Standard circuits for LM35DZ

Thread Starter

Chris Gray

Joined May 10, 2015
6
I've got the most basic circuit to work with this chip, 4-20V supply and reading the voltage with uC or meter. Works fine. But none of the other ones I've tried will work. The useful one to me is the "-55 to 150 degrees C" but it seems to do nothing. I haven't tried using a proper negative supply rather than the 2 1N914 diodes as shown in the datasheet. The output I got from this is really unstable and a negative going voltage with increasing temperature. I used 2 1N4148 diodes, which I believe are about the same as the 1N914.

Any clues? The rest of the example circuits don't have a large enough range.
 

Attachments

OBW0549

Joined Mar 2, 2015
3,566
Hmmm... I've been using LM34DZ and LM35DZ temperature sensors for years, and never had a problem with them; they're easy to use, accurate, stable and predictable. The only thing I can think of is you've got something connected wrong, somewhere.
 

GopherT

Joined Nov 23, 2012
8,009
I've got the most basic circuit to work with this chip, 4-20V supply and reading the voltage with uC or meter. Works fine. But none of the other ones I've tried will work. The useful one to me is the "-55 to 150 degrees C" but it seems to do nothing. I haven't tried using a proper negative supply rather than the 2 1N914 diodes as shown in the datasheet. The output I got from this is really unstable and a negative going voltage with increasing temperature. I used 2 1N4148 diodes, which I believe are about the same as the 1N914.

Any clues? The rest of the example circuits don't have a large enough range.
I agree with @OBW0549 . Double-check the pinout on the datasheet and compare to your circuit.


One warning if you are using the two-diode single supply circuit. You must measure voltage (temp) from the two points on the schematic you posted above, you cannot measure from ground to LM35 output. If you measure from the ground as your negative contact, then you will not get an accurate voltage because the two diodes change voltage with temperature as well (more than your sensor).
 

MikeML

Joined Oct 2, 2009
5,444
Right off the data sheet:

35.gif

Look how Vout is taken differentially.

If you want it ground referenced, you have to use a negative supply or follow the LM35 by an opamp wired for differential input (with extremely well-matched resistors), or a true instrumentation amplifier. It is likely easier to use a charge-pump to derive the negative voltage.
 

AnalogKid

Joined Aug 1, 2013
10,987
The problem might be at the measurement end. If this is going into an A/D, either a standalone chip or uC input, the - input might be grounded there unintentionally.

ak
 

Thread Starter

Chris Gray

Joined May 10, 2015
6
I'm next going to try the one at the start of the datasheet (certain datasheets), using a negative voltage and with bias to move the range into positive levels for very low temperatures. The pinout I use is not wrong, because the basic circuit works exactly right. For the -55 to 150 degrees example using diodes, I've used the output and anode of the top diode, like it shows. I also tried 1 diode, germanium, and got no likely output voltages. Awful. Tried with 12V, 9V, 5V. Tried different chips, all work fine on the basic circuit. Thanks for your time anyway, will get back to you soon with the negative voltage arrangement.
 

Dodgydave

Joined Jun 22, 2012
11,285
i have used that sensor using one In914 diode and a 10k resistor, always worked for me ,take the output from the sensor pins into a dmm on mV range.
 
Last edited:

Thread Starter

Chris Gray

Joined May 10, 2015
6
i have use that sensor using one In914 diode and a 10k resistor, always worked for me ,take the output from the sensor pins into a dmm on mV range.
Maybe the 1N4148 is no good for this , or germanium. My meter is a decent one, could be the plug-in vero block I was using.
 

Thread Starter

Chris Gray

Joined May 10, 2015
6
I've just tested using a 3V button cell creating a negative voltage on the 'ground' pin, a 59k resistor (yes, 59k!) from output to -V and the voltage on the output. All soldered, seems to work fine. Certain it was the plug-in vero board I was using that was the cause.
 

Thread Starter

Chris Gray

Joined May 10, 2015
6
I've just tested using a 3V button cell creating a negative voltage on the 'ground' pin, a 59k resistor (yes, 59k!) from output to -V and the voltage on the output. All soldered, seems to work fine. Certain it was the plug-in vero board I was using that was the cause.
Ammendment to that, on the other end of the resistor, not the 'gnd' pin.
 
Top