hi, is it possible to calibrate the sensor readings of an lm35 by connecting a potentiometer or trimmer resistor across it? i'm having trouble with results. Thanks.
Define what you mean by "calibrate".
Specify what trouble you are having. What is happening, and how is that different from what you want to happen?
That is the bare minimum necessary to help you.
sorry, i'll try again.
I have here an LM35 temperature sensor. This is a pretty accurate sensor but i'm using another thermometer and i want to match their values. Is it possible to connect a potentiometer or a trimmer resistor across it's data pin to alter it's results?
thanks..
Page 13 in the data sheet shows a Farenheit thermometer modification for it (http://www.ti.com/lit/ds/symlink/lm35.pdf). You might be able to adjust that to get whatever scaling you need, if you're clever enough.
That's a big might though.
First thing though, why do you need two different temp sensors to match like this?
If one is always different from the other by 0.5 degrees, simply add/subtract that from one of the readings.
What you want to do is feasible with the LM335, potentiometer connection is explained in the datasheet .
The LM335 gives a temperature in Kelvin, it is easy to convert Kelvin to Celsius by a simple translation:
0 ° C = 273.15 K
20 ° C = (20 + 273.15) K
The response of the LM335 is 10mV / K.
At 20 ° C :
- LM35 will give 200mV
- LM335 will give 2.931 V.
The LM335 is often cheaper than LM35 because it is used in greater quantity in the industry.
THank you all for sharing your ideas.
I think i already figured it out. 
what i did was connected the potentiometer directly to arduino and created a formula so that the voltage coming from my lm35 is directly proportional to the voltage from the pot, that way when i twist the pot from a set baseline it adds or subtract a certain value to the results of the sensor. It pretty much worked.
XD
The usual way of calibrating sensors attached to digital equipment is to make the adjustment in software, and store the adjustment data in EEPROM.
EssieN:
THank you all for sharing your ideas.
I think i already figured it out.what i did was connected the potentiometer directly to arduino and created a formula so that the voltage coming from my lm35 is directly proportional to the voltage from the pot, that way when i twist the pot from a set baseline it adds or subtract a certain value to the results of the sensor. It pretty much worked.
XDhow did you put the potentiometer? i am aso trying to calibrate our lm35 to get exactly the same with the digital temperature for our testing. thank you
EssieN:
hi, is it possible to calibrate the sensor readings of an lm35 by connecting a potentiometer or trimmer resistor across it? i'm having trouble with results. Thanks.
The LM35 is factory calibrated. If you want to fine-tune the results, simply do a two point calibration (ice and boiling water), then figure out M and B for the equation (Y=MX+B).
You cannot "calibrate" the sensor by varying the load on it's output. The output pin is an op-amp output which will do all it can to RESIST any changes you try to make to it.
Rule of thumb: Whenever possible, AVOID physical calibration (such as pots and resistors) and instead calibrate digitally via software.
EssieN:
THank you all for sharing your ideas.
I think i already figured it out.what i did was connected the potentiometer directly to arduino and created a formula so that the voltage coming from my lm35 is directly proportional to the voltage from the pot, that way when i twist the pot from a set baseline it adds or subtract a certain value to the results of the sensor. It pretty much worked.
XD
hi! what did you do? i am also having a hard time if it is possible to calibrate lm35.
Only LM335 can be calibrate by potentiometer, not LM35.
simply do a two point calibration (ice and boiling water),
That would assume you knew the local temperature at which water boils.
I am a hardware guy. I'm only learning coding. And I say, do the calibration in software.
If you use a potentiometer, now you have to worry about the change in resistance with time and temperature. Vibration can change the settings. It is another possible point of failure.
The ice thing isn't difficult. Freeze some distilled water, then smash it up and let it melt a little. It -must- be a mix of small bits of ice and water. Allow it to stabilize and have it in a reasonably insulated container like a small styrofoam cooler or large styrofoam cup. Put the probe right in the middle of the mix, don't let it get too close to the sides and -don't- let it touch the bottom.
In a pinch, if you have a mercury bulb glass thermometer, you can stick it and the probe under your tongue for 10 minutes or so and align those readings.
Otherwise, if you really want a super accurate boiling water point, you need a very clean glass or stainless steel pot, steam distilled or filtered water, and to accurately know the local barometric pressure. The water needs to be at a rolling boil, and needs to be left on the heat source. It will drop a few degrees within seconds of being removed from the heat. Same thing about the sensor, don't let it touch the sides or bottom.
It is probably good enough to know how far above sea level you are, and look up the boiling point of water on a chart.
Re the dipping it in the water thing.... presumably that's why distilled water, so there's no conduction? Or is the short across all three pins insignificant anyway?
Your LM35 better be made waterproof. You use distilled water because anything added to the water, salts, minerals, etc. will raise the boiling point.
Similarly, anything in water lowers the freezing point.
You have to ask yourself: How accurate is this other device? Are you sure the LM35 isn't already more accurate as it comes from the factory?
TI specifies that the LM35 is +-0.5 degree C at 25C and linear to +-0.25C.
Oh, and you cannot use a microwave to boil the water. It must be at a rolling boil when you calibrate, even a few seconds and the water cools.
@ Polymorph... Prezactly... DI water is the best, next distilled water and 'regular' water might well cause the "Calibration" to be worse than the Spec's for the sensor..
One other tip is to use a separate power supply for the LM35 Only, Clean and Well Regulated. I've used Many LM34/35's and any error is always In My Direct Experience due to either noise or a varying power supply, or a cheap China sourced part. A cheap fairly low current and Stable shunt regulated PSU for the device can be formed with a (Good) TIL431 and a current limiting resistor with a pot for adjustment.. This IS NOT a calibration pot... just the means to set the reference diode to a voltage within the Vcc limits of the sensor, The TIL431 only requires 100 uA or so for accurate regulation, The LM35 requires 90 uA quiescent current 250 uA is a good design target as it allows for minimum self heating of the reference diode while producing a stable output reference voltage... Don't forget to bypass the regulator well, both input and output. (min supply is 6V + 2V or about 7.5 to 8 Vdc input). This is really required to meet the base accuracy of the sensor... I attached the LM34 data sheet as it is identical to the '35 but for Fahrenheit measurement..
Doc
LM34XX.pdf (258 KB)