2SA1015 SRPP follower bridge mode

[dirkwright]

So, I'm putting this out there for comments. It's a simple circuit really, with a pair of 2SA1015's in SRPP formation driving an emitter follower on each side of the bridge. Feedback is taken from the emitter follower back to the emitter of the SRPP. I used ideal transformers in the sim because I wanted to know what the circuit was going to do. Obviously, performance will be much worse with real world transformers, which I already have (Jensen). I can't see anything wrong with this circuit so I want to build it but thought I'd post it first just to double check. I hope I didn't make any obvious errors. The only down side so far is that the input impedance is a bit low at 6k ohms.


I've tried different transistors and other arrangements but for 6dB of gain, this is just fine apparently. Performance seems to be really great.

[dirkwright]

I figured out how to fix the input impedance. I changed the bias arrangement to the input transistor by lowering the current through the voltage divider thus increasing the impedance. Overall input impedance is now about 12kohms.

[jan.didden]

Dirk, I was just looking at that.
Another option is to put a resistor between the bias string (top of the diode) to the b of the Q.
Not sure which is more stable - your lean current (which decreases stability of the ref voltage with T) or a series R (which causes some T dependence too).

jan didden

[dirkwright]

Originally Posted by janneman

Dirk, I was just looking at that.
Another option is to put a resistor between the bias string (top of the diode) to the b of the Q.
Not sure which is more stable - your lean current (which decreases stability of the ref voltage with T) or a series R (which causes some T dependence too).

jan didden

thanks for the advice! much appreciated!

I lowed the current of the bias divider to a bit less than 500uA and increased the resistor to 5k. The circuit sims almost exactly the same, but of course this may be different in real life.

[jan.didden]

I believe the 5k then sets the input Z. The real one will be lower because the dynamic impedance of the diode will be in parallel.
Or is it a current reg diode?
That's probably the best solution - an LM334 (?) current diode, that will work fine with only a few 100uA.

jan

[dirkwright]

Originally Posted by janneman

I believe the 5k then sets the input Z. The real one will be lower because the dynamic impedance of the diode will be in parallel.
Or is it a current reg diode?
That's probably the best solution - an LM334 (?) current diode, that will work fine with only a few 100uA.

jan

They are all current regulating diodes (CRD, CLD, CCD, etc). They are expensive, but they make circuit design more simple for me.