Bit of guidance for signal conditioning request

[Captain_Mainwaring]

Hi all,

I have a need for a circuit and thought I would post it onto this forum for some guidance. I have existing circuitry for a scale consisting of a strain gauge, wheatstone bridge and OPA333 op-amp. (I am stuck with this circuit so have to work with it).


I have fed the output of this op-amp into my ADC/microcontroller but the results aren't great, the main issue is with the signal out of the OPA333. (pls see attachment)


When the scales are unloaded the signal level is 0.95v. When loaded the signal level is typically around 1.3v with +/- 0.1v fluctuations for movement (actual levels depending on the animal being weighed and how much they can shift their weight about (the fluctuations)). I need to record the stationary loaded weight and also the variations due to the animal moving around.

However, the issue is with the poor use of range, I am using less than 10% of the signal range (and thus of the ADC range) and about 3% of the signal range for the fluctuations for movement. What I need to do is remove the 0.95v offset, then amplify to make sensible use of the ADC range. eg.


What is the best way to remove a DC offset in a DC signal? An op-amp Comparator?
I would then plan to use a non-inverting op-amp to amplify, is this the best way?

All/any opinions are warmly received.

Ta,

George

 

[barry]

Simply feed the output of your OPA333 into a circuit like this:



The actual implementation will depend on the required accuracy, but basically R2/R1 determines the gain and V+,R3,R4 determine the offset. R3 and R4 also affect the gain, so you want R3 to be very small compared to R1.

 

[Captain_Mainwaring]

Thank you Barry, this worked a treat!!!

Kindest regards,

George

Simply feed the output of your OPA333 into a circuit like this:

View attachment 95746

The actual implementation will depend on the required accuracy, but basically R2/R1 determines the gain and V+,R3,R4 determine the offset. R3 and R4 also affect the gain, so you want R3 to be very small compared to R1.

 

[pinout]

The offset suggestion looks good, but a quick look at your input circuit looks wrong to me.

The gain is being set by some function of the bridge impedance and the feedback resistor, perhaps this intentional, if so I'll try to examine it later.

Normally the bridge feeds a high impedance instrumentation amplifier or its 3 op amp equivalent.

 

[Captain_Mainwaring]

Do you think this sort of layout (2 stage) would be better? The first stage for the level shifting and the second stage for the amplification?

Thanks,

G

The offset suggestion looks good, but a quick look at your input circuit looks wrong to me.

The gain is being set by some function of the bridge impedance and the feedback resistor, perhaps this intentional, if so I'll try to examine it later.

Normally the bridge feeds a high impedance instrumentation amplifier or its 3 op amp equivalent.

 

[jiripolivka]

One problem can be that your circuit has no frequency limit and the fluctuations affect the ADC. Please use an oscilloscope so you can see the output fluctuations before ADC, then use integrating capacitors in your opamp to amplify the wanted slow signal and to reduce noise which you do not need.

 

[barry]

I agree with Jiri, your bridge amplifier is, um, not optimal. If you calculate your transfer function you'll see that it's kind of non-linear. You could use a differential amplifier arrangement to get better performance, without having to go to an instrumentation amp.

The two-opamp approach you show doesn't really buy you anything, other than making the transfer function calculation easier.

 

[pinout]

Hi,
Regarding my comment "The gain is being set by some function of the bridge impedance and the feedback resistor, perhaps this intentional, if so I'll try to examine it later.

Normally the bridge feeds a high impedance instrumentation amplifier or its 3 op amp equivalent." As per recent posts this is regarding your original circuit.

Strain gauges provide a relatively small signal and an Instrumentation amplifier would be the best choice, they provide good operating performance, high impedance inputs and an easily settable gain. If you are going for a low cost option the three opamp instrumentation amplifier is a good second choice, but performance will be let down by the resistor tollerances which are usually set (laser trimmed or similar in the IA version), but this may still be good enough.

A quick check on current products show the Microchip MCP6N11 start at just at under £1.50
If you follow the datasheet link from this page https://uk.rs-online.com/web/p/instrumentational-amplifiers/7559546/ it shows a bridge circuit in 4.4.4.
Texas Inst do the INA126UA at around £1.60, which again is a good price for what it does (this is a two op amp version).

From Microchip data sheet.

 

[crutschow]

Perhaps a two op amp diff amp configuration, which gives a high input impedance on both inputs, would work.