discrete opamp current sources?

There's probably a latchup mode. If they shared and the output rails high, Q4 could saturate and steal all the bias from Q3, turning off Q6-Q7, and it's dead forever.

Old original LM324s suffer from shared bias supply problems, but at least they recover.

John

You forgot Q5... Better to use matched transistor pairs in same package (for better thermal coupling), and use a resistor in place of that "extra" diode where the resistor matches (in value) that used for Q3, Q4 and Q5.

Thanks, I'd hadn't thought about fault conditions, maybe I'll LTspiceit, and see if that happens, but I still dont trust (LT)spice for wierd stuff, but thats probably just my ignorance

martin

"martin griffith" schrieb im Newsbeitrag news: snipped-for-privacy@4ax.com...

Hello Martin,

If the opamp is driven near saturation (Q6 off), a lot of current will flow into the base of Q4 and thus lowering the voltage across the two diodes. This would be a large unwanted change in the current for the differential input stage. The base current of Q4 will also change in the normal operation with every change of the output voltage. This will modulate the bias current source of the input pair too. Think on additional distortion here.

Over all two good reasons for seperate bias voltages.

Best regards, Helmut

The three 27K resistors could be reduced to two.

The input diodes prevent another latch mode, zenering an input b-e junction. It may have some other hazards if connected as a follower.

John

Probably to avoid a latchup if the -IN voltage ever gets high enough to bottom Q3. With joined bases the Q4 cc-source would then suddenly change current.