Total offset voltage in an op amp

[Kingcosmos]

I have searched high and low for an equation that describes the total
offset voltage of an op amp and I have yet to see one. I just read a
nice article on Planet Analog (
http://www.planetanalog.com/printableArticle.jhtml?articleID=30000023 )
and I have read through many books like Op Amps for Everyone, Op Amp
Applications Handbook, Operational Amplifiers etc. and there seems to
be different ideas or at least different assumptions being made when
offset voltage is concerned.

For instance, the article on Planet Analog mentions intial voltage
offset, voltage offset due to drift (over temp and aging), voltage
offset concerning PSRR and CMR but does not mention input bias
currents, input offset currents (intial or drift). Likewise, the books
I previously mention seem not to be concerned with PSRR and CMR voltage
offset errors but do mention input bias and offset currents.

My guess is that there are certain assumptions that are made in order
to do a first order approximation when analyzing circuits. In my
limited experience, I would be concerned with initial offset, offset
due to drift, and input bias/offset currents; however, it still would
be nice to find a gospel so-to-speak on voltage offset where in one
place it mentions all these things in one place.

So far this is my understanding on what can effect voltage offset:
1) Initial Voltage Offset
2) Offset due to drift from temperature and aging effects
3) Input bias currents
4) Input offset currents
5) PSRR
6) CMR

I am sure I am missing something. The more I dig the more 'gotchas'
crop up. Can anyone else offer their insights and comments? Is there
a be all to end all equation or does one make a first order
approximation and then refine their way to the solution?

 

[John Popelish]

I have searched high and low for an equation that describes the total
offset voltage of an op amp and I have yet to see one. I just read a
nice article on Planet Analog (
http://www.planetanalog.com/printableArticle.jhtml?articleID=30000023 )
and I have read through many books like Op Amps for Everyone, Op Amp
Applications Handbook, Operational Amplifiers etc. and there seems to
be different ideas or at least different assumptions being made when
offset voltage is concerned.

For instance, the article on Planet Analog mentions intial voltage
offset, voltage offset due to drift (over temp and aging), voltage
offset concerning PSRR and CMR but does not mention input bias
currents, input offset currents (intial or drift). Likewise, the books
I previously mention seem not to be concerned with PSRR and CMR voltage
offset errors but do mention input bias and offset currents.

My guess is that there are certain assumptions that are made in order
to do a first order approximation when analyzing circuits. In my
limited experience, I would be concerned with initial offset, offset
due to drift, and input bias/offset currents; however, it still would
be nice to find a gospel so-to-speak on voltage offset where in one
place it mentions all these things in one place.

So far this is my understanding on what can effect voltage offset:
1) Initial Voltage Offset
2) Offset due to drift from temperature and aging effects

This may add to or subtract from the original offset voltage. This is
only much of a concern if you add an offset zero adjustment to the
opamp, and wonder how stable that adjustment may be.

3) Input bias currents

Has nothing to do with the offset voltage between the two inputs
needed to hold the output at mid scale.

However, this current causes voltage drop across ant input of feedback
network it passes through, according to Ohm's law.

Note that many opamps have an input bias current that varies with the
input common mode voltage, especially some rail to rail input types.

Fet input opamps (both jfets and mosfets) generally have bias currents
that rise as temperature rises, so if you need a low bias current
opamp to work in a hot environment, check this carefully.

4) Input offset currents

Again, this has nothing to do with the offset voltage between the two
inputs. The input offset current is the mismatch between the two bias
currents. If you made the total ohms of each network connected to the
two inputs, and the two bias currents were equal, then the voltage
drops those two currents produced would make canceling input voltage
offsets to the signals. But if the two currents don't match, then
even if you match the resistance of the two networks, there will be
some mismatch between the voltages the bias currents generate, and
this will effectively (algebraicly) add to the input offset voltage.

5) PSRR

Power supply rejection ratio. This relates effectively an input
offset voltage that is proportional to the supply voltage, by this
ratio. For instance, if the PSSR is -80 db (1/10,000), and the supply
changes by 1 volt, then the input offset voltage will change by 1 volt
times 1/10000 or .1 millivolt.

6) CMR

Common mode rejection ratio. This relates essentially to an input
offset voltage that is proportional to the common voltage on the two
inputs. So if this ratio is -80 db and the two inputs change their
voltage by 1 volt (but keep the same actual difference between them)
the input offset voltage will change by .1 millivolt.

I am sure I am missing something. The more I dig the more 'gotchas'
crop up. Can anyone else offer their insights and comments? Is there
a be all to end all equation or does one make a first order
approximation and then refine their way to the solution?

Usually the worst case input offset voltage spec covers the sum of all
these components (except offsets generated externally by bias
current), unless some operating limitations are included as notes.

The difficulty of making an equation that tells anything other than
worst case, is that many times several of these errors may cancel each
other out under many conditions. But if you need a guarantee, assume
that they all add in the same direction. But, still, they should all
add up to less than the worst case input offset voltage (over time and
temperature).

 

[Kingcosmos]

Thanks John. Good explanations of the PSRR and CMR; I will put those
tools in my belt for future reference. It just irks me sometimes that
different books, app notes, and white papers are 'inconsistent.' Or
maybe it is just me.

 

[John Popelish]

Thanks John. Good explanations of the PSRR and CMR; I will put those
tools in my belt for future reference. It just irks me sometimes that
different books, app notes, and white papers are 'inconsistent.' Or
maybe it is just me.

Then I hope I got most of it right. There are some very sharp
posters, here, and I don't think anyone corrected me, (here's your
last chance, guys), so good luck with that.

 

[Chris]

Then I hope I got most of it right. There are some very sharp
posters, here, and I don't think anyone corrected me, (here's your
last chance, guys), so good luck with that.

Hi, John. I put on the green eyeshade last week when you posted.
Great answer.

I've use the the worst case input offset voltage (over time and
temperature) as a rule of thumb, too.

Cheers
Chris

 

[John Popelish]

Hi, John. I put on the green eyeshade last week when you posted.
Great answer.

I've use the the worst case input offset voltage (over time and
temperature) as a rule of thumb, too.

I appreciate you checking.