Composite INA Extends CMRR Frequency Range 10X

A resource for technical documentation. Datasheets, application notes, instruction manuals, books and links to resources are found in the Document Library.
Post Reply
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Composite INA Extends CMRR Frequency Range 10X

Post by mediatechnology »

"Composite Instrumentation Amp Extends CMRR Frequency Range 10X," Moshe Gerstenhaber, Chau Tran, Electronic Design, February 4, 2002.

This article shows how cross-coupled "double-balanced" instrumentation amps extend, due to symmetry, high frequency common mode range. In this thread viewtopic.php?f=12&t=700 Wurcer and Kitchen describe in 1982 a "3 dB" noise advantage of cross-coupled inputs due to stacking. An additional advantage would seem to be even-order distortion cancellation. Of the three arguments for cross-coupled inputs the extension of HF CMRR range seems to be the strongest.

Image
"Composite Instrumentation Amp Extends CMRR Frequency," Gerstenhaber, Tran, Electronic Design, February 4, 2002.

"Composite Instrumentation Amp Extends CMRR Frequency Range 10X," Moshe Gerstenhaber, Chau Tran, Electronic Design, February 4, 2002.
Full pdf:
https://proaudiodesignforum.com/images/ ... 4_2002.pdf

The top curve in Figure 3 shows the CMRR of the cross-coupled pair, the bottom curve shows the CMRR of the single differential amp U1.
Note that in audio applications the advantage that cross-coupling provides for high frequency common mode improvement doesn't require U3 (figure 1) since the line receiver of a following device or stage will likely be differential as well.
In this thread I discuss this circuit: viewtopic.php?f=6&t=422&start=45
For audio applications the AD623 is not the ideal part.
A pair of single op amp line receivers such as the THAT1246, INA137 or the dual versions of them make more sense for audio.

Gerstenhaber and Chau Tran published a later article showing how a cross-coupled input can be used to drive a differential A/D. In that article U3 is replaced by the A/D's differential input and Vcm injected into the INA's reference pins.

"Measurement Circuit Features High Common-Mode Rejection," Moshe Gerstenhaber, Chau Tran, Electronic Design, July 22, 2002.
Full pdf:
https://proaudiodesignforum.com/images/ ... 2_2002.pdf

Image
Measurement Circuit Features High Common-Mode Rejection," Moshe Gerstenhaber, Chau Tran, Electronic Design, July 22, 2002
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by JR. »

Hmmm...sorry to be contrary but I always thought instrumentation amps were very high impedance (+ and - inputs commented directly to op amp + inputs.)

Cross connecting two equal value differential amps makes a balanced input but it only extends the input common mode voltage 6 dB. So CM voltage could be 2x PS - signal voltage peak. By making the input resistors 300k delivers relatively high input impedance but at the cost of perhaps significant Johnson noise.

I killed some brain cells thinking about this decades ago after experiencing a PA stage ground with 50VAC on it relative to FOH (long story). Besides the obvious solution of using a transformer, the best topology I could come up with was connecting the inputs to two inverting opamps. The ratio of input R to feedback R would determine input CM voltage range, So 20k input resistors with 2k feedback resistors would tolerate 10x PS voltage CM but would require make up gain following that will come out of S/N. In other words there is no free lunch.

For an A/D input where the signal generally needs to be padded down this approach could deliver a lot of CM voltage.

JR

[edit= OK the AD623 really is an instrumentation amp, but at 35nV rt Hz input noise, and 12V rail max, a specialized part that may not be useful today. [/edit]
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by mediatechnology »

I think it would help to ignore the specific part numbers inside the triangles.

Treat cross-coupling as a concept.
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by JR. »

Another topology tidbit I didn't catch at first is that the AD629 differentials are 380k input but only 20k in the low side leg to references, so it will handle 19x the rail voltage for input voltage dynamic range, Using +/- 18v rails that is a bunch. :mrgreen:

The downside to this is the noise gain of the differentials are also 19x so not very quiet. :oops:

Of course there is little need for that much voltage headroom so building a version using low noise opamps and precision resistors with less of a pad ratio could handle tens of volts of signal swing from a modest CODEC power supply. The trick for good CMR is to balance one ground leg with the reference leg in parallel with the feedback R. In the standard part the 380k in parallel with the 21.1k resistor is very close to the other 20k resistor.

Interesting topology...

JR
Cancel the "cancel culture", do not support mob hatred.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by mediatechnology »

The thing I really like about the cross-coupled inputs are the fact that it's balanced in, balanced out and realizes common mode rejection in-between.
It allows a simple injection of Vcm.
It has symmetry.
There's beauty in symmetry...
User avatar
JR.
Posts: 3700
Joined: Sat May 24, 2008 7:21 pm
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by JR. »

Agreed.. it is especially useful for driving a dual input A/D since it will also convert a single legged input into dual polarity output, however I suspect if we look inside that A/D it is not two convertors, but using an internal differential (specualtion).

There is a noise (gain) trade-off for the extended input CM range, but for less difficult applications the amount of that penalty can be trimmed back.

Note: My dual inverting input approach has pretty much the same noise trade-off so no free lunch.

JR
Cancel the "cancel culture", do not support mob hatred.
emrr
Posts: 570
Joined: Sun Mar 02, 2008 10:27 pm
Location: NC, USA
Contact:

Re: Composite INA Extends CMRR Frequency Range 10X

Post by emrr »

Interesting read here, thanks.
Best,

Doug Williams
Electromagnetic Radiation Recorders
Post Reply