Load Cell Ratings?

[Lance]

I'm constructing a force measurement system, and I'm trying to pick a load cell that will do the job.

I'm taking a look at the following load cell:
http://www.meas-spec.com/downloads/FC22.pdf
http://search.digikey.com/us/en/products/FC2231-0000-0050-L/MSP6950-ND/809396

which I found by entering "Load Cell" into the search on Digikey. I've never really gone shopping for a load cell before, and I'm confused as to the units used to define the force ratings. It lists standard ranges as 0-10, 0-25, 0-50, and 0-100 lbf. I don't understand why the units are in foot-pounds, since that is a measurement of torque, and the cell from what I see moves linearly. I'm also failing to see how those ranges work.

Digikey also lists the "operating force" as 22.68kgf. Besides the odd mix of units, there is no "operating force" explicitly stated in the data sheet. Can someone please explain to me how load cells are rated, and what those ranges are?

[Rufus]

The 'f' in those lbf and kgf is for force not foot.

The force due to gravity on a mass expressed in pounds or kilogrammes.

1 lbf is about 4.45 Newtons, 1kgf about 9.81 Newtons.

[Lance]

So when it says that the operating force is 22.68kgf, does that mean that it takes 22.68kg to register 9.81 newtons of force on the load cell?

[Jimmy]

What sort of force you trying to measure you may want a strain gauge not a loadcell

[daedalus]

No, When it says 1kgf, it means that the sensor range is up to 9.8N (1*g). It is common for transducers for weighing applications to state their full scale range in weight units.

I only looked at the datasheet for FC22, so here goes:

This part comes in several configurations, so assuming you are going for unamplified:

Part is available in the following ranges:
0 to 10 (so full scale load is 10lbf)
0 to 25
0 to 50
0 to 100

Assume you buy the 0 to 10 part

10lbf = 44.48 newton

Span = 20mV/V
assuming you are powering the bridge with 3V3, then the output at 44.5N will be 66mV.

so ignoring the errors for now you have around 1.48N/mV response

Now assuming you have an ADC with an input range of 0-3v3, then you want a gain of 45 in order to get full scale output at max load.

assuming 12 bit ADC, your resolution works out around 10mN/lsb. As for accuracy, 1% of span on the bridge, plus whatever nonlinearity there is in your amplifier+ADC.

Thats my back of the napkin analysis, the rest is up to you.

[amspire]

So when it says that the operating force is 22.68kgf, does that mean that it takes 22.68kg to register 9.81 newtons of force on the load cell?

22.68kgf means it is a 50lbf load cell.  50lbf = 22.67962 kgf. Close enough.

22.68 x 9.80665 = 222.41 N

So  222.41 N or 50lbf or 22.68kgf will produce between 3.88V and 4.12V on the load cell on the particular Digikey part you specified.

After you calibrate your circuit, you should get better then 1% accuracy.

Richard.

[Rufus]

So when it says that the operating force is 22.68kgf, does that mean that it takes 22.68kg to register 9.81 newtons of force on the load cell?

22.68kg = 50lb.  As if imperial units were not confusing enough without mixing in metric.

Load cells are rated for a maximum force. Exceed the maximum force (and usually not by much) and the bit that strains gets bent permanently and the cell is trash.

The cell output is specified in volts per volt of excitation applied to the bridge at maximum rated force. Typically 2mV/V. The bridges are usually specified at 10v excitation, a compromise between bridge resistor dissipation (and temperature associated errors) and getting a large output signal.

Load cells need mechanical overload protection (a few have it built in). Usually mechanical stops to limit travel combined with springs or bevelled washers which 'give' at just over the rated force to greatly increase the travel and make mechanical stop design and setting much easier.

[amspire]

The cell output is specified in volts per volt of excitation applied to the bridge at maximum rated force. Typically 2mV/V. The bridges are usually specified at 10v excitation, a compromise between bridge resistor dissipation (and temperature associated errors) and getting a large output signal.

The particular Load Cell on Digikey has a built in amplifier so its nominal full load output is 4V.

However, you are much better off using the non-amplified version that Rufus is describing as it reads accurately down to 0. The amplified cells are useless in the bottom 20% of their range.

Richard

[Lance]

So if I'm understanding this correctly, 22.68kg is the maximum force that can be applied to this cell, and it will give a reading of between 3.88 and 4.12V when at this load.
So is the span just the maximum applied load? I don't see a definition for this term on the sheet anywhere. I'm also confused when it comes to the mV/V unit...what is that ratio?

Assume you buy the 0 to 10 part

10lbf = 44.48 newton

Span = 20mV/V
assuming you are powering the bridge with 3V3, then the output at 44.5N will be 66mV.

so ignoring the errors for now you have around 1.48N/mV response

Now assuming you have an ADC with an input range of 0-3v3, then you want a gain of 45 in order to get full scale output at max load.

assuming 12 bit ADC, your resolution works out around 10mN/lsb. As for accuracy, 1% of span on the bridge, plus whatever nonlinearity there is in your amplifier+ADC.

I'm unsure about how that calculation was made, mostly because of the mV/V unit.

[amspire]

The Volts and mV stuff  is talking about the non-amplifier version.. The more volts you apply to the resistive bridge, the more mVolts you get out.

The span is the reading at the rated load.

If you do use  the version with the inbuilt amplifier, remember that below about 20% of the full load, the reading is useless.

Richard

[Rufus]

The Volts and mV stuff  is talking about the non-amplifier version.. The more volts you apply to the resistive bridge, the more mVolts you get out.

The amplified version is still ratiometric and should have been specified at about 800mV/V. The 4v full scale (or span) output is for a 5v supply (specified at the top of the table). I doubt the amplified version is useless at low outputs, it just has a large zero offset which is also ratiometric and not included in the 4v span specification.

The data sheet is a bit crap. I would expect on 5v supply the amplified version will swing between about 0.5v and 4.5v for zero to rated applied force.

[Lance]

Thanks for the clarification. I see on the connections diagram it specifies the connections for both versions. Under the ordering information, the output space is "3", which means 0.5 to 4.5V. So this is  the amplified version?

It looks like the FC22 doesn't come in an unamplified version, at least not that digikey carries. Why exactly are the amplified cells useless at the bottom 20% of their range?

[amspire]

I was probably wrong there. Misreading the data sheet.

I think Rufus is correct - it reads down to zero pressure.

Richard

[Lance]

I was probably wrong there. Misreading the data sheet.

I think Rufus is correct - it reads down to zero pressure.

Richard

So the value is fine in the bottom 20%?

[shadewind]

I don't really understand why they don't just use Newtons since that's what is really being measured here...