08-27-2012 12:58 PM
The cal values are read directly off of the cal sheet for the cells. They are 1~mV/V. These cells are only for compression - no tension. So are you recommending that I eliminate the middle calibration points and use 0lbs. = 0 mV/V and 1000lbs. = 1>103...mV/V?
Also do I need an output step to the Zero Offset step and then program an input into the Zero Offset?
08-27-2012 01:00 PM
I will give these a try. The excitation is 0-10V.
08-27-2012 01:26 PM
OKors,
The same thing is going on with your program as well as mine. I made the suggested changes to mine still no dice and I used yours with no success. It there a step that I am need to do to send the signal to the zero offset or is this automatic in the programming?
08-27-2012 01:48 PM
Its automatic when dealing with one signal type like you're doing. In the zero step input/output tab your input signal should be the Force Bridge channels.
There's nothing more you need to do at that point.
Your signals of interest at that point appear as "Zero Offset Data".
So in the zero offset box on the left you'll see the input signal, you can click the + and see the channels. Below that is the Zero Offset Data. This is the data where the zero has been applied.
At this point I'm wondering if its a hardware issue, because even with no sensors attached to my device I can see a signal... its bogus but its there, it shows up as a sinewave because there's nothing connected but I can still zero. The zero button is at the top of the Zero Offset Step Setup window above the graph.
Take a look at this screenshot. It was a couple seconds of pre-zeroed data then you can see the zero shift. This is when I clicked Zero Now.
I've included another proj file also. It includes an Amplitudes and Levels step because sometimes the data is easier to view when its filtered.
08-27-2012 03:21 PM
This is not working either. I just got off with Tech Support and they are stumped as well. They said they will work on it offline and see what they can come up with. Is this really that difficult of a program???Reading four loadcells???? Getting them to 0......Perhaps this is too easy for NI.
08-27-2012 04:30 PM
No, its not that difficult at all. In fact, after tweeking the scaling of each load cell in your channel list the project file even works on my end. At least I'm able to zero the dummy signals.
Can you give a breakdown of your hardware?
Chassis
9237 module (d-sub or RJ50 connector)
Load cell mfg.
I'm betting on it being hardware/wiring related at this point. The d sub is connector is easy to miswire.
Who did the wiring for your connectors? Can you list a pinout? are you using remote sense or teds or anything?
In its simplest form, a non-teds load cell should have only 4 wires.
Really, it shouldnt be this difficult. I can generate a project with 4 load cells in less than 5 minutes. That's including tailoring my graphs to my personal liking....
08-28-2012 06:59 AM
The load cells were purchased with pick tails, I then put RJ50 connectors on them. I am using the NI9237 USB with these. The load cells are Toledo Transducers. It is a small custom mfg. in the area. There are - & + signal wires along with - & + excitation. I am pretty sure I have them configured correctly but I will verify.
08-28-2012 12:42 PM
Assuming your wiring is correct lets back up and take a different measurement approach:
Now you’re going to create Custom Scales for each load cell; if you’ve just created one channel then you’ll only need one custom scale, if you’ve created channels for all 4 then you’ll need 4 different custom scales since the cal values will all be different:
Now you need to calculate the y=mx+b “slope” for each channel using this formula: m=y/x (y=full scale of the xducer, x=sensitivity, m=slope, b=y-intercept or “offset”)
I’ve copied the mV/V value from the tables in your project file. Channel 1 (ai_0) in your list had a mV/V value of 1.1357. This is the electrical output of the sensor at full scale (1000 lbs). So using the above formula: m=(1000 lbs)/(1.1357 mV/V)= 880.5142. This is your SLOPE.
I find that with certain hardware I have to multiply this slope by a factor of 1000 because SE’s custom scales use Volts as a pre-scaled units and not mV. But with other hardware (like the ratiometric 9237 module) it wants a factor of 10, so your slope may be 8805.142. Honestly, if you find your measurement to be a decimal place off just move the decimal on the slope around till it’s right.
Note: I’ve found a quick and dirty approach to determining if your slope is scaled properly; I happen to know the grip strength between my thumb and fingers is roughly 20lbs. Squeeze the load cell and you should see 20lbs. If you see 2.0 or 200 lbs your decimal is off… This really only works with smaller load cells and is no way a reliable check.
Close the custom scales window leaving the y-intercept blank
From here you should be able to click RUN and view your measurements. In the DAQmx Acquire Step Setup screen I change the graph to a table and view the numerical display.
Try that for now and see if it works.
If it does you’ll want to move on to manually zero’ing the load cell offsets which you can do in the custom scale for each respective channel:
If none of this works you may have a hardware limitation. Page 14 of this manual explains the excitation limitations when working with multiple sensors requiring 10volts.
08-28-2012 01:22 PM - edited 08-28-2012 01:29 PM
You can also try with this proj file.
You'll need to import the scales into your system. Open MAX, click File> Import... or just save the file to the desktop and double click it.
This is a list of linear scales I've created based on your load cell values you had entered in that previous proj file.
~EDIT~
Youll need to delete the .txt from the file name. It appears this forum doesnt like the .nce extension.