Go Down

Topic: Why is my temp off? (Read 394 times) previous topic - next topic

iwrk4dedpres

Ok ... New to all this electronic stuff.  Bought Uno couple years ago ... collected dust while I looked at it ... Now got a sensor kit and trying out stuff.


Sitting here in my living room .. Wall thermometer reads 69*F and this thermometer reads 59.74... I'm within 8ft and 2 ft lower than the one on the wall.

Can anybody shed information on how to:


Procedures they do to ensure that readings are accurate.
Procedures to calculate an offset?



Just information in general.


Sensor I'm using is a DS18B20 from this kit:  Click Here!

quamikazee

I've found the same thing with sensors.
Have you checked to see if it can be calibrated?

iwrk4dedpres

Well the kit was pretty cheap and finding information about the actual discrete parts used is not something I've got great skill @. 

Was kind of wondering if I was being an idiot or if I'm just another chump in the crowd lol.  Being new to something makes you question your intelligence alot.

rogertee

first off what is the wall thermometer an old bulb type (red Liquid) or the furnace thermostat. You are 8 feet from it you are measuring air temp while it is measuring wall temperature. To get a totally accurate reading you need to physically attach the two together i.e. tape them together or something. To get a 99% accuracy you would need a laboratory grade thermometer. What is normally done is to put both sensors in a glass of water either with ice, room temp or boiling water then compare them. And different thermistors read different values at same temperature. Your temp sensors needs to be calibrated to read right. If this topic interests or concerns you study Arduino thermistor and various libraries and the different formulas to read them

iwrk4dedpres

first off what is the wall thermometer an old bulb type (red Liquid) or the furnace thermostat. You are 8 feet from it you are measuring air temp while it is measuring wall temperature. To get a totally accurate reading you need to physically attach the two together i.e. tape them together or something. To get a 99% accuracy you would need a laboratory grade thermometer. What is normally done is to put both sensors in a glass of water either with ice, room temp or boiling water then compare them. And different thermistors read different values at same temperature. Your temp sensors needs to be calibrated to read right. If this topic interests or concerns you study Arduino thermistor and various libraries and the different formulas to read them
Not sure but I think throwing this circuit into a glass of water would be detrimental to my uno.  I thought of that as well ( putting it near the wall thermometer btw it's a digital one ) just to remove that variable.


Thanks for responding.

Nick_Pyner

I would be more inclined to trust the DS18B20 than the thermometer on the wall, and no, they don't need to be calibrated. They are not thermistors.

quamikazee

Well the kit was pretty cheap and finding information about the actual discrete parts used is not something I've got great skill @.  

Was kind of wondering if I was being an idiot or if I'm just another chump in the crowd lol.  Being new to something makes you question your intelligence alot.
Don't worry...everyone starts somewhere.   I'm only slightly ahead of you.   Still a vast amount to learn.
But I have found that some of the Chinese sensors can vary by so much that they're almost useless.

Some work pretty good...some are garbage.   I usually order 3 or 4 or any sensor I get and see which ones I'll keep and which ones are paper weights.


quamikazee

I would be more inclined to trust the DS18B20 than the thermometer on the wall, and no, they don't need to be calibrated. They are not thermistors.
Here's a thread on calibrating the sensor
http://www.avrfreaks.net/forum/calibrating-ds18b20-temperature-sensors-better-accuracy

iwrk4dedpres

Well I tend to think that if 2 things read the same thing then they should read the same value.


I get that they won't be the same .... but in the same GENERAL environment readings of 68F and 59F seem to be wildly out of place ( i'm in my living room ).


Trying to read the datasheet for the D18B ic now.

iwrk4dedpres

Here's a thread on calibrating the sensor
http://www.avrfreaks.net/forum/calibrating-ds18b20-temperature-sensors-better-accuracy
Well holy bat crap batman .... that's way more information than this general layperson has the ability to understand.  LOL.   


But I appreciate the link.  BTW .... I had to break out the crayola's with my UNO.  KISS applies to me at least the "stupid" part

FullOfBadIdeas

my quick $.02 - the quality of the sensor matters - in terms of library support and usability.  I spenmatht way to many hour cursing temp sensors, and getting BS readings, until I switched the sensor type I used. 

I'm now using an AM2302.  cost a few bucks more, but worth it in that it works, and gets working quickly without a ton of funny calculations. 


iwrk4dedpres

my quick $.02 - the quality of the sensor matters - in terms of library support and usability.  I spenmatht way to many hour cursing temp sensors, and getting BS readings, until I switched the sensor type I used. 

I'm now using an AM2302.  cost a few bucks more, but worth it in that it works, and gets working quickly without a ton of funny calculations. 


yeah I'm sure after sometime, if I keep this up, I'll come to the same conclusions.  For now just trying gain understanding to what I'm doing.  Will look up your part reference.

Far as the library support ... yeah I keep looking at the libraries that are used like the dallasxxxxx something or other and wish I understood what was in it.  But that will take time as well.  Sticking in searial prints where I'm trying to understand something.

Nick_Pyner

Well I tend to think that if 2 things read the same thing then they should read the same value.
This is true, but not enough of a reason to doubt the DS18B20.  You make no comment about rigorous comparison, and are just talking generalities. The wall thermometer may be reading a convection stream rather than the typical room air.  One thing you can be sure of is that any perceived NEED to calibrate DS18B20 is pure horseshit.  These things are one of the most popular sensors on this forum, and for good reason. If any calibration was needed, everybody would have known about it for years - and would be using something else. Indeed about the worst thing that can happen with the DS18B20 is that people have complained that they are having a hard time with what they think is a DS18B20, and it turns out to be something else. In that event, no amount of calibration will fix the problem.

I have a couple of dozen DS18B20s and I have to admit that I do test every one of them in water against a lab-grade 0-50 mercury thermometer. I also have to admit that I don't actually have a good reason for doing this, as I have never found one that doesn't do what it is supposed to do.

GolamMostafa

The DS18B20 sensor does not require any 2-point calibration like LM35 to find gain and offset. The DS18B20's reading is directly mapped to normal decimal value (natural binary). The Following Table has been extracted from the data sheets of the sensor.


TomGeorge

Hi,
You do have the 4K7 resistor fitted?


Can you post your code please?

Tom... :)
Everything runs on smoke, let the smoke out, it stops running....

Go Up