Hi,
I’m looking to do a project that measures the weight of an object. I understand from reading various articles that a compression load cell could do the job. I’m very new to all this. The first thing I’m struggling with is I make the correlation between a load cell and the maximum weight it can measure.
What should be I looking for when reading through the datasheets. One I found mentions ‘Range’’ and gives an lbf value?
Any pointers appreciated?
First, how about sharing a link to the data sheet? Or the supplier you’re looking at …
Many load cells require an excitation voltage to be applied and return tiny voltage changes millivolts/volt of excitation such that for industrial load cells they need signal conditioners.
LBF often refers to ‘pounds force’.
There are many ways to weigh something.
How heavy is it?
How accurate do you need it to be (please don’t say as accurate as possible – that’s not a real answer)?
How much can you spend?
Ok thanks, here is a link to the one I had in mind:
http://www.mouser.com/ds/2/261/FX1901-8147.pdf
The object I’m trying to weigh (bee hive) can be up to 150kg. What I’m particularly interested in is getting an initial measurement and then monitoring the deviation + / - to this. Ideally I’d like to measure this in grams.
The hive itself sits on 4 wooden legs, the simplest solution I guess would be to have a load cell under each leg, as this is only a hobby that particular approach looks to be cost prohibitive, so not entirely sure how you’d get the weight of something with 4 legs and centre the load to single point?
This project is still in the early stages and my background is more software, however just looking at it one step at a time
Other things that I need investigate are drift, if you stick a 50kg weight on a load cell and come back in month would it still report 50kg (assuming constant power supply). Also this would be outside albeit in fairly moderate climate would that impact it?
I also not sure how multiple loads cells are connected. Found this online example for connecting a single load cells
http://edg.uchicago.edu/tutorials/load_ … ematic.png
If you connect two does it mean that you need another amplifier and that then connect that amplifier to another analog port, then leaving you to combine the two readings in software?
As for cost, I already have Arduino Uno, so for the rest I’d to keep it under £100 all in.
Happy to receive any suggestions / pointers
Measurement Specialties has an amplifier that will connect up to 4 load cells in parallel. I don’t know how this works.:
http://www.meas-spec.com/product/t_product.aspx?id=9524
I would think you would need to at least put an inverted cup over the load cell to try an protect it from the elements (or contact support for environmental capabilities). They are most often used in industrial applications but those environments can get pretty icky.
Drift (as in aging) would be addressed by recalibrating the amplifier. Simple using a shunt resistor. Nearly ALL strain gauge amplifiers have that capability.
You could possibly sum all the inputs with a summing amplifier, but these days it’s more useful to use an A/D converter with multiple inputs.
The first hurdle you’ll run into is that measuring a 150kg mass with a 1g resolution will need a minimum of 1:150000 scale. The Arduino’s onboard analog inputs aren’t remotely capable of that, so you will need an offboard 18-bit A/D converter at least. I would rethink that level of resolution: first, because you say your background is in software and you are getting deep into low-level analog measurements which are far from easy, and because it will likely be swamped by noise, vibration, swaying in the wind, etc.
The actual steps are pretty straightforward: you need a reference voltage (may be supplied by the ADC), a couple of instrumentation amplifiers to handle the bridge inputs and 4 channels of multiplexing, and then an 18-bit or higher ADC. In some cases, the reference and differential multiplexers will be built into the ADC so it becomes a single-component design. That at least makes it easier.
I’ve done some further investigation. Came across a picture that look to be offering something close to what I want. I guess the load cell here is a beam one?
https://www.dropbox.com/sh/0idzsdf4dc6v … 2.JPG?dl=0
I intend to purchase one of these:
http://www.zemiceurope.com/Products-Sin … t-l6e.html
Which can take a 200kg load and try to use it in the manner shown in the picture. I guess you screw one end to the frame and then bolt a tray on to the other end. Then the weight on the plate causes stress in the bar which translate to a change in resistance.
Based upon that load cell can someone suggest a decent amplifier? This one seems to be used in a few examples I’ve seen:
Odds are, any Burr-Brown (I think TI owns them now) instrumentation amplifier will be fine. But don’t focus on any one component, look at the system. Before you can decide how much gain you need, you have to figure out how you’re going to read the signal. As I mentioned, your application the way you describe it needs at least 17 bits of data conversion. Have you selected an A/D converter yet?
Also, I’m pretty sure you’ve already blown your budget with the one in that first picture 
The only way you are ever going to compensate for the drift over these time scales is with two measurements. One with the load applied and one with zero load. This will let you remove the inevitable offset errors.
Many Sigma-Delta converters include programmable gain amplifiers on the front end that are more than capable of handling your signal. If you use the same voltage for its reference as you use to excite the sensor, you will remove yet another source of error. The data sheet for the AD7781 shows an example on page 14.
