T4 - Issues with RTD PT100 probe measuring temperature with LJTick-Resistance1000

6 posts / 0 new
Last post
jdinnh
jdinh's picture
T4 - Issues with RTD PT100 probe measuring temperature with LJTick-Resistance1000

Hi,

I'm having some consistency issue with the readings of with a PT100 RTD sensor on the T4 with LJTick-Resistance1000. The spec-sheet for the RTD sensor is found here https://cdn.automationdirect.com/static/specs/psrtdprobelwt.pdf

The sensor is 3-wires (2 red, 1 white). I've attached the white wire to the Vref slot, and the other 2 red wires both to the INA slot. Below is the configuration I'm using.

       ('AIN0_EF_INDEX', 40); // RTD Pt100

        ('AIN0_EF_CONFIG_A', 1); // Celsius

        ('AIN0_EF_CONFIG_B', 4); // Circuit 4, LBJ-Res

        ('AIN0_EF_CONFIG_C', 0); // AIN0

        ('AIN0_EF_CONFIG_D', 2.497); // Vref

        ('AIN0_EF_CONFIG_E', 1000); // Shunt resistor (ohm)

 

Two issues I'm seeing:

1. When monitoring temperature using Kipling, I'm seeing flickering between 2 different values returned by the T4 (see attached gif). I've made sure to completely to submerge the probe in the medium. I also scoped the output of Vref, and it seems stable. Measuring the resistance across the sensor directly with the MM and I see no jump in values.

2. Temperature does not update accordingly. (ie. temperature reading at 53C will stay there even when the medium cooled down to 30C over time).

Any help would be greatly appreciated. 

File Attachment: 
LabJack Support
LabJack Support's picture
The signal from the LJTR is 0

The signal from the LJTR is 0-2.5 volts, so you are going to get better results using a low voltage analog input (e.g. FIO4) rather than a high-voltage analog input.

https://labjack.com/support/datasheets/t-series/ain

You always have some noise with an ADC.  That is normal and in fact it appears that your value is only bouncing between 2 different codes which is very low noise (+/- 0.5 counts).

Start by looking at raw voltages instead of the AIN-EF feature.  Just look at voltage readings from AIN4 (for example) and see if the noise is as expected and then does the voltage change as expected.

 

LabJack Support
LabJack Support's picture
I did a little math for you

I did a little math for you using the normal 0.385 ohms/C for a PT100 with the LJTR-1k:

0C          100 ohms           2.2727 volts
30C        111.6 ohms        2.2490 volts
53C        120.4 ohms        2.2312 volts

https://labjack.com/support/datasheets/accessories/ljtick-resistance

https://docs.google.com/spreadsheets/d/1daW-Q-fFmkqI9-ekp-yhjkXitzB0CZM4...

 

jdinnh
jdinh's picture
I took your suggestion and

I took your suggestion and used FIO4 instead to get the analog voltage input, and used the voltage divider eq. to get the resistance on the sensor and manually calculate the temperature from there. Everything seems to be working good so far. Thank you!

jdinnh
jdinh's picture
I'm adding this here because

I'm adding this here because I thought this might be helpful for someone in the future.

I noticed a ~1.2C resolution per 0.001V change which is quite a lot. I added a 200 ohms resistor on top of the R102 on the LJTR-1k, bringing the shunt resistance down to 166.6 ohms from 1000. This gives a total of 266.6 ohms with the RTD, just slightly above the minimum 250 ohms required for 10mA. This results in a ~0.3C resolution per 0.001V, which is a lot more managable.

LabJack Support
LabJack Support's picture
I believe the main reason we

I believe the main reason we don't offer the LJTR with a smaller ground resistor is current draw.  If we had a 100 ohm version and someone is using a 100 ohm potentiometer, the current drawn from Vref will be 12.5 to 25 mA, but Vref is only rated for 10 mA.