RTD PT100

[Jim90]

I recently bought a thin film pt100 temperature sensor and i am trying to measure the temperature value.

I built a voltage divider (100ohm + RTD PT100) and in the middle of the two i connect the received value which is going to the microcontroller (measure the temperature value).

The code which i am using is presented below:

[temtronic]

'very high values' is not a good description to go on.
Typical RTD PT100 devices are 100r@0*C, 138r@100*C, if we assume a Vref of the ADC section to be 5.000 ( PIC power supply), the range is 2.50 to 2.89 volts.

You should tell us the temperature of the sensor and the raw ADC bits.

Also be aware that actual wiring of the sensor, board layout, etc. have a very important part in accurate measurements. Also RTDs are NOT linear devices so you need to do some 'math' to get the best results.

http://www.picotech.com/applications/pt100.html is one quick random 'google' for more information that should be useful.
You don't tell us what PIC, which ADC setup (8 bit, 10 bit, ??), conversion rate, etc. All this is important to know....

hth
jay

[Mike Walne]

You don't tell us which way up your divider is wired.
I'm assuming RTD to GND, 100R to +5V.
Have you read the data sheet for the RTD?
If it's thin film it could be getting hot.
Give us more to go on.

Mike

[Jim90]

I use dspic33fj128gp802 with voltage Supply equal to 3.3V and ADC: 10 bits

I put a 200ohm resistor in series with rtp and in the middle is connected it to the adc input.(At one side of the resistor i connect a supply and at one side of the rtp i connect ground).

I know that i need to do some maths so to receive the correct value but i have little confusion what exactly i need to do.

I am waiting for your respond.

[Mike Walne]

Is this what your circuit looks like?

[Jim90]

My circuit is :

200R RTD
+3V3 _______WWWWW__________WWWWW______ GND
|
|
|
dspic33

What are the necessary calculations for the above circuit?

[Jim90]

200R RTD
+3V3 _______WWWWW__________WWWWW______ GND
|
|
|
dspic33



Now is clear . Sorry for the confusion.

[temtronic]

basic OHMs law voltage divider

Vout=(rtd/(rtd+200)) * Vin

rtd=+-100r@0*c
vin=3.3

Of course you must measure vin (it won't be 3.3), rtd value varies according to temperature obviously.

Simple test is to freeze the sensor then let it warm up, measure ADC input and record ADC bits on a chart, say every 10 seconds.

hth
jay

[Jim90]

I changed the circuit and I did it as Mike Walne presented.

My confusion is that the Mike Walne and the temtronic explained different ways to processing the input signal.


Can you make it more clear please?

[Mike Walne]

[Jim90]

I came back to continue the project. I use the following circuit:

http://obrazki.elektroda.pl/1446047100_1392586975_thumb.jpg

The Vout = (RTD*VCC)/(RTD+R)
where VCC=3.3v and R=10K


According the following link :


http://www.michshur.co.il/image/users/190375/ftp/my_files/Files/TD_TV_PT1A.pdf?id=8609087


for temperature : -10 c -->R=96.09 and with the use of the Vout equation V=2,98
for temperature: 110c -->R=-->V=3,083

for a 12 bits adc (1023) 120/1023 = 0,111.


what is the next step???

[Mike Walne]

[asmboy]

i suggest using external analog processing of the sensor data, and
an external precision 2.5 volt ADC ref. like the lm4040, since if your last post were correct, it shows you are trying to map a 120 degree span into 127 bits. with a reasonable assumed uncertainty of any single ADC count,
of +/-1 bit, you won't resolve any less than 4 degrees on average.

unless your pic power supply is providing a highly accurate and compliant 3.3v, your accuracy is also never going to make you happy either.


your problem is in circuit design, or lack of.
Direct reading an RTD, unless the temp range is extremely wide, is tough with only 10 to 12 bits of conversion,..l..

analog processing is the cheaper way to get better resolution, rather than using a lot more expensive bits of external ADC to achieve the same ends.

[Jim90]

You are right Mike Walne,

the values are :

for temperature : -10 c -->R=96.09 and with the use of the Vout equation V=0.031
for temperature: 110c --->V=0.046.


asmboy i don't want to use an external adc.

these values look like ok?


(to be honest i expect to receive something like 0.3 instead of 0.03 )

[asmboy]

you might google this phase and look at the images:

rtd schematic

with a 10bit adc - you are simply not going to resolve the temperature range you hope for, or with a noise level in the readings you get, to want to brag about it . at least not with the simple circuit you hope to use.

to work well with so few bits you need a constant current source for the RTD and analog gain. Its just math.........
over and out

BTW: (1023) 120/1023 = 0,111.

actually there is general agreement on this forum anyway,
that 1024 is 10 bits worth not 12

i think 1024 is 12 bits only on ATMEL MCUS