Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

Themperature measurement with nV-meter 2182

2182 Series Support
Post Reply
hajek1
Posts: 4
Joined: December 13th, 2010, 6:09 am
Country: Czech Republic

Themperature measurement with nV-meter 2182

Post by hajek1 » December 13th, 2010, 6:24 am

Hello, I am measuring temperature in a furnace with K thermocouple and Keithley 2182 nanovoltmeter (ch1). For this measurement I use internal CJC (cold junction ccompensation) of the instrument (i.e. I am _not_ using simulated junction). However, whenever temperature gets near the value 499.94 °C (centigrades) I see abrupt jump in the data. Heating or cooling - the direction of temperature development does not matter. I suspect it has something to do with K thermocouple calibration. Or a change of resolution ... ? I am sure the jump in the data is not physical effect in the furnace and it is not numerical problem in my computer. The jump is seen regularly from temperature 499.98°C to 499.947°C +- 0.004°C. I do not see any similar discontinuity from room temperature up to about 700°C. Can you please help me to clarify this issue? Thank you very much.

Jennifer C
Keithley Applications
Keithley Applications
Posts: 49
Joined: September 17th, 2010, 7:55 am
Country: United States

Re: Themperature measurement with nV-meter 2182

Post by Jennifer C » December 13th, 2010, 6:57 am

Hello hajek1,

Can you please be more specific about the "jump" that you see in the data? What is the temperature difference across the readings? How does this compare to how fast you are changing the temperature of the furnace? How often are you making temperature measurements? Do you have data that you attach to this post?

Jennifer

hajek1
Posts: 4
Joined: December 13th, 2010, 6:09 am
Country: Czech Republic

Re: Themperature measurement with nV-meter 2182

Post by hajek1 » December 13th, 2010, 2:34 pm

Dear Jennifer, thank you for prompt reply. I have made a graph of my data, please have a look: [img]http://material.karlov.mff.cuni.cz:people/hajek/keithley/jump.png[/img]
http://material.karlov.mff.cuni.cz/peop ... y/jump.png
here is the setup of the instrument:

*rst
:init:cont off
:sens:func 'TEMP'
:sens:temp:tran tc
:sens:temp:rjun:rsel int
:sens:temp:tc k
:sens:temp:nplc 1
:init:cont on
And reading is done by :data:fresh? query.
I do have the data, but the file is rather long (about 740000 data points). So I will separate interesting parts and attach it somewhat later (I do not think that will bring more qualitative understanding than the above graph though). As for different speeds of measurement - I did not test it, so I do not know. Now I am doing about 16 measurements per second. But why should speed matter? The jump appears only at that one specific temperature. Nowhere alse. And I have the same speed for whole temperature range.
Last edited by hajek1 on December 13th, 2010, 11:55 pm, edited 1 time in total.

hajek1
Posts: 4
Joined: December 13th, 2010, 6:09 am
Country: Czech Republic

Re: Themperature measurement with nV-meter 2182

Post by hajek1 » December 13th, 2010, 11:53 pm

Ok, here are the data. First column is time [s] , second column is temperature [oC]. If you plot the time range 7000:15000 seconds and 499.9:500.1 centigrades on a big enough screen (or zoom as necessary), you will notice the jumps I am talking about. Particularly, around 9000-9300 s. or 11000-11500 s. But as far as I have noticed, the jumps are present whenever temperature crosses 499.94 oC. Feel free to check yourself. For the sake of complete clarity, I add the snippet of C code for reading the numbers from device (2182 nV-meter):

Code: Select all

 
        lptgpib_write(K2182N, "data:fresh?\x0a");
        lptgpib_read(K2182N, buf, sizeof(buf));
        clock_gettime(CLOCK_REALTIME, &now);

        y =  atof(buf);
        x = now.tv_sec-start.tv_sec + 1e-9*(now.tv_nsec-start.tv_nsec);
        printf("%Lf %Lf \n", now.tv_sec, now.tv_nsec, x,y);
But as far as I understand, the effect must come from the Keithley 2182 device (firmware included) itself. Otherwise one would see it on other occasions as well.

Oh, the 8MB attachment is too large for this site. So I uploaded the data here:
http://material.karlov.mff.cuni.cz/peop ... ithley.dat


I am looking forward to hear peoples openion on this phenomena. Thank you in advance!

Jennifer C
Keithley Applications
Keithley Applications
Posts: 49
Joined: September 17th, 2010, 7:55 am
Country: United States

Re: Themperature measurement with nV-meter 2182

Post by Jennifer C » December 14th, 2010, 6:59 am

Hi hajek1,

Thank you for the additional information. All I needed was the graph.

It is interesting that the readings jump back and forth for 40 seconds. However, you should know that the jump observed is within specification for the instrument. For K type thermocouples, the Model 2182A Nanovoltmeter has a specified accuracy of +/- 0.2 deg C. The 0.03 degC jump observed in your data is well within that. But, let's see if we can find out any reason why this jump is occurring

I checked the K type thermocouple reference chart and it does not appear that the instrument is changing ranges during the measurement, as the thermoelectric voltage should only be in the range of 20.6 mV.

The code snippet that you sent indicates that you are programming the instrument correctly.

Can you try running the test again, with the same hardware configuration, but using the voltage measurement function of the Model 2182A instead? To measure voltage, use the command ":SENS:FUNC 'VOLT'". I'd like to see if we see a similar jump. You will probably want to use autorange or use the 100mV range of the instrument to make the measurement.

--Jennifer

hajek1
Posts: 4
Joined: December 13th, 2010, 6:09 am
Country: Czech Republic

Re: Themperature measurement with nV-meter 2182

Post by hajek1 » December 14th, 2010, 7:19 am

Dear Jennifer, thank you for your effort.

I am fully aware that this effect appears well below the limits given by the specifications, that is - I do not blame Keithley :) It just seems to me worth investigating as I am curious why this effect happens.

I will try running the test again with voltage measurement instead of temperature measurement as suggested. But this will happen only after a few days, since I am doing some other important tests right now.

Post Reply

Return to “Nanovoltmeters”

Who is online

Users browsing this forum: No registered users and 6 guests