Power Supply Depepdence of 741 Op-amp

[PumaCat]

Hello,

I was wondering if anyone could explain why the open-loop gain of the 741 is so dependant on the supply voltage. The spec sheet I have show it dropping by about 20dB between Vcc = 2V and 18V.

As I understand it, the gain in the 741 comes from the Common Base stage after the input and the Darlington Pair near the output, both of which have active loads and shouldn't exhibit much power supply dependence. Can anyone help me with this or point me in the right direction?

No this isn't homework. Homework is over. I'm studying for my final, and one of the things we are suposed to do is be able to explain where every graph on the 741 spec sheet came from. I can understand all of them but this one. If it helps this is from the Motorola MC1741 sheet, but all of the 741s I've looked at show this, usualy as the last graph. Thanks for any help!

 

[Bobthelost]

Other than the obvious question of where the 741 draws power from...

*goes to dig out his notes*

 

[PumaCat]

As far as I know it assumes ideal power supplies for Vcc and Vee. Vcc is exactly +x volts and Vee is exactly -x volts, where both can provide any amount of current needed.

I don't know if its important, but the CB amplifier right after the input stage is biased (at the base) by a transistor followed by a resistor and connected to the low power rail. This bias current seems voltage dependant since this transistor's base is connected to a Widler current source. Significant? Thanks again to everyone.

 

[djhuber82]

In general, open loop opamp gain is a poorly controlled (and relatively unimportant) parameter, so I'm not supprised is varries so much with supply voltage. This is definately coming from supply-dependent bias currents. The circuit used to gererate the bias reference is a 39k resistor in series w/ two diode-connected transistors (1 npn, 1 pnp).

 

[TuxDave]

To be clear, are you saying that the open loop gain drops by 20dB when you move the supply from 2V to 18V?

 

[PumaCat]

Thanks for your help djhuber, I do think its something like that, but anything more specific or quantitative would be great. From what I can figure out the biasing on the CB amplifier reduces Ic, reducing gm and the gain. Similarly, the same refrence current is used for the darlington pair, and similarly reduces their gain when it drops at low voltage.

And no TuxDave, it's the opposite, the gain drops from about 100dB at 18V to about 82dB at 2V. Any ideas? Thanks!

 

[bobsmith1492]

I believe transistor in general have a lower hfe at lower voltages; that may simply be propagating through to the overall chip. Look at Figure 1:
http://www.ortodoxism.ro/datasheets/on_semiconductor/2N4410-D.PDF

At lower voltage, hfe drops off more rapidly. That may explain it.

 

[djhuber82]

Puma, the CB amp you keep reffering to is only a common mode loop. I'll reffer to the schematic here: http://www.uoguelph.ca/~antoon/gadgets/741/741.html
Q11, Q12, and R5 set the main bias current to Ibias = (Vcc+Vee-2*0.7)/39k, which is directly dependent on supply voltage. This is mirrored to the input stage via the Widlar current source Q10 and R4. Q8 measures the current in the input stage; Q9 compares this to the Widlar source and adjusts the base voltage of the CB amp to make the two equal.

 

[TuxDave]

Originally posted by: PumaCat
Thanks for your help djhuber, I do think its something like that, but anything more specific or quantitative would be great. From what I can figure out the biasing on the CB amplifier reduces Ic, reducing gm and the gain. Similarly, the same refrence current is used for the darlington pair, and similarly reduces their gain when it drops at low voltage.

And no TuxDave, it's the opposite, the gain drops from about 100dB at 18V to about 82dB at 2V. Any ideas? Thanks!

Edit What djhuber82 said...

Ok, makes more sense. Theincrease in supply voltage should lead to an increase in bias current through the transistors. Mainly because you have a large voltage drop across whatever active or passive load you may have in the circuit.

Since transconductance of both BJTs and MOSFETs are generally proportional to bias current (gm = 2*Id/Vdsat for MOSFET), and gain is proportional to transconductance (assuming equal load), a gain is expected.