Calibrating 10cm range LVDT connected to Omega process controller using gauge blocks

Thread Starter

capivara

Joined Nov 8, 2018
10
I have an Omega LVDT (model is LD620-50) that I want to calibrate to use in lab tests. I'm using it alongside a process controller (model is CNiS16D22-C4EI) and I have gauge blocks that I want to use to calibrate it. I have 4 gauge blocks (1.3mm, 1.9mm, 7.5mm, and 18mm) and although the range of this LVDT is +/- 5cm, I only need a range of 2cm for my application.

I tried calibrating the LVDT by using the Input/Reading part of the controller's menu, and Load enabled so that the calibration is done online. I understand that I have to place the LVDT at the specified lengths, thus inducing a voltage that the controller reads (Input), and then I have to make the Reading the appropriate measurement in millimeters.

I tried using 5 linearization points (my four gauge block lengths and 0) and setting it up with the controller as described above. However, once I start measuring in run mode, I get very wild numbers that are off by as much as 9% error even when I simply place the gauge blocks used to calibrate it under the probe.

Am I doing this the right way? It seems very strange that right after setting up the input/reading points, I get numbers that are so wrong.
 

Reloadron

Joined Jan 15, 2015
5,646
Do you have a link to the manual you are using? The units I have used are slightly different in the InScOf menu. I am familiar with where you get a prompt and enter the number of points between 0 and 10, in your case 5 and as you go through each point similar to In 1 and there is a number based on the actual input so for example with zero travel the meter displays a number. Then you step to Read 1 and you enter what you want the meter to read, in this case 0 and when you enter that number hit enter and the meter should flash StRd (Stored) . The meter advances to In 2 and you repeat the process with the first gauge block inserted.

The only manual I am finding is this one which does not get into the multipoint setup, only two points. It's similar to many of the older meters which only offered a two point InScOf (Input Scale and Offset).

My guess is that the meter works but something is not taking in your points during the setup.

Ron
 

Thread Starter

capivara

Joined Nov 8, 2018
10
Do you have a link to the manual you are using? The units I have used are slightly different in the InScOf menu. I am familiar with where you get a prompt and enter the number of points between 0 and 10, in your case 5 and as you go through each point similar to In 1 and there is a number based on the actual input so for example with zero travel the meter displays a number. Then you step to Read 1 and you enter what you want the meter to read, in this case 0 and when you enter that number hit enter and the meter should flash StRd (Stored) . The meter advances to In 2 and you repeat the process with the first gauge block inserted.

The only manual I am finding is this one which does not get into the multipoint setup, only two points. It's similar to many of the older meters which only offered a two point InScOf (Input Scale and Offset).

My guess is that the meter works but something is not taking in your points during the setup.

Ron
Thank you for your reply. I was using that same manual. My guess is that the online calibration had such poor results because of a bad LVDT. So I am now trying with just 2 linearization points and offline calibration -- so setting 0V to 0mm and 5V to 50mm -- and it is looking better.

There is some error, and I am now trying to adjust the slope or maybe add a vertical offset to get the correct measurements (I'm using the gauge blocks to check). So far the best I've gotten to is setting 0V to 0mm and 5.028V to 50mm, where I get errors of +/- 0.05mm. Which is the minimum requirement for my application so I think it might work out with a little more tuning.

However, I have noticed that the linearization only looks good for a segment of the LVDT's full range, about 40% of it. Outside of that range the errors shoot up to +/- 0.1 mm or worse. Honestly I'm starting to think I have a bad LVDT because the linearity should be better than 0.2% for the full range according to the datasheet. But thankfully I only need 2cm for my application so I'm still hoping I can make it work.
 

Reloadron

Joined Jan 15, 2015
5,646
Thank you for your reply. I was using that same manual. My guess is that the online calibration had such poor results because of a bad LVDT. So I am now trying with just 2 linearization points and offline calibration -- so setting 0V to 0mm and 5V to 50mm -- and it is looking better.
OK, that is good and something I forgot to suggest. If you have a good DMM (bench meter or any meter) I would measure the input going to your process meter. Insert your gauge blocks and see if the input repeats up and down scale for the travel. That should tell you if the LVDT is repeatable.

Something else you can do is call Newport or Omega and get one of their applications engineers on the phone. Ask them to walk you through it step by step. Also request an email from them with a link to a manual which covers the multi-point setup. The idea being the meter, in multi-point sets up a Y=M*X +B linearization curve where it should be much more accurate.

Ron
 

Thread Starter

capivara

Joined Nov 8, 2018
10
OK, that is good and something I forgot to suggest. If you have a good DMM (bench meter or any meter) I would measure the input going to your process meter. Insert your gauge blocks and see if the input repeats up and down scale for the travel. That should tell you if the LVDT is repeatable.

Something else you can do is call Newport or Omega and get one of their applications engineers on the phone. Ask them to walk you through it step by step. Also request an email from them with a link to a manual which covers the multi-point setup. The idea being the meter, in multi-point sets up a Y=M*X +B linearization curve where it should be much more accurate.

Ron
Is it correct to assume that if the LVDT datasheet states the linearity is <0.2% FSO and the range is 20mm, the uncertainty on any measurement is taken to be 20mm*0.2% = 0.04mm (so +/- 0.04mm)? Additionally, if the datasheet states that the linearity error is +/-0.15% for FSO for a 20mm range instrument then the uncertainty on any measurement would be 20mm*0.3% = 0.06mm (+/-0.06mm)?

I'm asking because these uncertainty values seem rather high for "high accuracy" Omega LVDTs, while I have mechanical Mitutoyo dial indicators that are basically as good as +/- 0.01mm. Shouldn't the LVDTs be at least as good as the mechanical indicators?
 

Reloadron

Joined Jan 15, 2015
5,646
You may want to call an Omega App Engineer as I read it the same as you but yes, that seems a little high for the uncertainty.

Ron
 
Top