Common mode rejection ratio with opamps?

[cupoftea]

Hi,
If we have an opamp with 0v and +5V supply, then its common mode voltage is +2.5V, and so the Common mode rejection ratio will be max for signals near 2.5V?
If we have an opamp with +5V and -5V supply, then its common mode voltage is 0V , and common mode rejection ratio will be greater for signals near 0V?
If its an InAmp with 0V and +5V supply, then the CM rejection ratio is the same no matter where the signal is between 0V and +5V?

 

[barry]

No.

 

[cupoftea]

Thanks,
Though would you agree that the attached is an instrumentation amplifier with a gain of 1?
Its common mode rejection ratio will be as good as an Inst Amp because there is no amplification to a common mode signal at its inputs.?

Also, its difficult to find an OTS In Amp with a gain of 1, so the attached, which could be done with a cheap quad OpAmp, looks like a game changer?
LTspice and jpeg attached

 

[danadakk]

CM performance highly dependent on R matching and Aol of the OpAmps.

An example sim :

Above the R's represented by the fdbk factor blocks.

IAs with G = 1 : https://www.analog.com/en/parametricsearch/13010#/


Regards, Dana.

 

[dick_freebird]

DC CMRR generally comes from the forward gain and a
stable Vio (front end gain and balance), through the
feedback network. The same is true of PSRR. Of
course poor design choices can degrade things, from
there.

As inputs or outputs get close to a supply rail there is
likely to be some degradation to AVOL, Vio or both. If
that happens then CMRR will degrade as well.

Now many op amp topologies will be fine at one end
of the range and bad at the other - look for hints in the
datasheet regarding input headroom, at where the
output load-driving attributes are measured, etc. It's
a good bet that outboard of those ranges, occurs
"stuff" that the vendor isn't real proud of. Like the
"good" ol' LM124 with its Vcc-(2Vbe+Vce(sat)) upper
bound in input voltage for meeting main specs - no
way around needing to keep the front end lit. Though
more modern "rail-rail-I/O" architectures are better,
you can still see (if you peel off a couple layers of the
datasheet onion) that gain is depressed within a volt
or so of the rails. They probably neglect to mention the
other "fellow traveler" attributes which suffer the same.

 

[KlausST]

Hi,

The descriptions - especially about common mode voltage - of post#1 and post#3 don't match at all.

Often INAs are made for higher gain.
And the higher gain stage is in the first stage and it designed to amplify common mode with gain of 1 and differential modd with the desired gain.

Your circuit of post#3 thus is a "special case" of instrumentation amplifier. It does not benefit from the classical INA design.
One can say it is a classical differential amplifier circuit with buffered inputs.

I can not find any "game changing".

So I did a search on "instrumentation amplifier datasheet" and checked the first 5 hits:
* INA115 ... can be operated the same way as circuit of post#3
* INA114 ... too
* AD8223 ... is something "special" can work with gain>=5
* AD622 ... can do
* AD623 .. can do

What "game changing" do you see?

Klaus

 

[cupoftea]

Thanks,
The "game changing" is in that INA115 and co cost $5+ whereas LM124 (etc) Quad opamp is <$1

And the higher gain stage is in the first stage and it designed to amplify common mode with gain of 1 and differential modd with the desired gain.

Thanks, and since CMRR is (DM gain/CM gain), then if the first stage only has a gain of one, then the InAmp has very poor common mode rejection?

 

[KlausST]

Hi,

You really compare a LM124 with an INA115?
Vos: 2mV vs 0.05mV
VOSDRIFT: 7uV/°C vs 0.25uV/°C
IB: 20nA vs 2nA
CMRR: poor 70dB (and where are the resistors? Dana's post) vs 115dB including laser trimmed resistors

Just the cost: What is the cost for LM124 plus 4 laser trimmed resistors to get a CMRR of 115dB?
(Indeed I'd say it is impossible to get close to the CMRR (only!) performance of the INA115)

My opinion: one could say the invention of an INA115 was "game changing".

I can not find a single parameter where 3(out of 4 parts of) LM124 beat the INA115.
Not even power supply 3× 1.5mA vs 3mA .... nor speed (2 x 1.3MHz in series) vs 1 MHz

Klaus

--- Updated Feb 15, 2023 ---

then the InAmp has very poor common mode rejection?

Even then the INA has 45dB better CMRR. This is about a factor of 30,000.

It's the other way round: if you have a higher input gain stage, then the INA has additional benefit.

Klaus

 

[cupoftea]

It's the other way round: if you have a higher input gain stage, then the INA has additional benefit

Thanks, yes, but we only want a gain of 1.
We just want to pass a feedback signal between PWM controllers for whom ground varies slightly.
We cant just pass the error voltage with respect to ground, as there are ground drops between the PWM controllers....(due to current flows)

 

[KlausST]

Hi,

A "current" feedback from a shunt?

--> Which solution to choose depends on many things, like frequency, and source impedance... and so on.

We have no idea how your source looks like nor how the output signal should look like.

A "gain of 1" application isn´t the usual target for an INA.

Klaus

 

[cupoftea]

It's the other way round: if you have a higher input gain stage, then the INA has additional benefit

Thanks, yes i think its best for a "classical differential amplifier circuit with buffered inputs", as you kindly called it.

 

[danadakk]

If you want to grind thru the CM G, or CMRR :

http://www.ittc.ku.edu/~jstiles/412/handouts/2.4%20Differnce%20Amplifiers/Differential%20and%20Common%20Mode%20Gain%20lecture.pdf

OR https://masteringelectronicsdesign.com/differential-amplifier-output-common-mode-voltage-calculator/


Regards, Dana.

 

[barry]

Your initial question was basically: ’does CMRR change with common-mode voltage’. The answer is, maybe a little, but not enough to matter. Then you totally change the discussion to your “game-changing” standard 3 op-amp inst amp