0

I'm developing a project where an op-amp is configured to bring a input voltage difference ranging from 1.8V to 3V to a output voltage ranging from 3V to 5V. The output of the op-amp will be fed into a ADC input of a MCU. the circuit looks as follows:

schematic

simulate this circuit – Schematic created using CircuitLab

Taking following considerations:

  • The op-amp will be powered in a single supply configuration with a voltage of up to 36V.
  • In- can range from GND to Vcc-1.8V.
  • In+ can range from 1.8V to Vcc.
  • In+ - In- can range from 1.8V to 3V.
  • The input signal will be really low frequency, there will be hardware and/or software low-pass filtering.

How can I find a op-amp based on it's Common-mode input voltage range?

Am I missing something or making something wrong?

For transistors there is the alltransistors page where you can make a parametric search, I need something like this but for op-amps.

What I have so far is:

At the lower end of the input range I would have In+ = 1.8V and In- = 0V this gives that the common-mode input of the op-amp should be 1.8*20/(12+20)=1.125V.

At the upper end of the range I see plenty of room for the Common-mode input voltage range.

But from the datasheet of a typical UA741 page 6 the Common-mode input voltage range VICR is ±13V in a ±15V test condition. This means that my 1.125V figure is outside it's input range.

An other way to solve this would be to connect R3 to some fixed voltage reference instead of GND and re-think the values of the 4 resistors. Adding a extra voltage reference would add more components to the board, increase size, cost, power consumption and make it more vulnerable to disturbances.

Falk
  • 201
  • 2
  • 13

2 Answers2

1

If you need to only use a single supply as shown but still must be able to handle inputs which are close to ground, I would suggest looking for a rail-to-rail op amp, which are specifically designed for their inputs and/or output ranges to swing very close to both supply voltages. There are R-R input op amps, R-R output op amps, and there are both (typically referred to as RRIO op amps).

However, all this to say that most modern op amps, even non R-R input ones, should be able to handle the common mode input you describe, which is over a volt above ground. It sounds like you might want a precision DC op amp for your purposes, and there are lots of good options. Chopper/autozero amps effectively get rid of their intrinsically-large low-frequency noise & drift through internal switching methods which essentially shift the 1/f noise content to the switching frequency, making them excellent for precision DC measurements.

I’d suggest a part like the LTC2057HV, which is a great low-noise, precision DC op amp with a wide supply range and a CM input range that actually can handle even 100mV below its negative supply! The data sheet is here: https://www.analog.com/media/en/technical-documentation/data-sheets/2057f.pdf

Jack Davis
  • 11
  • 1
  • I don't need exactly a rail-to-rail op-amp, a margin of 1V would be acceptable. The LTC2057HV is quite expensive for my budget. What about how to find it? how did you find it? – Falk Jun 07 '20 at 08:57
  • Hey Falk, point taken, it’s a bit pricey for sure. I just recall the part from using it before at work. A great place to start looking is at analog’s catalog of op amps, which they have separated based on application/usage. Assuming you care about low-frequency performance, I’d take a look at their precision op amp table here: https://www.analog.com/en/parametricsearch/11094#/ for instance, the first op amp on that list, the ADA4523-1, is under a buck and it looks like its common-mode range would be suitable for your purposes. – Jack Davis Jun 07 '20 at 15:31
  • Thank you Jack, ADA4523-1 looks suitable for my project, I will consider using it. – Falk Jun 07 '20 at 15:53
1

Many modern opamps that do not have rail to rail inputs have a common mode range from V- (or slightly below) to (V+ - 1.5). This is because the input structure is a PNP (or P channel) pair and we have come a long way in the design of these devices.

As you are measuring a low frequency signal, you may need to consider flicker (1/f) noise. Zero drift amplifiers can eliminate this noise.

Most precision amplifiers have a common mode range that extends to V- (you can see some of them in this table) (Most of the usual suspects have such devices - this is a typical table).

One that stands out is the ADA4523-1 which has pretty impressive specifications and much less expensive than the LT2057HV.

Looking at TI, there are plenty of offerings with budgetary (1k) prices below $1.

The table has a rail to rail column which specifies both the input and output and most of the items here have input to V-.

Rail to rail inputs can have their own very unusual problems if driven by a high impedance source as many of these devices have overlapping input pairs (a PNP and a NPN pair) and if the input is not actively steered the crossover point is not particularly well defined. I have experienced low level oscillations with some of these devices when the input was near the (thermally sensitive) crossover point.

If you do not need full rail to rail input performance it is best not to use one.

Before choosing one, check to see if it is available in distribution - the most expensive part is the one you cannot buy.

Peter Smith
  • 21,923
  • 1
  • 29
  • 64