Push-Pull cross-over distortion ?

Put the bases on the output of the opamp and pick the feedback two emitters. That will minimize the crossover distortion. But may not completely eliminate it. Be careful driving reactive loads or you may get oscillation.

Why not just use a part with two independent NPNs?

On a sunny day (Wed, 14 Apr 2010 08:05:48 -0400) it happened Hammy wrote in :

Depends on how you drive it. Sometimes you can just put a resistor from the bases to the emittors, at low signal levels (where the transistors are still off) the driver then drives the load via the resistor.

• | c ---b | e

----- drive ---------- R1-| | |---------- out | e --b c | -

You can reduce the distortion and push it up in frequency like this:

---! NPN -------!+\ R1 ! !\e ! >--+--/\/\----+ ! --!-/ ! ! +-------------+------------ ! +--/\/\----------+ ! ! ! R2 ! ! ! +---!!--- ! !/e ! ! ---! PNP ! ! R3 ! -----/\/\------------------------------

You need a very fast slewing op-amp like the LT1351 or 1354

The local feedback capacitor needs to be just enough to prevent oscillation at the gain cross over frequency.

R1 is perhaps as low as 10 Ohms. At very high frequencies the output of the op-amp looks like a weird LC circuit. You need to deQ this a little.

R2 lets the op-amp drive the load during the cross over.

Yup, that's a good one, inside a feedback loop.

John

Thank you Jan. I'll try that.

Can you stand the dissipation if you class-A biased the output by pulling a DC load current?

I used to do LM324's that way, so they could be used in active filters. ...Jim Thompson

I'm trying to keep the power consumption to a minimum both to minimize heat and to keep the power supply small. I'd like to get as wide a swing (as close to +/- 5V) from a +/- 6V rails or less.

The THS422 can accept 15V max but the lower the rail voltage the less dissipation in the output transistors. The less current the THS422 sinks/sources the wider the swing for the push/pull input which is why I was looking at the dual NPN/PNP high gain low Vce sat duals.

This is for the output of a AD9834 10MHz signal generator I built a while ago. Its only using a AD8057 to bump the filtered output up to about 2Vpp. I'm considering ways to get this upto 10Vpp into 50 ohms which seems to be the standard max output on most of the commercial ones I looked at. I plan on using the THS422 as an adjustable gain stage so I can have some amplitude adjustment. I haven't got that far yet though ;-).

Bridge feed forward compensation:

If the bridge is balanced (R1 x R2 x C = L), then the compensation is perfect.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

-------

=A0 =A0 =A0 !

=A0 !

=A0 =A0 !

=A0 =A0 =A0 =A0 !

I used to build some audio line drivers similar to that. Transistors were a push-pull emitter follower following an op-amp, Negative feed back for the op-amp was picked off after the transistors. This worked better than several other biasing schemes I tried.

Jimmie

Thanks

I'll try that as well.

That's pretty cool Vladimir. Can I think of this has balancing two times R1*C and L/R2? And if that is correct, do you pick the time to be faster than than the highest frequency of interest? Or is there a better way to think about the circuit.

George H.

Looks quite similar to the "current dumping" feed forward system used in the Quad 405 power amplifier a few decades ago.

When the Quod 405 was introduced, there was a lot of discussion about this kind of topologies for instance in Wireless World. This might help googling for similar systems.

On a sunny day (Fri, 16 Apr 2010 08:24:01 +0300) it happened Paul Keinanen wrote in :

Exactly, and I used it in a 150 kHz 'RF' amp around 1982 or so.

Thanks for the hint. For any interested there is a link to a paper by PJ Walker down the bottom of this site that does a good job of describing what's going on.

George H.