Maker Pro
Maker Pro

Feedback resistor in non inverting unity gain opamp?

The OPA2134 datasheet indicates that, in non-inverting configuration,
the impedance seen by both inputs should be matched to minimize
distortion if either of them is > 2 kohm. So, if I have a unity gain
Sallen Key filter where the source impedance (in the passband) is
around 27 kohm, should I put a 27k resistor (rather than a short)
between output and inveting input? I guess in theory I should but I
don't think I've ever seen this done, and if I'm not mistaken this
problem of increased distortion with different Zs at the inputs is
common to many FET opamps, not just the 2134.

Many thanks in advance,

Carlos
 
P

PeteS

Jan 1, 1970
0
For the scenario you have, you are correct.

It should be noted that setting the inputs to 'see' the same input
impedance minimises input bias current errors (a difference in input
impedance will mutiply the error by the amount of mismatch to give you
an input offset voltage, which then appears at the output as an error).
Input offset current errors are always going to exist (which is a good
reason to have a low impedance driving source).


The error you will get (signal) is roughly I(bias) x Z(diff) x (Av)
where Zdiff is the difference in the effective difference in input
impedance seen at each input and Av the closed loop gain of the amp.

At DC, the output error (for this issue) would be I(bias) x R(diff ) x
Av.
Only resistive issues apply to the large signal model, of course.

As for the mismatch, it's usually more pronounced for FET input amps
but it exists as an issue in all op-amps.


Note that for your case, you will have a signal error of I(offset) x
27k x Av.

Apart from trimmers (which I won't use unless I must), the only simple
way to minimise input offset current errors is to keep the source
impedance low.

Cheers

PeteS
 
P

Pooh Bear

Jan 1, 1970
0
PeteS said:
For the scenario you have, you are correct.

It should be noted that setting the inputs to 'see' the same input
impedance minimises input bias current errors (a difference in input
impedance will mutiply the error by the amount of mismatch to give you
an input offset voltage, which then appears at the output as an error).
Input offset current errors are always going to exist (which is a good
reason to have a low impedance driving source).

The error you will get (signal) is roughly I(bias) x Z(diff) x (Av)
where Zdiff is the difference in the effective difference in input
impedance seen at each input and Av the closed loop gain of the amp.

At DC, the output error (for this issue) would be I(bias) x R(diff ) x
Av.
Only resistive issues apply to the large signal model, of course.

As for the mismatch, it's usually more pronounced for FET input amps
but it exists as an issue in all op-amps.

Note that for your case, you will have a signal error of I(offset) x
27k x Av.

Apart from trimmers (which I won't use unless I must), the only simple
way to minimise input offset current errors is to keep the source
impedance low.

I don't see how you addressed the issue of *distortion* that the OP
mentioned. The issue re DC < damn I used *that* word again > conditions is
well understood.

That's a new one on me for sure !

Graham
 
J

John Larkin

Jan 1, 1970
0
For the scenario you have, you are correct.

It should be noted that setting the inputs to 'see' the same input
impedance minimises input bias current errors (a difference in input
impedance will mutiply the error by the amount of mismatch to give you
an input offset voltage, which then appears at the output as an error).
Input offset current errors are always going to exist (which is a good
reason to have a low impedance driving source).


The error you will get (signal) is roughly I(bias) x Z(diff) x (Av)
where Zdiff is the difference in the effective difference in input
impedance seen at each input and Av the closed loop gain of the amp.

At DC, the output error (for this issue) would be I(bias) x R(diff ) x
Av.
Only resistive issues apply to the large signal model, of course.

As for the mismatch, it's usually more pronounced for FET input amps
but it exists as an issue in all op-amps.

But fet opamps have picoampere bias currents, and they usually aren't
guaranteed to have even the same polarity, much less the same
magnitude.

John
 
Top