45+80? Why not 62,5+62,5? What is ideal for dissipation?

Thanks: sajti

[phase_accurate]

What is ideal for dissipation?

That is actually depending on program material. The voltage ratio that gives the highest efficiency for continuous full power isn't necessarily the one that is best for normal program material.

Regards

Charles

[sajti]

Originally posted by phase_accurate


That is actually depending on program material. The voltage ratio that gives the highest efficiency for continuous full power isn't necessarily the one that is best for normal program material.

Regards

Charles

OK, but how I can deteminate it, in practice?

sajti

[phase_accurate]

Douglas Self has written an article about class-G amps where he presents some curves generated by playing around with the voltage partition. I you like it I can mail it to you.

In general you would have to be able to perform integral calculus to calculate things like that properly. But there are also the possibilities of simulation (like P-SPICE) or you can do an approximation with programs like Excel.

Regards

Charles