BJT differential amplifier biasing

[likephysics]

I was looking at input stage of 741 internal ckt - http://en.wikipedia.org/wiki/File:OpAmpTransistorLevel_Colored_Labeled.svg
how exactly are the 2 transistors Q1,Q2 biased?
I looked at bjt differential input stage wiki here - http://en.wikipedia.org/wiki/Differential_amplifier#Long-tailed_pair
Fig 2 shows the biasing using collector and emitter resistors. Is there no need to put the base at >0.7v?

 

[yungman]

Bias current of the entire op-amp is set up by Q11 and Q12 and 39K resistor. The current setting on the differential pair is set in a very funny way that it almost bias on it's own but is controlled by Q3 and Q4 through their base that connected to collectors of Q9 and Q10.

It is a funny design where most of the op-amp differential stage are not like this. Current in Q5 and Q7 where Q7 is a buffer, mirror to Q6 back to Q2. Q3 and Q4 is biased by Q9 and Q10. Notice emitter of Q1 and 2 drive emitter of Q3 and 4 resp. For example, a positive change of base voltage of Q1 will increase the collector current of Q1which reflect into Q3 and to the current mirror Q5 and Q7 to Q6. Q6 current will increase and pull the collector voltage down. Because current of Q1 increase, current of Q8 increase and cause current of Q9 increase and pull the voltage of the collector of Q9 to increase. This will decrease the current of Q2 and Q4. If you think it's confusing, it is, it goes round and round. All these just to change the current of Q9 and affect the voltage of Q3 and Q4...Back in the circle again!It is a funny design as it's the first and oldest opamp out. That's the reason almost nobody use it except maybe students as it's cheap. I never once use 741 as the speed, offset, noise are so much inferior than any of the newer op-amp. Don't use it as a guide line of op-amp as this design is border line "stupid"...no offense.

 

[DragonPetter]

I'm curious, what is the modern counterpart of the 741 that is a good general purpose opamp?

 

[yungman]

There are too many to name. Go to DigiKey and look up that's in your price range, speed range, noise range.....

http://search.digikey.com/us/en/cat/integrated-circuits-ics/linear-amplifiers-instrumentation-op-amps-buffer-amps/2556125

That's the reason there are so many of them around. For multiple opamp, LM324, TLO82,LM358, LM4558.....Too many popular ones to list out.

It is very important to understand the specification of op-amps. Take the time to study the meaning of each specification and choose one that fit you.

 

[likephysics]

Bias current of the entire op-amp is set up by Q11 and Q12 and 39K resistor. The current setting on the differential pair is set in a very funny way that it almost bias on it's own but is controlled by Q3 and Q4 through their base that connected to collectors of Q9 and Q10.

It is a funny design where most of the op-amp differential stage are not like this. Current in Q5 and Q7 where Q7 is a buffer, mirror to Q6 back to Q2. Q3 and Q4 is biased by Q9 and Q10. Notice emitter of Q1 and 2 drive emitter of Q3 and 4 resp. For example, a positive change of base voltage of Q1 will increase the collector current of Q1which reflect into Q3 and to the current mirror Q5 and Q7 to Q6. Q6 current will increase and pull the collector voltage down. Because current of Q1 increase, current of Q8 increase and cause current of Q9 increase and pull the voltage of the collector of Q9 to increase. This will decrease the current of Q2 and Q4. If you think it's confusing, it is, it goes round and round. All these just to change the current of Q9 and affect the voltage of Q3 and Q4...Back in the circle again!


It is a funny design as it's the first and oldest opamp out. That's the reason almost nobody use it except maybe students as it's cheap. I never once use 741 as the speed, offset, noise are so much inferior than any of the newer op-amp. Don't use it as a guide line of op-amp as this design is border line "stupid"...no offense.

741 may not be a great pick to understand opamp internal ckts.
What about the biasing in fig.3 here - http://en.wikipedia.org/wiki/Differential_amplifier#Long-tailed_pair
Just biasing using collector and emitter resistors (or current mirror) is enough? Base doesn't need to be 0.7v?

 

[yungman]

Your figure two is a much more simplistic circuit that really not used in real design as the gain is too low because it use resistors Rc1 and Rc2 as load. And the bias current is set by Re that current change with supply voltage. So the gain is low, common mode rejection is bad.

May I suggest you to look at this one:

http://www.national.com/ds/LM/LM124.pdf

This is more the kind I would design if I need an op-amp. I design BJT IC before and This is very typical circuit I use, very simple, straight forward. I don't even use darlington at the front end most of the time.

Ignore Q1 and Q2 where they serve to lower the input bias current and increase input resistance. Q3 and Q4 is the differential pair, Q8 and Q9 is the current mirror and Q10 is the second gain stage the drive the darlington Q11 and Q12 to the output push pull stage. Cc is the dominant pole compensation cap that make the amp unity gain stable.

This one is simple, straight forward, it is still being used in designs unlike 741 that is more being talked about in school than really being used just because it was the first one out.

 

[likephysics]

Your figure two is a much more simplistic circuit that really not used in real design as the gain is too low because it use resistors Rc1 and Rc2 as load. And the bias current is set by Re that current change with supply voltage. So the gain is low, common mode rejection is bad.

May I suggest you to look at this one:

http://www.national.com/ds/LM/LM124.pdf

This is more the kind I would design if I need an op-amp. I design BJT IC before and This is very typical circuit I use, very simple, straight forward. I don't even use darlington at the front end most of the time.

The LM124 is even better. I was going to suggest LM358,but both the 124 and 358 are similar.
I had some problem with the sinking current of LM358. When used as a unity gain buffer, Q13 is off for output voltages less than 0.8v. So sinking current is drastically reduced. Not a good output stage.

Ignore Q1 and Q2 where they serve to lower the input bias current and increase input resistance. Q3 and Q4 is the differential pair, Q8 and Q9 is the current mirror and Q10 is the second gain stage the drive the darlington Q11 and Q12 to the output push pull stage. Cc is the dominant pole compensation cap that make the amp unity gain stable.

This one is simple, straight forward, it is still being used in designs unlike 741 that is more being talked about in school than really being used just because it was the first one out.

You meant ignore Q1,Q4?

 

[skeptic2]

I'm curious, what is the modern counterpart of the 741 that is a good general purpose opamp?

One opamp that I like is the CA3140. It has mosfets for the input transistors so the input impedance is 1.5 teraohms. This has advantages for some kinds of high impedance sensors. It has a slew rate of 9V/uS which is enough for audio purposes whereas the slew rate of a 741 is only 0.5 V/uS. (A 1V pk signal at 20 kHz requires about 8V/uS). Unfortunately they only come one to a package.

http://www.datasheetcatalog.org/datasheet/intersil/fn957.pdf