Why don't BJT IC opamps use resistive degeneration in the input stage?

[godfrey]

They often use resistors at the output for current limiting and there's normally at least one somewhere to set the idling currents, so why no input stage degeneration?

In fact chip designers seem to go out of their way to avoid using resistors as far as possible. Any thoughts?

 

[DF96]

Resistors use a lot of silicon? Better to use three or four matched BJTs than a resistor.

 

[dirkwright]

Yeah, considering all of the benefits of at least some degeneration in a differential input stage, I don't get it either. Even if the transistors are matched, there is further benefit from degeneration, at least based on my experiments in simulation.

 

[richie00boy]

How do you know they don't? The schematic on the data sheet is just conceptual.

 

[dirkwright]

How do you know they don't? The schematic on the data sheet is just conceptual.

Well, here's the NE5532 schematic. No degeneration in the input pair, but plenty of other resistors shown. I don't claim to understand this schematic very well.

 

[AndrewT]

Is it noise specification that over-rules all other design choices?

 

[dirkwright]

Is it noise specification that over-rules all other design choices?

Do degeneration resistors actually add that much noise, if any at all? I don't know really. It doesn't seem to me that they'd add much if any noise.

 

[dirkwright]

I can only think that they don't use them because they want the maximum gain in the first stage. Wouldn't max gain (in addition to the CCS in the tail) translate to higher CMRR for the differential input pair?

 

[Bonsai]

1. Noise - especially an issue where the opamp is used in a high gain configuration
2. Maximize loop gain at LF (remember the DC performance of opamps is a large part of their utility value)
3. They most often use Cdom MC, so provided output loads are not unduly capacitive, they tend to be stable, so no need to use input stage degen to reduce loop gain enough to put the output stage pole below the UGF (pole splitting effect of Cdom MC helps here as well)

Others already mentioned above as well.

 

[dirkwright]

There also may be a game of "specsmanship" going on here.

 

[godfrey]

Resistors use a lot of silicon?

Possibly, but it's hard to imagine that chip real estate is such a big issue. Perhaps there's some difficulty in creating decent quality or close tolerance resistors. I notice that specifications that rely on resistors tend to be quite badly defined (e.g. quiescent current or max output current).

Better to use three or four matched BJTs than a resistor.

That does seem to be the approach sometimes taken, but I'm still wondering why.

 

[SY]

Close tolerance is hard but close matching is easy. Of course, the resistors also degrade noise specs since their Johnson noise is uncorrelated.

edit: I see that one of the smart guys has already mentioned the noise issue.

 

[godfrey]

Close tolerance is hard but close matching is easy.

Thanks, SY - That's good to know.

So for an opamp circuit like the one below, we don't know exactly what the idling currents of the two input stages are, but they will be closely matched (when the common mode input voltage is close to zero).

The main feature of the circuit is that the common mode input voltage range extends about 0.5V beyond the supply voltage in both directions.

The benefit of biasing the input stages with resistors is that as the common mode input voltage moves from rail to rail, control is handed over gradually from one input stage to the other, with the total transconductance remaining constant.

I'd been wondering how guilty to feel about the resistors, but it should be fine.