Symmetric Amp, Interesting Output Stage

[Hentai]

Hello,

Looking back through some of my old notes, i found a schematic of a power amp i drew a while back. It was based on a 3 transistor buffer designed by John Bowers an engineer from Dynatronics. I have used this buffer back then in preamps and loved it. The components nowadays may render it obsolete but it can achieve low drift and low offset, better than darlington

Current gain of the stage formed of Q44,46,47 is Beta44*Beta46*Beta47 +Beta44*Beta46+Beta44+1.

A basic sketch of the amp is attached. I haven't played with solid state in a good time so I'm a little behind with components available today. I would like to hear from you comments about it, if it is worth pursuing and i would like help on choosing the best transistors.

I have looked on Sanken, Toshiba, On Semiconductor offer and have some ideas but would like to hear more from you.

Based on my experience with this buffer in line level applications i think it will be worth looking into using it with power applications.

Thank you.

 

[AKSA]

Hi Hentai,

This is a very interesting output stage, unique.

However, and the double phase shift is similar to the Bryston, I suspect this would create a distributed phase shift.

This might cause errors for depth of image and possible with resolution. You might know if this is right or wrong, I would be interested in your insights of the sonic qualities of this amp.

Cheers,

Hugh

 

[Dave Zan]

Problematic front end

I think the front end will suffer from the usual problem of ill defined current in the VAS with this IPS, as seen in the Slone amp and similar. Have you ideas for this?
An unfamiliar buffer is fun to see. What benefits do you think it has?

Best wishes
David

 

[magnoman]

Hentai

I don't understand how this output topology can set the idle current through Q47/48.
It's a CFP but with an emitter follower added, however as shown there doesnt appear to be feedback directly to the driving transistor.

Thanks
-Antonio

 

[Hentai]

Hi, thanks for replies.

However, and the double phase shift is similar to the Bryston, I suspect this would create a distributed phase shift.

This might cause errors for depth of image and possible with resolution. You might know if this is right or wrong, I would be interested in your insights of the sonic qualities of this amp.

I will study this and will post some sims soon.

I think the front end will suffer from the usual problem of ill defined current in the VAS with this IPS, as seen in the Slone amp and similar. Have you ideas for this?
An unfamiliar buffer is fun to see. What benefits do you think it has?

I've looked at the Slone amp, indeed similar front end. Can you offer more details about its problem? Front end will suffer modifications i'm sure.
As far as benefits at a first glance high current gain, low offset and drift, keeping base-emitter voltage at Vbe and lower saturation voltage than darlingtons.

I don't understand how this output topology can set the idle current through Q47/48.
It's a CFP but with an emitter follower added, however as shown there doesnt appear to be feedback directly to the driving transistor.

CFP with emitter follower, yes thats a good description. By driving transistor you mean Q45/46? if so it is just like CFP.

 

[jerluwoo]

Will probably have to add some low value resistors from Q43/44 collectors to the rails since they will draw all of their current through Q45/46 BE junctions and may cause them to burn out.

 

[Hentai]

Don't think it will be necessary as the current drawn by Q43/44 is very small, that's why it can connect directly to VAS stage.

 

[TANDBERGEREN]

Will probably have to add some low value resistors from Q43/44 collectors to the rails since they will draw all of their current through Q45/46 BE junctions and may cause them to burn out.

For the same reason I neither could figure out that output stage, but actually there will not be drawn too much current through there, as the Q45/46 will ensure that the Q47/48 weill be driven much earlier than the Q43/44.

All in all a near pointless topology for this output-stage, unless one solely looks at the purpose of leaving as little load to the VAS as possible.
But to obtain that, there is quite a number more efficient ways to do exactrly that.

I would really say that Q45/46 is in fact a part of the VAS.

 

[Hentai]

Well Q45/46 output current, dont think it as part of Vas. I'm not sure what you mean by more efficient and i would like to read more about it.

 

[Dave Zan]

I've looked at the Slone amp, indeed similar front end. Can you offer more details about its problem?

Start here http://www.diyaudio.com/forums/solid-state/16796-unstable-vas-current-amp-slone-book.html

and here http://www.diyaudio.com/forums/soli...ror-image-topology-construction-troubles.html

Bob Cordell's book discusses this and shows an effective solution. Edmond Stuart's CMCL circuit concept also appeals to me. But I am still interested in new ideas.

Best wishes
David

 

[Dave Zan]

part 2.

As far as benefits at a first glance high current gain, low offset and drift, keeping base-emitter voltage at Vbe and lower saturation voltage than darlingtons.

The buffer is a three transistor loop so instability is a real risk. You have not shown any compensation even for the overall loop. Both will need careful attention.

Best wishes
David

 

[Hentai]

Hi David, thanks for the link, i am studying them.

Yes it is just a basic schematic. I'm running sims that i will post soon there are some nice features tho that i will show about this buffer.