Differential thermistor inputs

Having problems ? let us know the details here
Post Reply
JohnWT
Newbie
Posts: 0
Joined: Tue Jul 22, 2014 7:08 pm

Differential thermistor inputs

Post by JohnWT »

I have a ADC20 with terminal board.

I am attempting to measure 3 temperatures using thermistors. I have wired them in series with a 10k precision resistor and fed the circuit from the 2.5v on the terminal board and grounded the circuit at (AG)

I connected each component to inputs (1-2), (3-4), (5-6) and (7-8) respectively.

As a test I used 10k precision resistors in place of the thermistors.

I setup 4 channels in differential mode, however the logged voltages were:

1 -593.710; 3 -605.856; 5 -630.060; 7 -668.694

Whilst these add up to ~2.5000 they are not all the same i.e. +625mV

What am I doing wrong please?

Martyn
Site Admin
Site Admin
Posts: 4491
Joined: Fri Jun 10, 2011 8:15 am
Location: St. Neots

Re: Differential thermistor inputs

Post by Martyn »

Can you please supply a sketch of your circuit and wiring.

Thanks
Martyn
Technical Support Manager

JohnWT
Newbie
Posts: 0
Joined: Tue Jul 22, 2014 7:08 pm

Re: Differential thermistor inputs

Post by JohnWT »

Hi,

Please find attached a sketch wiring diagram.

Is there an equivalent input circuit for the differential mode please?

Thanks for your interest.


John
Attachments
Pico Test Connections.jpg

PeterF
Advanced User
Advanced User
Posts: 435
Joined: Thu Jun 07, 2007 10:53 am
Location: Cambridgeshire

Re: Differential thermistor inputs

Post by PeterF »

Hi John,
What you have not realised is that on the ADC-20 there is an input impedance of 1Mohm between each individual input and AG. To calculate the expected voltages, you have to add the seven 1Meg impedances to the circuit diagram. (Channel 1 impedance is shorted out).
Apart from that, the whole circuit layout is not suitable as the calculation of temperature will be very complex since each thermistor will affect the others.
You need to set the ADC-20 so as to use only four channels in "Single Ended" mode. The -Ve inputs of these four channels need to be joined together with the "AG" pin. The four +Ve inputs are separate & each needs an individual "Pull-up" load resistor to the +2.5V supply on the ADC-20. Each thermistor then needs to be connected across the +Ve & -Ve inputs of each channel. You will now have four identical channels, each with a load resistor pulling up & a thermistor pulling down. Choose the value of the load resistor to be approximately the same value as the thermistor at the centre of your required temperature range. At this temperature, the voltage seen by each channel will be about +1.25V. When you do the calculation to generate a "Look-up" table of voltage against temperature, don't forget that there is a 1Meg impedance in parallel with the thermistor.
Below is an example of just such a setup, use suitable values for your own application:-
(In this example the thermistor to be used was:-
https://www.elfa.se/elfa3~se_sv/elfa/in ... &toc=19509
and the customer wanted to measure in the zone of 60 - 110 degrees Celsius.)
The centre of the range is 85 degrees Celsius so i will optimise for that value.
The thermistor above has a resistance of 8,376 ohms at 85C so Ra needs to match that for a centre scale reading of volts (+1.25V) at this temperature. Use an 8.2 Kohm resistor as the nearest standard value. Do NOT use a variable resistor as you will never know where you are for calibration. The ADC-20 has an input resistance of 1Mohm and this is in parallel with the thermistor thus the thermistor resistance value is reduced to (8.376 x 1,000)/(8.376 + 1,000) kohms = 8306.4 kohms (@85C). Assuming you are using the +2.5V sensor supply of the ADC-20, this is fed to the 8.2 kohm resistor and thermistor in series and the voltage at their junction wil be (2.5 x 8306.4) / (8306.4 + 8200) Volts = 1.2581V.
Doing the same calculations from 60C to 110C gives values:-

60C = 1.7997V
65C = 1.6987V
70C = 1.5931V
75C = 1.4828V
80C = 1.3707V
85C = 1.2581V
90C = 1.1473V
95C = 1.0419V
100C = 0.9415V
105C = 0.8464V
110C = 0.7579V

Assuming you have connected up the sensor, resistor & supply as described above, you just need to set the scaling for the channels in question. I will illustrate how to set-up channel 1 for temperature measurement from 60C to 110C.
Start PicoLog and set-up your sample rates as required.
When you get to setting up Channel 1 ("ADC-20 Channels"), highlight "Channel 1" and click "Edit".
Type "Temp Chan 1" as the Name, leave Conversion time as 60mS & select +/-2500mV as the Voltage range.
Leave "Differential input enable" un-ticked.
Now click on "Options". (You might be asked to save the present setup, click "OK").
In the "Parameter Options" dialogue box tick "Use parameter Formatting", put "degrees C" in the Units box. Field Width = 4, Decimal places = 1, Minimum value = 50, Maximum value = 120. Now click on "Scaling" to go to the next dialogue.
In the "Parameter scaling" dialogue box, select "Table Lookup" as the Scaling method and in the visible text box, type:-

1799.7 60
1698.7 65
1593.1 70
1482.8 75
1370.7 80
1258.1 85
1147.3 90
1041.9 95
941.5 100
846.4 105
757.9 110

(Please note there is a single space between the pairs of numbers.)
This tells the program to equate 1799.7mV to 60 degrees C on the graph and 757.9mV to 110 degrees C on the graph.

Go "OK" to go back through the open boxes and channel A should now be scaled from 50 to 120 degrees.

The sensor you have chosen is not the most accurate, the data sheet quotes +/-2.4degrees at 60C, +/-3.5degrees at 85C and +/-4.6degrees at 110C. You may want to choose a better one.

Look through the PicoLog manual for information on using PicoLog or click on the "Help" button at any stage in PicoLog. This is a link to the PicoLog manual:-
http://www.picotech.com/document/pdf/plw.en-3.pdf

Regards,
PeterF.

JohnWT
Newbie
Posts: 0
Joined: Tue Jul 22, 2014 7:08 pm

Re: Differential thermistor inputs

Post by JohnWT »

Many thanks.

Yes, it is now obvious. I intended using 3 separate constant current sources which would have involved a PSU and significant component count.

The recorded data is processed off line. I intend keeping the simple series circuit and solving the network to yield the 3 resistance values, and hence the temperatures.

Thanks again


John

PeterF
Advanced User
Advanced User
Posts: 435
Joined: Thu Jun 07, 2007 10:53 am
Location: Cambridgeshire

Re: Differential thermistor inputs

Post by PeterF »

Hi John,
Your series circuit can work but you only save two resistors and you lose a lot of resolution and accuracy.
PicoLog cannot display the temperatures directly either.
Please contact me again if you have any other questions.
Regards,
PeterF.

Post Reply