a doubt on stabilty of temperature measurement

[wwert]

I have a SCXI-1102,and two SCXI-1112 in a SCXI-1000 chassis with a SCXI-1303 terminal block. I connected one thermocouple at Channel 0 in SCXI-1300, and also connected other TCS to SXCI-1112. And I tried to measure ambient temperature by T-type thermocouples. In VI, I set CJC type to be constant value, and it automatically set default temperature at 25 degree cellsius. Then I test it, but temperatures by thermocouples are differert from the value of thermometer. So I changed CJC constant values accordingly, and I testest again, It still did
not show the right value.

BTW: When the data acquire system run for a long time, such as half an hour, does it affect the data
acquired?

[Conseils]

You don't say how much different....

You really need a temperature simulator to 'Calibrate' the system.

If you use a type K thermocouple you can almost assume a linear 40mv per °C for calibration. Thus it's easier to use a voltage source to get an idea of operation.

You say that you set the cold junction manually, this could be your problem because the room air temperature could change significantly. e.g. an air conditioning system such as a Hitachi unit for example will blow air at 13°C into the room. So in conclusion your room is not at an even temperature especially if you have a sunny window at the opposite side. As a result things in your room, the cold junction, terminals and air are all at very different temperatures.

P.S. In my experience, norma
lly you don't have problems with calibration of the NI kit. It's use of it that's more often the problem.

Regards

[wwert]

The difference is about a few degree.
I am new for Labview, can you tell me how to caliberate the system. And I am using T-type
thermocouples.

[Conseils]

To CALIBRATE you must use a reference which is traceable to a national standard.

I suspect what you want to do is check.

Warning the following involves water which could be near electricity - take EXTREME CARE!!!!! There is also hot water and electricity - very dangerous if care is not exercised.

1) Get some ice
2) Smash it up
3) Mix in a flask 50% water 50% ice
4) Leave for 10 minutes
This is 0°C (Zero)


1) Boil a pan of water
2) Keep it boiling for 5 minutes
This is 100°C (Don't put your sensor in the bottom of the pan, near the top third)

The above are quite precise, if your not at some very high elevation. One can compensate for elavation if required.

[wwert]

I have do something like this. However, I did not get a good result. For example, I keep the CJC constant to be the room temperature. then I put the thermocouple into ice watter system. It did not show zero. Should I change the constant? Or Change it to be 'bulit in', or anything else?

[wwert]

I forgot one thing, When the thermocpuple in the ice water system, the reading decreaced after a few second. Do you know the reason.

[Randall_T]

Both of the modules you mentioned have built in cjc sensors. You mentioned the 1303 terminal block then later called it a 1300. The 1303 should have a more accurate cjc reference than the 1300. I have not used the 1112 but since it is designed specifically for Thermocouple measurements they should have a decent cjc sensor network. I will mention that the miniconnector plugs should be the right type for the T type you are using but I imagine you have done this.

In the scan mode your 1102 module will continue to scan the cjc so changes in ambient temperature at the chassis should be compensated for. Since the 1112 are newer module they should do the same. Some of the older isolated modules only picked up the cjc reference at on the initialization of the sc
an. I have used the 1102 modules with the 1303 terminal blocks for several years. The accuracy of this combination usually keeps me within a couple deg F. when I check the system.

If you are using the 1303 terminal block make sure you have the right resistor network for your application. Check your documentation for if you have the 10 ohm bias resistor in and your tc is grounded you will see a significant error.

There are ways of using an external CJC to improve your accuracy if needed. These methods will increase the cost of your system, as they will require additional hardware.

[wwert]

I am new for Labview. So I didn't use the miniconnector. You mean the module can update the
CJC temperature, it cause the decrease in reading from thermocoupes, right?
Also I have not use resistor and my tcs were not grounded. So I can use the resistor, right? and how?

[Conseils]

Why are you setting the Cold junction temperature, it must be the reading from the compensated cold junction to correctly calulate the actual temperature.

[wwert]

I am new for NI equipment. Can you explain the procedure about how to measure the temperature and thing should be pay attention to?
For example:I know the CJC temperature always changes with time.But How can I set it, there are three choices:'constant','build in',and 'channel'.
You said I can't set CJC temperature, so I have to set it either 'build in'or'channel'. right,
Now suppose I set 'build in', if the temperature measure by the thermocouple in ice-water system is not zero, what can I do?

Thanks.