How do I zero 4 load cell channels with Signal Express?

[La-Z-Man]

OKors,

Thanks for the helps. This seems to have taken care of the issues. If I understand correctly, even when I load the calibrations in under the configuration scale I don't need the custom scaling feature.....

[La-Z-Man]

Also OKors,

Is there anything else I need to program or configure before I start running my testing?

Does the order of the DAQmx Acquire and the Zero Offset matter on the left hand side?

Do I need to click on the Offsets as Inputs or Export Offsets?

Another question is why does Signal Express keep timing out?

[OKors]

If I understand correctly, even when I load the calibrations in under the configuration scale I don't need the custom scaling feature.....

Correct. Use one or the other, not both. In fact (rereading your original post), dont' even use the Bridge Calibration, AKA Shunt Calibration Wizard, under the Device tab when using the Configure Scale. Its not necessary since your mV/V sensitivity is determining the slope and your Zero Offset is handling any signal offset. It sounded like in your phone discussion with NI the AE was having you use the Shunt Calibration wizard.

Is there anything else I need to program or configure before I start running my testing?

• If you're looking to save your data, you'll need to log or Save to ASCII your data files.
• Be sure your Signal Input range equals the min/max range of your load cells or you'll clip your signal if the range is too low; I.E. if using a 1,000lb cell youd enter +/- 1k
• Be sure to use whatever excitation the device was calibrated with or your slope will be off slightly. This should be on your datasheet or cal sheet for each load cell. 10v is pretty common but there is no real standard.
• Not sure what you have at your disposal but it's always a good idea to check the output of the load cells and make sure the scale is proper. Just a gut-check... A couple ways to do this would be to stack a "known" weight on the load cell itself. If you have a calibrated weight, or any KNOWN weight, stack it on the load cell and monitor the output on the screen. I happen to have a custom made breakout box that I put inline with between the load cell and the cable going back to the chassis. I can put a 100k ohm resistor in the breakout box and monitor the output. This is applying a "simulated" load. I have shunt cal info available on my calibration records which tell me what that simulated load is at 100k, you may not and you may not have a means of injecting a shunt into your system (something to think about in the long term).

Does the order of the DAQmx Acquire and the Zero Offset matter on the left hand side?

I believe the DAQmx Acquire step has to be first in the order. In practice this would be the first step anyway since channel setup is usually the first thing you do.... then you start adding steps as necessary.

Do I need to click on the Offsets as Inputs or Export Offsets?

Depends on your acquisition needs; usually you'll leave both of those options unchecked.
Hint: if you click Help> Context Help, a help pane will open up on the RH side. This will display help information on whatever options your cursor is hovering over; nice touch from NI and is great for getting a feel for what's going on within each screen. You'll know by reading the help context whether or not this option applies to your needs. Probably not is my guess.

I normally leave everything default settings (Setup Type= On Demand). Log a couple seconds of "dummy" data and stop. While monitoring the Input Output graph within the Zero step setup click Zero Now at the top of the screen.
Be sure when saving data or viewing signals on the graph you're saving/viewing the "Zero Offset Data". This is the signal to which the offset is applied.

Another question is why does Signal Express keep timing out?

Sounds like a timing thing. 

In the DAQmx Acquire settings be sure your Timing Settings are set to Continuous Samples and not N Samples or Samples on Demand.

Hope that helps.

PS, if you'd like. Id be happy to take a look at your .seproj file to make sure everything is in order. Wont be able to take measurements since our sensor types may be different, but if anything out of the norm appears that may create bogus data I may be able to spot it.

[La-Z-Man]

OKors,

When I open my project in Signal Express LE a License Information box says "This project contained unlicensed feature tabs. In order to successfully load the project, these unlicensed feature tabs were removed from the project." Then after about 5 mintues the program shuts down.

When I strated this project I had the SE 30 day trial and when that expired I registerd the LE version. Do I have to redo the project with the now registered LE?

How do I send you a copy of my program?

Ted

---

@OKors wrote:

If I understand correctly, even when I load the calibrations in under the configuration scale I don't need the custom scaling feature.....

Correct. Use one or the other, not both. In fact (rereading your original post), dont' even use the Bridge Calibration, AKA Shunt Calibration Wizard, under the Device tab when using the Configure Scale. Its not necessary since your mV/V sensitivity is determining the slope and your Zero Offset is handling any signal offset. It sounded like in your phone discussion with NI the AE was having you use the Shunt Calibration wizard.

Is there anything else I need to program or configure before I start running my testing?

• If you're looking to save your data, you'll need to log or Save to ASCII your data files.
• Be sure your Signal Input range equals the min/max range of your load cells or you'll clip your signal if the range is too low; I.E. if using a 1,000lb cell youd enter +/- 1k
• Be sure to use whatever excitation the device was calibrated with or your slope will be off slightly. This should be on your datasheet or cal sheet for each load cell. 10v is pretty common but there is no real standard.
• Not sure what you have at your disposal but it's always a good idea to check the output of the load cells and make sure the scale is proper. Just a gut-check... A couple ways to do this would be to stack a "known" weight on the load cell itself. If you have a calibrated weight, or any KNOWN weight, stack it on the load cell and monitor the output on the screen. I happen to have a custom made breakout box that I put inline with between the load cell and the cable going back to the chassis. I can put a 100k ohm resistor in the breakout box and monitor the output. This is applying a "simulated" load. I have shunt cal info available on my calibration records which tell me what that simulated load is at 100k, you may not and you may not have a means of injecting a shunt into your system (something to think about in the long term).

Does the order of the DAQmx Acquire and the Zero Offset matter on the left hand side?

I believe the DAQmx Acquire step has to be first in the order. In practice this would be the first step anyway since channel setup is usually the first thing you do.... then you start adding steps as necessary.

Do I need to click on the Offsets as Inputs or Export Offsets?

Depends on your acquisition needs; usually you'll leave both of those options unchecked.
Hint: if you click Help> Context Help, a help pane will open up on the RH side. This will display help information on whatever options your cursor is hovering over; nice touch from NI and is great for getting a feel for what's going on within each screen. You'll know by reading the help context whether or not this option applies to your needs. Probably not is my guess.

I normally leave everything default settings (Setup Type= On Demand). Log a couple seconds of "dummy" data and stop. While monitoring the Input Output graph within the Zero step setup click Zero Now at the top of the screen.
Be sure when saving data or viewing signals on the graph you're saving/viewing the "Zero Offset Data". This is the signal to which the offset is applied.

Another question is why does Signal Express keep timing out?

Sounds like a timing thing. 

In the DAQmx Acquire settings be sure your Timing Settings are set to Continuous Samples and not N Samples or Samples on Demand.

Hope that helps.

PS, if you'd like. Id be happy to take a look at your .seproj file to make sure everything is in order. Wont be able to take measurements since our sensor types may be different, but if anything out of the norm appears that may create bogus data I may be able to spot it.

---

[OKors]

Read through THIS thread; specifically message 2 and see if that helps.

Regarding attaching a file just save the SE project someplace and use Add Attachments Browser at the bottom of your next post.

[La-Z-Man]

 Hey OKors,

I have down loaded the trial version of SE with licensed on the way. I was able to run, save, and drop in Off set Nulling. However it is not doing anything. My data is still all out of wack. I have attached a copy of my program for your review and would appreciate your insight.

Thanks,

La-Z-Man

[OKors]

I'll take a look at your proj file. In the meanwhile can you elaborate on "its not doing anything"?

what's it not doing?

[La-Z-Man]

The zero offset is not doing anything. (Sorry for not elaborating) All four load cells are reading the same 150, 230, 280???, and 340???. They are not reading zero with no load applied.

[OKors]

Unfortunately I cannont connect 4 load cells to my system to check your signals so I cannot see the 150, 230, 280, and 340. But a few things come to mind.

Where are you getting your cal values for the load cells?
In your table you're showing a max of 1~ish mV at full scale. Usually full scale us upwards of 20 or so mV (2.0 mV/V). Double check your cal tables and make sure the units are correct.

Also, is there no mV/V/FS "sensitivity" for your load cells? In my experience, this value is more common (and preferred) than a table as you've defined in your test. With a table range you've hard set a value of 0 as a minumum. Load cells are designed to go in tension and compression so with a hard 0 defined you will never see anything in the negative. Not a big deal if that's what you're expecting and do not desire anything in compression, but a mV/V sensitivity will allow your signal to go positive and negative.
And I may be totally wrong (because I never use a table scale), but that may be the reason you cannot zero.
As a test I changed your "Force_0" channel to use a two-point linear scale rather than a table. I entered 2.08mV/V as an electrical Second value and 1k as a physical second value. Set yourSignal Input ranges to min= -1k max= 1k. 
Run some data and not only is the signal not flatlined at zero but I can also zero the signal in the zero step.

Do you have your calibration sheets in electronic form? I'm curious to see what information you have available.

[OKors]

OK, hang on a minute, I think my coffee just now kicked in...

Get rid of your tables for all those channels and use the Two-Point Linear scale option.Take your last value entered in your table (so for your Force_2 channel the last value is 1.09007 mv/V) and enter that as your Electrical Second Value and enter 1000 as your Physical Second value.
Do this for all your channels, again your value of interest is your full scale mV value, whatever that is; ignore the values leading up to full scale.

Change your Signal input range to +/- 1k, not 0-1k.
Take some data and see if your signals zero.

You've got some funky stuff going on in your graph also...

I've attached a modified proj file, see if that works and note the changes I made.

Also, please be sure your the values on your load cell datasheet were created using a 10v excitation as this will affect the scale if not.