I want to understand the problems in the following design (instrumantation amplifier using 3 op amp

crutschow

Joined Mar 14, 2008
26,793
One problem with that design is that the common-mode rejection depends upon how well the resistors are matched.
For 1% resistors the common-mode rejection could be as high as -40dB.

IC instrumentation amps use laser trimming of the internal resistors to get much higher common-mode rejection.
 

Ian0

Joined Aug 7, 2020
1,628
I would have called that a differential amplifier (but with input buffers). I wouldn't have called it an instrumentation amplifier unless it had resistors between pins 1 and 2, and pins 6 and 7; with the gain set by a resistor between pins 2 and 6.
 

MisterBill2

Joined Jan 23, 2018
8,246
It is a special case of an IA. Normally yes, there are resistors in those feedback loops AND a gain setting resistor between the two inverting inputs. THAT is the IA that we are most familiar with.
 

Veracohr

Joined Jan 3, 2011
749
It looks like someone did a poor job of making a quad opamp in their CAD software, so that each individual opamp shows the supply connections. Where did that screenshot come from?
 

crutschow

Joined Mar 14, 2008
26,793
each individual opamp shows the supply connections. Where did that screenshot come from?
Looks like LTspice.
LTspice does not have any information about how many op amps are in a package since it only does circuit simulation, not PCB layout, so the supply pins are shown for each op amp used.
 

Ian0

Joined Aug 7, 2020
1,628
Looks like LTspice.
LTspice does not have any information about how many op amps are in a package since it only does circuit simulation, not PCB layout, so the supply pins are shown for each op amp used.
Pulsonix (pcb design) does the same thing. It's a pain to have to connect Vdd and Vss to each inverter in a 74HC04!
 

MisterBill2

Joined Jan 23, 2018
8,246
OK, so now I can provide a complete summary of the flaws. First, the CAD model is wrong in showing power connections on each section of a quad package. THAT was the CAD error. The other errors were in the resistors that were left out, which made it not really an instrument amplifier. In an actual instrument amplifier there are feedback resistors from the output to the inverting input of both of the input amplifiers, AND there is the single gain-setting resistor then connected between those two inverting inputs. Without those three components the assembly is not an actual instrument amplifier, it is something else. It could be called an abnormal special case, at best. The comments about resistor matching are OK, because one of the manin applications of an instrument amplifier is to reject the common mode voltage at the input.
 
Top