strain gauge calibration

1. strain gauge calibration

Hi All,

Im new in the strain gauge world.

Im looking for a low cost strain guage that can feel 1g to 1kg tension that preformed on a small aluminum plate.

I understand that there is a several methods to sense the change in the resistant of the strain guage. the most common is a Weston bridge.( if there is a better way ill be to get some information)

Can some one recommend me what is the best to use ( half bridge, full bridge ?) , and also how to calibrate the bridge? ( if there is a IC that do it it will be grate)

thanks

•

2. Re: strain gauge calibration

Hi. Try and understand your theoretical concept about weaston bridge circuitry. Note: there are reference point which can be used for calibration. It's an analogue circuit which it output can feed to an instrumentation amplifier or monitor with your multimeter. You can also go microcontroller to process the analogue sensor reading and display result on LCD or pc.

3. Re: strain gauge calibration

Hi,

If you need temperature stability, then a full bridge is best.
There are signal conditioning ICs especially made for strain gauge applications.

What do you want to calibrate? Gain, offset? In whag range?
What precision do yo want? ( Yes, I know you like to answer: The best one can achieve.. But with this answer nobody can calculate. We need values)

Klaus

4. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

If you need temperature stability, then a full bridge is best.
There are signal conditioning ICs especially made for strain gauge applications.

What do you want to calibrate? Gain, offset? In whag range?
What precision do yo want? ( Yes, I know you like to answer: The best one can achieve.. But with this answer nobody can calculate. We need values)

Klaus
Hi Kalus,

Thank you for your quick respond.

I read that before i start to work with the bridge i need to do a calibration that the V+ and V- differential voltage will be 0V when is no tension. im looking for a low cost small size signal conditioning ICs .

regarding the precision, i dont need to sense the exact tension i need only to sense the change in tension ( if tension applied or not- with threshold )

i hope i explained my self correctly .

thanks for the help :)

5. Re: strain gauge calibration

Hi,

Two different things.
i need to do a calibration that the V+ and V- differential voltage will be 0V when is no tension
This means laser calibration of the applied strain gauge. Very difficult and high effort.

i need only to sense the change in tension
This is totally different. Use any ADC, measure the initial value and store this value in an EEPROM.

You need to define what accuracy, what precision and what resolution you want. (No text. Values)

How many sensors? How often do you need/read/store a value? Mobile or stationary solution. What power supply, what communication?
Tell us more.

Klaus

6. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

Two different things.

This means laser calibration of the applied strain gauge. Very difficult and high effort.

This is totally different. Use any ADC, measure the initial value and store this value in an EEPROM.

You need to define what accuracy, what precision and what resolution you want. (No text. Values)

How many sensors? How often do you need/read/store a value? Mobile or stationary solution. What power supply, what communication?
Tell us more.

Klaus
Hi Kalus.

thanks for the help.

My main goal is to check if tension that preformed on a small aluminum plate.my product is mobile.and i only need one sensor because the aluminum plate is small (~ 4X4 cm) -the tension that preformed is small.

i want to sample the strain Gage every 200mS. with an arm micro-controller( my micro controller have ADC inside). I working with li-on battery 3.7V.

i dont need to measure the exact tension that preformed- i want to set a voltage threshold (1 or 2 values) and to have the ability to know if the tension pass each one of the thresholds ( the thresholds not close to each other- so i dont need an exact measurement)

for example my ADC ( in the micro controller) is 12bit and the analog pin can have 0-3.3V
so if i divide 3.3 to 3 thresholds i get 1.1V -> so i can have 3 thresholds :
0-1.1V - state 1
1.1-2.2V - state 2
2.2-3.3V - state 3

( the voltage that i measure on the bridge)

i want to know what stage im in in the bridge.

i read that the strain gage with a bridge have an drift and temperature issues. So I have to circuit knows how to overcome these problems.

thanks for the help

7. Re: strain gauge calibration

If this can help. You can buy any cheap wheighing scale and *****-open it. There's 8 or 16 strain gage in it, so you can see how they are build.
For calibration, I also recommand full bridge for temperature compensation.
For setting V+=V-, you can use a potentiometer. Connect one pin to Vdd, one to GND and the center one to V+. This is how I do, but maybe there are better ways of doing it, so I am also interested by this topic.

Concerning the precision, you say you don't need an exact measurement but still you need to set the treshold value. So how precisely do you need to measure this treshold?

8. Re: strain gauge calibration

Hi,

* With three states: you need no precision. You even donīt need to worry about temperature drift.
* With a microcontroller you donīt need to calibrate on the analog side. Do it with software. It is more easy.
* Just use a simple OPAMP circuit to amplify the signal. Add a C for low pass filtering.

Klaus

•

9. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

* With three states: you need no precision. You even donīt need to worry about temperature drift.
* With a microcontroller you donīt need to calibrate on the analog side. Do it with software. It is more easy.
* Just use a simple OPAMP circuit to amplify the signal. Add a C for low pass filtering.

Klaus
Hi Klaus,

thank you for your quick respond.

so if i understand you correctly:
i need to take a full bridge ( 2 strain gauges) . connect the bridge power supply to my 3.3V LDO. the V+ and V- bridge pins in need to connect to a differential OP AMP. and the output of the op amp i need to connect directly to my micro controller ADC pin.?

thanks

10. Re: strain gauge calibration

Hi,

some questions:
* what is the max expected strain gauge output signal range (including offset)?
* what is the (decoded) max. ADC input range ("differential" input range, if possible)?

Klaus

11. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

some questions:
* what is the max expected strain gauge output signal range (including offset)?
* what is the (decoded) max. ADC input range ("differential" input range, if possible)?

Klaus
Hi Kalus,

*The STM micro controller has a 12bit ADC - the ADC is not differential.
*I can not define it now- i have a 3.3V LDO that power the board so i can supplay to the bridge 3.3V the differential voltage that goes out for the bridge can be in the range of the input voltage of the differential OP AMP .
*The max ADC read can be the STM micro controller VDD that currently is 3.3V.

I thought to use something like this:

do you think it will work?

12. Re: strain gauge calibration

Hi,

your circuit is incomplete. There are no gain setting resistors.

Could you use a second ADC_input to act as differential input? (connected and read_out independently- interleaved-, but subtracted in your software)

There are many possible solutions.
For a single ADC input solution maybe an instrumentation amplifier is a good choice. Set the output offset to ADC_range/2.

***
In any case:
* use low pass filters for the strain gauge signal
* and for the VRef
Both need to have the same characteristic and fc.

Klaus

13. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

your circuit is incomplete. There are no gain setting resistors.

Could you use a second ADC_input to act as differential input? (connected and read_out independently- interleaved-, but subtracted in your software)

There are many possible solutions.
For a single ADC input solution maybe an instrumentation amplifier is a good choice. Set the output offset to ADC_range/2.

***
In any case:
* use low pass filters for the strain gauge signal
* and for the VRef
Both need to have the same characteristic and fc.

Klaus

yes i can use 2 ADCs on the micro controller and to make them to act as differential.

http://www.ti.com/lit/an/sloa034/sloa034.pdf

also:

do you think it will be a proper solution to my PCB?

thanks

•

14. Re: strain gauge calibration

Hi,

in sloa034.pdf see Figure5.
(Maybe use two equal resistors with twice the initial value of R3. Connect one to GND and the other to VRef)

I think this is a simple and good solution. Using only one ADC input.

Klaus

15. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

in sloa034.pdf see Figure5.
(Maybe use two equal resistors with twice the initial value of R3. Connect one to GND and the other to VRef)

I think this is a simple and good solution. Using only one ADC input.

Klaus
Hi Klaus,

i found some components that i want to implement in my circuit ill be happy to read your feedback :

1. i found a low cost full bridge strain gauge:

2. For the instrumentation amplifier i have 2 options :

2.2 TI :

or

the circuit will be as you recommended :

thanks for the help

16. Re: strain gauge calibration

Hi,

I canīt say anything about the strain gauge.
But the electrical parameters seem to be good.
Expect a 10mA of strain gauge current. Making 33mW of disspated power.
***
I didnīt check them. But your application ist not that critical, so I assume all will work.
***
Yes this is a good solution.
To suppress 3V3 noise you need to add a second LPF from VCC to VREF.

Only if both filters have the same characteristic the errors are compensated.

An unfiltered VRef generates the same error as an unfiltered signal.

good luck.

Klaus

17. Re: strain gauge calibration

Originally Posted by KlausST
Hi,

I canīt say anything about the strain gauge.
But the electrical parameters seem to be good.
Expect a 10mA of strain gauge current. Making 33mW of disspated power.
***
I didnīt check them. But your application ist not that critical, so I assume all will work.
***
Yes this is a good solution.
To suppress 3V3 noise you need to add a second LPF from VCC to VREF.

Only if both filters have the same characteristic the errors are compensated.

An unfiltered VRef generates the same error as an unfiltered signal.

good luck.

Klaus
Hi Kalus,

For better power dissipation i can work with 1.8V ( all the components can work with this voltage) but it will cause a minor change in voltage( when tension applied on the bridge) at the bridge output and ill be more sensitive to noise.

regarding the LPF please chack if i added the second LPF (VCC to VREF) currently.

how do i choose the LPF( C and R ) values ?

A general quotation. you told me that i will not need to calibrate the bridge . but if some reason ill need to do it there is an automatic way to do it? or an IC that do it?

18. Re: strain gauge calibration

Hi,

For better power dissipation i can work with 1.8V ( all the components can work with this voltage) but it will cause a minor change in voltage( when tension applied on the bridge) at the bridge output and ill be more sensitive to noise.
Because the ADC_Vref is 1.8V also the digital output of the ADC wonīt change. --> The ADC value is independent of VCC, it is proportional to the tension.
You may see worse signal-to-noise. I donīt think itīs worth considering it.

***

second LPF:
Not correct, because now the unfiltered 3V3 go to the ADC_VRef.
But the ADC_VRef needs to be filtered.

so: 3V3 -> R -> C -> GND.
The voltage of the C goes to ADC_VRef and the same voltage goes to the voltage divider.

***

you told me that i will not need to calibrate the bridge
Thatīs what i wrote:
* With a microcontroller you donīt need to calibrate on the analog side. Do it with software. It is more easy.
--> if you want to calibrate for offset, then just subtract a constant value. Usually stored in the EEPROM. It can be initiated with a key press.
--> if you want to calibrate for gain, then just multiply with a constant value. Usually stored in the EEPROM. It can be initiated with a key press.

Klaus

--[[ ]]--