A question about diamond buffers

[darian]

Hello everyone,

Recently (since this morning actually) I got very interested in diamond buffers and found several schematics. I understand the principle for the Walt Jung version and some derived ones but not this one for which I post the schematic. I may be a dumb question but how is the second stage correctly polarised? To me it looks like a class B or even C stage... Probably I'm missing something in the design...

[darian]

The Walt Jung design is limpid, with the emitter resistors on the first stage setting the second one in class A...

[darian]

And finally this one used in an Accuphase amplifier is very elegant with the first stage biasing the second one...

[juma]

Take a peek here:
http://www.diamondstar.de/dDB_design.html

[darian]

Thank you for your answer,

These ones have the emitter resistor for bias as it should. I just wondered about the scheme without them. I found it on a forum about headphones. They could be Sijosae design or something. I just wondered about the crossover distortion in a strict class B like that. It is better in class A.

And if the Vbe of the first stage is lower than the second, you get a nice class C one.

I may be wrong but I wanted a confirmation.

Anyone?

[darian]

And I wanted to share the Accuphase one which I find very interesting with the autobias acting like a faint NFB too.
Very elegant...

[forr]

Hi Darian,
You are right, the first schematics you show is wrong.
I would be curious of the results obtained with the Accuphase circuit.

[wakibaki]

Darian

I read the circuit (1) like this: -

The circuit is symmetrical and collector voltages in the first stage will be equally offset from the supply rails when the Hfe of the transistors is equal.

Since the bases of the input transistors are at the same voltage (connected), the differential voltage between the collectors is Vbe(337) + Vbe(327). This is insufficient to bias the second stage into conduction, because of the voltage drop across the 2*4R7 emitter resistors and because the transistors used are identical to the ones in the first stage.

So this is a class B push-pull amplifier and will have pronounced crossover distortion.

w

[darian]

Thanks for your answer,

So I was right about it. I guess no one complained about it in the headphone forum (Headwize I think). There may be some people listening to that badly designed circuit and thinking it's a good one. Too bad for them...

Anyway, I have my answer. If someone wants to try the Accuphase circuit, I'd be pleased to hear some feedback. It's so smart and so simple at the same time.

Thanx

[wakibaki]

Originally posted by wakibaki


The circuit is symmetrical and collector voltages in the first stage will be equally offset from the supply rails when the Hfe of the transistors is equal.

Since the bases of the input transistors are at the same voltage (connected), the differential voltage between the collectors is Vbe(337) + Vbe(327)

Sorry

I was about to go to bed when I wrote that. I went to sleep and woke up thinking about it.

In both instances this should read emitter, not collector.

This is 2 cross-coupled emitter followers, so voltage gain < 1. It's purely a current buffer.

Since the Vbe's of the stages are likely to be close, the xover is probably fairly smooth.

If you wanted to run it AB, you could do this:-



w