Go Down

Topic: Interfacing a 7 Wire Load Sensor having two sense wires (Read 198 times) previous topic - next topic

atf_atf_atf

Hello everyone, I have been unsuccessful in finding a solution to my problem of interfacing a 7-wire load-cell (strain-gauge type. 2 for Excitation, 2 for Signal/Output, 2 for Sense and a Shield) to my Arduino UNO. I have successfully interfaced and received weight data from it using a HX711 module but have recently realised that the HX711 IC does not support four-terminal sensing.  :smiley-eek-blue:

I have decided to modify the circuit by using an instrumentation amplifier and an 24bit A/D converter instead, as suggested by this topic in the Arduino Forum. This, and the Wikipedia page on four-terminal sensing have been the most helpful so far, but I still have some questions before I dig another hole and jump in...

The Wikipedia page says that four-terminal sensing has 2 pairs of wires " current-carrying" and "voltage-sensing". So, the Sense wires are the former and the Excitation wires are the latter, since these two pairs (Ex+/Sen+ and Ex-/Sen-) are shorted inside the load cell.
BUT
.. the Excitation wires are the ones that supply the stabilized voltage to the Wheatstone-bridge, then how can they be the ones that are the "voltage-sensing" ones? Aren't the Signal wires used for that?

Also, as I understand (correctly, I hope), that supplying a constant current is more important than a constant voltage in order to compensate for wire resistance and temperature changes so that the calibration remains intact over time. Is this correct? If yes, then which pair should I use for that? Sense or Excitation, or both?

Summing up all the above blathering, HOW to interface a 7-wire force transducer to an Arduino UNO?

Any suggestions or ideas are welcome...

outsider

Can you post a link to that sensor's datasheet?

atf_atf_atf

Sorry, this load cell is part of an OEM in line checkweigher in an factory. I have searched intensively throughout the web for any information specific to the cell but there isn't any.

The resistances between the wire-pairs are as in the attached file "ohmPairTable.png", deduced using a digital multimeter.

DaveEvans

The sense wires are only needed if the leads between the load cell and ADC/microprocessor are long, as they allow the micro to account for the voltage drop in the current-carrying excitation wires.

You will find more information if you google six wire load cell  (rather than googling seven wire load cell).

If it is important for your application to use the sense wires, I suggest trying this: http://www.ti.com/tool/ADS1232REF

atf_atf_atf

#4
Oct 17, 2018, 11:16 pm Last Edit: Oct 17, 2018, 11:24 pm by atf_atf_atf Reason: Inclusion of more data related to the problem
Quote
The sense wires are only needed if the leads between the load cell and ADC/microprocessor are long, as they allow the micro to account for the voltage drop in the current-carrying excitation wires.
Thank you so much DaveEvans, so is it safe to continue using the HX711 as per my first design because I need to make sure that the accuracy is maintained across seasons, i.e. temperature changes which might be as high as 25 degrees C from winter to summer? This doubt arises as per this formula from Wikipedia:



It is intended to be used in a food packaging industry where the packages are as small as 25g. Unethical not to consider it I think, even if it is as small a quantity as a fraction of a gram.

Also, the load cell is 15kg C3 class. Sorry, should have mentioned it before.

MorganS

The HX711 IS a 4-wire device, with an extra terminal available for the shield.

I'm not sure why you would need 7 wires except perhaps it has additional sensors in there. It's not uncommon to put a NTC type thermistor onto the strain gauge and use a second channel of the strain gauge instrument to measure temperature. There's a few tricks to make the thermometer read with exactly the same circuit as a strain gauge.

Is there no ability to calibrate it? A special 25g weight kept in a glass case and only handled with white gloves once per season?
"The problem is in the code you didn't post."

DaveEvans

Unethical not to consider it I think, even if it is as small a quantity as a fraction of a gram.
If your application is that important, you should be getting advice from an industrial weigh scale consultant, not a hobby forum.

atf_atf_atf

#7
Oct 17, 2018, 11:52 pm Last Edit: Oct 18, 2018, 12:10 am by atf_atf_atf Reason: Added a point
Quote
Is there no ability to calibrate it? A special 25g weight kept in a glass case and only handled with white gloves once per season?
Yes, I could do so. The problem is that the complete work involves multiple machines (I dont know the precise number but its atleast >30) spread across hundreds of kms (>450 ) across 14 factories.

And then comes Mr. Bureaucracy.... :smiley-sad-blue:

Quote
If your application is that important, you should be getting advice from an industrial weigh scale consultant, not a hobby forum.
You are right about this. I have already contacted one. The reason I posted it here was because I was using Arduino, also because the forum has brilliant people and far quicker answers.

Quote
It's not uncommon to put a NTC type thermistor onto the strain gauge and use a second channel of the strain gauge instrument to measure temperature.
Thanks MorganS, this combined with DaveEvans post is what I needed.  :D

DaveEvans

The resistances between the wire-pairs are as in the attached file "ohmPairTable.png", deduced using a digital multimeter.



It doesn't make any sense ( :) ) for the resistance between the two excitation lines (green and black) to be zero.  It also doesn't make sense for the resistance between the two sense lines (blue and grey) to be zero.    I think you screwed up the colors and names.

It would make sense for the resistance between one of the excitation lines and a sense line to be zero, and zero resistance between the other excite and other sense.

Go Up