Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Accuracy of differential opamp?

Status
Not open for further replies.

Flyback

Well-Known Member
Hello,
We wish to monitor the current from a battery as in the schematic shown of a differential opamp with LT6220 opamp.
However, the LT6220 datasheet only gives the CMRR for Vin from 0 to 3.5V, and only when the supply to the opamp is 5V.
However, we have an 8V supply to the opamp, and the common mode voltage can go up to 4.8V.
Therefore, do you think our setup will be very accurate? We don’t need fantastic accuracy, but if the opamp output voltage is more than 10% away from what it should be, then that may be a problem.

LT6220 Opamp datasheet
https://cds.linear.com/docs/en/datasheet/622012fc.pdf

Please also find attached the LTspice simulation of it. (this shows no problem but I think the opamp models are idealised in the simulator, and have infinite CMRR over the entire input range, so the simulation cannot be relied upon.
 

Attachments

  • Differential opamp schematic.pdf
    7.8 KB · Views: 174
  • Differential opamp.asc
    1.9 KB · Views: 120
Sorry but zxct1009 cannot be used down at the lower end of the battery voltage, ie, 2.4V.

zxct1009 datasheet
**broken link removed**

an NIMH cell can be 0.8V at lowest, and also, with wiring cable drops, the 2.5v rating of the zxct1009 etc is unfortunately not sufficient
 
Last edited:
Do you think that this instrumentation opamp version of the above circuit will have a better CMRR and better input offset voltage?
 

Attachments

  • instrumentation opamp.asc
    2.4 KB · Views: 115
  • instrumentation opamp SCHEMATIC.pdf
    8 KB · Views: 182
Another post with a ridiculous question that has been answered. What's the purpose?

John
 
Therefore, do you think our setup will be very accurate? We don’t need fantastic accuracy, but if the opamp output voltage is more than 10% away from what it should be, then that may be a problem.

I don't see any problems. The opamp is overkill for this kind of application, but should work fine. Note that LT, Unitrode (TI), and Maxim make integrated high-side current monitor ICs, some with the shunt resistor built-in.

ak
 
From the datasheet: "Input Common Mode Range Includes Both Rails"

The input offset voltage and current change notably above 3.5V CM. If the offset voltage/current are larger than you'd like, you can just put an extra 1k resistor from each input to ground to reduce the CMV and adjust the 10k resistors if necessary.
 
From the datasheet: "Input Common Mode Range Includes Both Rails"
thanks, but then why do they only quote CMRR up to 3.5v when the rail is 5v......3.5v is not "rail to rail" when rail is 5v.

also will the instrumentation amp version be more accurate?

purpose is to measure battery current as the converter actually regulates the discharge current of the battery.
 
1. Probably because the CMRR degrades with higher input voltages, something they don't want to tout. Given your accuracy requirement, don't care.

2. No, less accurate. It's the same opamp and presumably the same tolerance resistors. The difference between 1K and 1M input impedances to a 0.1 ohm source impedance is trivial, and overpowered by the input offset voltage errors of two additional amplifiers. The only instrumentation amplifier designs that are a measurable improvement over a well-done diff amp design are the ones that use fully integrated, laser-trimmed instrumentation amp ICs.

3. With a 10% error budget, an LM358 would work in this application.

ak
 
Last edited:
You could halve the sense resistor value and double the amp gain.
 
AnalogKid said:
3. With a 10% error budget, an LM358 would work in this application.
No. But an LM358 can sense grounded current. a TL081 can sense positive supply current, if its supply is 7V or more.
The error due to offset for an LM358 in the given circuit is 30mV (given worst-case offset of 3mV and gain of 10), which is 30% for a current of 1A (proportional error is worse for lower currents).
 
Worst case input offset voltage occurs under conditions that are not present here. The DC resistances seen by the two inputs differ by only 0.1 ohm. I think that's low enough to consider the typical offset values.

ak
 
Worst case input offset voltage occurs under conditions that are not present here. The DC resistances seen by the two inputs differ by only 0.1 ohm. I think that's low enough to consider the typical offset values.
How do you mean? The resistances could differ by ~10 ohms with 1% resistors... What does that have to do with the input offset voltage?
I think that's low enough to consider the typical offset values.
Typical is 2mV (https://www.ti.com/lit/ds/symlink/lm158-n.pdf), or 20% error at 1A.
 
Worst case input offset voltage occurs under conditions that are not present here. The DC resistances seen by the two inputs differ by only 0.1 ohm. I think that's low enough to consider the typical offset values.

ak
Maybe you do not know that there are two causes of input offset voltage:
1) The difference between the input currents that you are talking about.
2) The inherent input offset voltage of the input transistors even if they are Jfet and have no input current.
We are talking about the second which is 7mV max for the cheap LM358 and most other general purpose opamps but is only 2mV max for the more expensive LM158A and less for other more expensive opamps with the inputs shorted together so the input currents have no effect.

EDIT: Do you know where to buy opamps with "typical" spec's, I don't. If I buy a few thousand at different times and test them all then maybe I will find a few typical ones, if their production yielded some typical ones. The "maximum" spec's are the ones that are guaranteed.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top