Cheap, accurate thermometer

Thread Starter

tracecom

Joined Apr 16, 2010
3,944
I just completed assembling a thermometer built from an LM34DZ, a Harbor Freight DMM, and a 9 volt battery. Attached is a photo of the completed assembly.

The external battery provides power to the LM34DZ, and the output from the LM34DZ is fed back into the DMM. I compared the measurement of my assembly to a thermometer of known accuracy and found a significant discrepancy. Turns out, there is a trimpot inside the DMM, so a whim, I decided to try to calibrate the assembly. I placed the LM34DZ (in a plastic bag) in a styrofoam cup of ice and water and put the cup in the food freezer. In about 5 minutes, the reading on the DMM had stabilized and I then adjusted the trimpot in the DMM for a 320 millivolt reading, which represents a 32 degree F. temperature.

Thus with this adjustment, I matched the DMM to the LM34DZ and nullified any built in errors in both. Now, my assembly is within .15 degrees F. of my reference thermometer (which itself is only claimed to be +/- .1 degrees F.)

And best of all, I used one of several Harbor Freight 90899 DMM's that I have, most of which cost $1.99 or less. All told, my thermometer cost less than $10 including two new Duracell batteries. :D
 

Attachments

Last edited:

wayneh

Joined Sep 9, 2010
17,496
I've used the LM35 and the identical meter. I got mine free with a coupon. :D

Just curious: How far off was the multimeter? Any idea if the meter was off on any other voltage measurements? Also, did you load the output of the LM34 with a 220 ohm resistor? The LM35 output is unreliable without that.
 

someonesdad

Joined Jul 7, 2009
1,583
Wow, nice job! How well does the calibration keep at the upper temperature end, say around 80 °C or so? (I use a little second-hand toaster oven for such checks.)

If others build this, the catch will be the need to find a good temperature reference. Personally, I have a number of RTDs, so I'll use those because I trust them. If I didn't already have those, I'd get a Measurement Specialties 44007RC thermistor, as they are rated to around ±0.2 °C and you can get them for around $10-$12. To use them, you just need an accurate ohmmeter. BTW, Measurement Specialties sells the parts that YSI (who went bankrupt) used to sell (we used to have their precision thermistors in the lab stock where I used to work).

By the way, this little project is such a good idea, you might want to write up a little blurb on the construction etc. for newbies who haven't built such things before.
 

bertus

Joined Apr 5, 2008
22,270
Hello,

Nice setup.
For your information.
The LM34 is for Fahrenheit and the LM35 for Centigrade measurements.

Bertus
 

wayneh

Joined Sep 9, 2010
17,496
If others build this, the catch will be the need to find a good temperature reference.
Yes and no. I think for many folks, the LM35 (much more easily available) or LM34 device would be more accurate than anything they can easily get their hands on. Constructed properly, it's rated to ±1°C over its entire usable temperature range.

Now the meter readout is another matter. Unless you plan to dedicate the meter to this purpose (it's cheap enough), it would be better to calibrate the meter for its 0-200mV range using a calibrated meter, rather than build in any additional calibration of the sensor.

I'm not so sure how linear and/or adjustable, and or precise, this cheap meter is. It may drift, be affected by temperature, who knows. I trust the IC more.
 

Thread Starter

tracecom

Joined Apr 16, 2010
3,944
How far off was the multimeter?
I checked two DMM's with the same LM34DZ against my "accurate" thermometer. One differed by .7° F and the other by .4° F, so I used the one with the smaller discrepancy. Then, when I decided to calibrate the meter at 32° F, it was off by 3 mv. My conclusion now is that my "accurate" thermometer is actually off by .1° F, but I realize that conclusion may not be correct given the cheapness of the DMM.

Any idea if the meter was off on any other voltage measurements?
No, I didn't try to check it in any other way.

Also, did you load the output of the LM34 with a 220 ohm resistor? The LM35 output is unreliable without that.
No, I wasn't aware of this requirement. I have reread the LM34 datasheet (as well as the LM35 datasheet) and didn't find such a requirement. However, I'm not very good at reading datasheets. :) I have been using the DMM/LM34DZ thermometer continuously for the past 48 hours or so, and it seems very stable.
 

Thread Starter

tracecom

Joined Apr 16, 2010
3,944
How well does the calibration keep at the upper temperature end, say around 80 °C or so? (I use a little second-hand toaster oven for such checks.)
I have been monitoring the interior of a temperature controlled container (about 100° F.) for the past 48 hours, and comparing the DMM/LM34DZ thermometer with my "accurate" thermometer. Except for a slight difference in response time, they have stayed exactly .1° F apart for the entire time. I haven't tried it above 108° F.
If others build this, the catch will be the need to find a good temperature reference.
Except for the freezing point of water, I haven't found an absolute reference point for calibration. I tried boiling water, but the DMM only shows two decimal places above 2 v. I thought about body temperature, but that is actually quite varied from person to person and time to time as well. Google didn't provide an answer, either.
By the way, this little project is such a good idea, you might want to write up a little blurb on the construction etc. for newbies who haven't built such things before.
Thanks for the kind words. I actually did take some photos with the idea of perhaps an "instructable," but then decided it was too juvenile for this forum. About a hundred years ago, I made my living as a tech writer and thought I was pretty good at it.
 
Last edited:

someonesdad

Joined Jul 7, 2009
1,583
I actually did take some photos with the idea of perhaps an "instructable," but then decided it was too juvenile for this forum.
Remember that people come in all kinds of skill levels. This little temperature device is such a good idea and should be simple to build, even for people who have zero electronics experience. Thus, for example, a good beginning PDF on the topic could help some school kids or a single mom build something of use to them.

If you're not interested in doing the work, PM me and I'll do the work for you if you supply the info. I have lots of little technical/project tidbits stashed on my computer and I'm slowly getting them into digestible form and posting them on my Google code sites for open source stuff (here's one). I like to use this Google outlet because I suspect it will be around longer than I will -- and I'd like the stuff I put out there free for folks to remain available even after I croak (my wife says if I keep mouthing off to her, that day could be real soon now :p).
 

Thread Starter

tracecom

Joined Apr 16, 2010
3,944
This little temperature device is such a good idea and should be simple to build, even for people who have zero electronics experience. Thus, for example, a good beginning PDF on the topic could help some school kids or a single mom build something of use to them.
Actually, I am about half way finished with it. Maybe you would be interested in reviewing it prior to a general release?
 
Top