Hello everyone,
I have this new LM35 temperature sensor 10pcs. This is my very first encounter with this specific component so I don't know what readings to expect from it. I manage to find this component in the Proteus list of components and in simulation, it showed a linear or equivalent increase in output voltage with LM35 temperature. More specific, 10mV per 1°C.

In reality, I'm not getting this result at all. I'm getting 0V reading.
I start to think my components are bad?
But I also manage to make them work !!! at least to collect a more correct voltage from it's output, Im not sure how correct is the temperature reading converted in mV, but is very close to what direction of working the simulator showed me it should be like.

So if the opamp is unpowered from either + or - rail, the output reading of LM35 will get back to 0V.
I also don't have to collect the voltage from the output of the LM358.
I could amplify x10 and collect 250mV from LM358 output, to correctly display the value as shown in my simulator... I didnt do it yet. Is a thought.
x10 amplification
I also suspect a uA741 opamp will also do the same work, because I have a lot of them and unused. I didn't test it yet.
It's enough to only connect that LM35 output to one of the inputs of the (powered) LM358 and I can read the LM35 output in mV scale with my DMM, some value different than 0V as in the top naked case.
My question to you: - is this normal or correct ? I feel it is not.
I wish that you make a reading of your LM35 like in my first experiment and show me a picture of your output reading. To clearly show me how a good LM35 should work. Because Im not convinced at this point how it should work.
Thank you !

 

Hi q12,
With a 5V supply and the basic circuit, the LM35 Vout should be 10mV*20Cdeg = 200mV.

If you add an OPA you can amplify the voltage.
What temperature range do you want to measure and how do you want to display it?
E

 

What temperature range do you want to measure and how do you want to display it?

hi, The range I need it is up to 100*C in steps of 10*C's, like 30, 40,50... for example.
Can you test your LM35 and read directly its output and tell me or make a picture if it is indeed 1mV or 10mV per 1*C ? or as you mentioned 10mV*20Cdeg = 200mV ?

 

Hello,

Does this page help?
https://danyk.cz/lm35_en.html

Bertus

 

hi q12,
This is a more complex circuit that covers -20C through +100C and gives a ~0V through +5V out suitable for a microcontroller ADC input.
E

 

hi q12,
This is the basic LM35 0C to 100C
E

 

This is a more complex circuit

hold your horses @ericgibbs
All I want to know is if this LM35 temperature sensor can be read in reality with a DMM !!!
Exactly (or close enough) to my simulation reading.

In my case, I only got a 0V reading when tested directly like that.
Or maybe there is a trick I dont know how to properly read or setup these LM35 sensors? From my simulation, it doesn't seem so.

 

All I want to know is if this LM35 temperature sensor can be read in reality with a DMM !!!

hi,
Yes!
Add a 10k load.
E

 

hi,
Yes!
Add a 10k load.
E

Exactly what gut told me when I read the OP.

 

Add a 10k load.

Smart idea !
Here are my results:

It would be very good to show me your --real-- readings !!!
It is clearly some voltage divider going on, when Load Res is connected to +5V rail.
Although, in simulator is showing a 250mV output and not 0V as in reality:

 

hi q,
When trying to open your attachment, I get an Error 404 message>
E

It would be very good to show me your --real-- readings !!!

What is wrong with reading from a graph plot????

 

When trying to open your attachment, I get an Error 404 message>

refresh this page - I edited a lot of stuff - must have remained something from before....

 

hi
I don't understand why you find this confusing??
Add a 10k from the Vout to 0V.;
E

 

I don't understand why you find this confusing??
Add a 10k from the Vout to 0V.;

Nothing is confusing.

In my circuit I did just that ! Pay more attention to my circuit and you'll see its exactly YOUR circuit, It's only drawn differently a bit.
I also added supplementary test to + rail and floating.

 

I was wondering why it worked fine with LM35 output linked to one of the inputs of LM358... hmmm
Those are pnp transistors and the current goes from the input UP towards to + rail through those transistors. Is how I see it.

 

Видеть

 

I am partially convinced that my LM35 are bad in some way. Either from factory as reboot, or from ESD or even used, although they look brand new.
I am also suspecting that the internal circuitry of my LM35 is incomplete or faulty somehow. Thats why I can read some derated results. In principle they are good, but in practice ... they are bad. Is my opinion.
All I wanted from you is a confirmation of how they should really be !
A reality check with some models that on your side are doing what my simulator is telling me they should do. (clear as mud?)

 

Did the TS actually use the circuits shown in the applications portion of the data sheet??
One caution I offer is that with such very low voltages and small currents, attention to the connections is very important.

 

My experience with LM35s has been with using them in the Arduino environment. They always had to be equilibrated to bring their range into calibration by scaling equations. @ericgibbs circuit with 2 opamps and a potentiometer does that without using the signal scaling equations needed in the Arduino's microprocessor analog input. I too was surprised at this and still wonder how a "standard" sensor that has many uses but needs external scaling to bring it into calibration has such a large market share. The TMP36 is just as bad. But then they are in a 3 terminal TO-92 package and not on a larger chip format with an input pin for voltage calibration adjustment. They are cheap components so the old adage of "You get what you paid for" is applicable.

 

This is what my LM35CA is showing (without a 10k resistor)


And with a 10k resistor:



As we can see 10k makes it more "accurate".