Load cells

[PLC]

I have a beam with a load cell on each side (under each end)

If the maximum weight being pulled down on the beam is 1000lbs


Do you buy two 500lb load cells or two 1000lb ?

Asking for a friend

 

[James Mcquade]

what is the application?
for example, if you are dumping a load into a hopper, you must take the load shock into consideration. i do not think an accurate answer can be given without knowing the application details.
regards,
james

 

[ASF]

This sounds like a very mechanical problem, but if it were me I'd be buying two 1000lb load cells.

I know that when you're rigging a load using a four chain assembly, you have to size the chains such that any two are rated for the entire load, because as the load shifts and swings, most of the load will be observed by two chains at a time.

Similar thing here. Unless the load is exactly and perfectly centered between the load cells every time (which will never happen), one load cell will be exceeding its 500lb limit.

Depending on how uneven or unbalanced the load could get, you could possibly get away with e.g. two 750lb load cells. But at that point, why waste time trying to pinch a few pennies and risk losing thousands of pounds (pun fully intended) when practice doesn't quite match theory? Just do it properly to start with and then you don't have to worry.

 

[PLC]

Ok
Just say for a weight, do you get 2- 500 lb ones or 2 - 1000 lb

What’s best practice?

Hypothetically speaking
For accuracy

 

[ASF]

It's not my field of expertise, but my very-slightly-educated guess is that the accuracy difference between 2x500lb and 2x1000lb load cells will be so infinitesimally small as to be completely outweighed (pun fully intended again) by inaccuracies caused by general installation conditions, like termination resistance, length of cable run, interference, alignment of the load cells, accuracy of the calibration weights, whether the load cells are 4-wire or 6-wire, accuracy of your A/D converters if there's any sort of 4-20mA or 0-10V analog output/input involved, and so on. And, of course, you'd have to consider whether repeatedly overloading one of your load cells will eventually lead to more inaccuracies.

I mean, if you were saying "should I use 750lb load cells or 75,000lb load cells", then sure, I should imagine there would be a loss of accuracy if you went the 75,000lb route. But 500lb to 1000lb? Not really.

What you're really changing is not the accuracy, but the resolution. A load cell has a mV/V rating, typically in the range of 1-3mV/V. Let's consider a 500lb load cell with a rating of 2mV/V, connected to a weight transmitter with an excitation voltage of 10V (which is fairly common). A rating of 2mV/V means that the signal output by the load cell will vary by 2mV for every V of excitation voltage, across the full rated load. So, the difference between no load and 500lb is 2(mV/V) * 10(Excitation V) = 20mV. This means each lb gives a difference of 0.04mV.

If you consider the same setup but with a 1000lb load cell, it's now only 50% loaded at 500lb, so the difference between no load and 500lb is 2(mV/V) * 0.5 (50% load) * 10 (Excitation V) = 10mV, so each lb gives a difference of 0.02mV

Now consider a 50,000lb model. At 500lb it's only 1% loaded, so the difference between no load and 500lb is 2(mV/V) * 0.01 (1% load) * 10 (Excitation V) = 0.2mV. This means each lb gives a difference of 0.0004mV.

From the perspective of your weight transmitter, it's just converting a mV signal to a digital signal internally, which is then displayed on the screen and output via 4-20mA or 0-10V or some kind of fieldbus. It has an A/D converter to convert this mV signal to an internal number, and that A/D converter will have a certain resolution. Let's say that the resolution is 0.0005mV ( a figure that I pulled out of thin air for the purposes of illustration).

So, for your 500lb load cells, where each lb = 0.04mV, your maximum resolution - i.e. the smallest change in weight that you can measure - is 0.0125lb.
For your 1000lb load cells, where each lb is 0.02mV, your maximum resolution is 0.025lb.
For your 50,000lb load cells, where each lb is 0.0004mV, the smallest change in weight that you can measure is 1.25lb.

Now in practice this looks like the 50,000lb load cells are inaccurate, because you can put an extra pound of weight on the scale and it won't change. But the fact is they're accurate, they're just not as precise as the 500lb or 1000lb in that range.

Again, comparing between the 500lb and 1000lb options, they will both be as accurate as your installation conditions allow, and the inherent difference between them in accuracy will almost certainly be so negligible as to not make any practical difference. If the resolution is of critical importance to you, then you would have to carefully select a solution based not just on the rating of the load cells, but also on the type and resolution of your weight transmitter.

Again, I'm not an expert on this. I've just messed about with load cells enough to absorb a bit of information about how they work. Take all of the above with a grain of salt.

One final anecdote about how equipment selection is only one very small part of a weighing solution. I was commissioning a batch weighing system where batch weights ranged from 150g to 30kg. Acceptance criteria was +/- 2%. +/- 2% on 30kg is over half a kilo, so we had no problems whatsoever with the larger batches, but +/- 2% on 150g is 3g.

We struggled badly to get the required accuracy on the 150g batches - we'd get two in a row, and then the next three would fail. Until I noticed something. Across the room, about 20m away, there was a fan on the wall to keep the operators cool. It oscillated back and forth slowly. If it was pointing at our equipment when the 150g batch when through, it would fail every time. I turned the fan off, and we put through nine 150g batches in a row to within 1g. On the tenth batch, a forklift drove past and the batch failed again.

All that goes to say - weighing is an inherently complicated task, and when you look at all the variables in play, the difference between a 500lb load cell and a 1000lb load cell is often completely academic from the perspective of "how accurate will my entire solution be"

 

[Steve Bailey]

How sure can you be that the two 500 pound load cells won't be overloaded? Can you be sure that the load will always be equally distributed between the two cells? Because with two 500 pound cells and 1000 pounds of weight, you're giving yourself no margin for error beyond what the manufacturer of the load cells builds in. I personally would not cut it so close by using two 500 pound cells.
Sure, a pair of 1000 pound cells may sound like excessive overkill and you will lose some precision in the weight. Are you using a signal conditioner to convert the millivolt signal from the load cell to 4 - 20 mA for a PLC analog input? If so you may be able to scale the signal conditioner signal over less than the full range of the load cells.

Finally, the only choices available can't be just 500 pound or 1000 pound cells. A pair of 250 kg cells would give you more margin than a pair of 500 pound cells.

 

[PLC]

Thanks
I found what I was looking for in a Mettler Toledo example

 

[parky]

Most definitely you need loadcells at least 1.5 times the max weight, it will depend on how many loadcells for example a batching tank of 1000kg capacity with the standard configuration of 3 cells (often 3 is better as 4 can cause bab readings unless the legs are balanced), you have to take into account of the weight of the vessel or what ever it is for example take a 1000kg capacity tank, the empty weight could be almost as much as it's capacity so although in your code or setup of the loadcell interface you will require zeroing, so already the loadcell is reading the weight of the vessel, add to that the weight of the product or force plus an allowance for instantaneous shock you would probably need 3 1000kg loadcells.

 

[cardosocea]

Ok
Just say for a weight, do you get 2- 500 lb ones or 2 - 1000 lb

What’s best practice?

Hypothetically speaking
For accuracy

The capacity of them would have a small effect in comparison with how you use them... the weight transducer, whether they'd go on a 4-20mA loop, etc...

 

[parky]

I just so happens that a collegue of mine did a project in the late 80's, this was a shear beam test rig, so something probably similar to your system, the load cells (2 off) were each rated at the capacity of the max load to be applied, so in effect if the load was 5000 kg then each loadcell was 5000kg, a total of 10000kg. there were tests for non destructive & destructive test so either load it to the weight required for a pass & load it to destruction.
I have worked on many batching vessels and although the mechanical design was not in my scope in every case in 3 point setting each loadcell was at least the capacity of the vessels in kg. so although not an actual expert it seems that this is the case. Not sure on your application but as it seems to be that there will be a standing weight i.e. the weight of the beam or supports as well as the load that is to be applied, so 500lb loadcells I think would almost certainly over scale & probably fail, idealy, you need the max test weight to be somewhere in the middle to 75% of the loadcell capacity.

 

[PLC]

This is what I was looking for
... an example

 

[parky]

Yes that is correct, so that on 4 loadcells you are looking at 2 so assume a load of 1000 lb & what ever vessel weight is plus 1.25 divided by 2, however, you stated that this is some sort of beam system, is there a possibility of sudden shock i.e. if it is a shear test or sudden load then you need to factor that in as well, we had 6 vessels ranging from 300 to 1500 kg capacity, the shock of adding ingredients was none existent except on the 1500kg vessels, certain ingredients were dropped in via a lift & bin the bin weight could be 400kg the shock of the sudden addition was causing loadcell failure on odd occasions this only happened on this vessel as it was the only one with the large bin addition, the original loadcells were 1500kg x 3 this was actually more than was required for a vessel of that size, as were the other 1500kg vessel, we ended up fitting 2000kg cells this cured the problem so bear this in mind, also one other little warning, if there is a possibility of the load being present on only one loadcell then this needs to be taken into consideration, our other systems with 2 loadcells (weighing bins) although these were 500kg loadcells with a product weight of 500kg & empty weight of 200kg would fail on some occasions due to a design fault where one loadcell would not be balanced correctly, engineers would not check this & re-calibrate using 400kg of weights, the weight therefore was on only one loadcell of 500kg these often fail or at the very least saturate as it was reading above the rated total weight this gave rise to overfilling of the vessel as it would not reach the final weight.