Is my LM35 sensor damaged or not?

Projects General Guidance
1

I received 10pcs of LM35 at the end of 2024.
image

LM35-Pinout
LM35-Pinout299×380 8.52 KB
and I tested them with a very basic program in arduino:

const int input = A0; 
void setup() {
  Serial.begin(9600);
}
void loop() {
  float temp = analogRead(input);
  Serial.print("\ndata = " + (String)temp);  //°C
  delay(1000);
}

The circuits on the web suggests direct connection so I did the same.

LM35-Arduino
LM35-Arduino687×658 48.9 KB

I obtained this results:
image
image168×983 17.1 KB
image
image254×984 20 KB
where the reading was on and off and on and off, repetitive like this. When is off I get all 0's, when is on I get some random numbers - I was expecting them to stop at a certain value, whatever that value is, and not fluctutate too much from it. If its a small deviation, its fine, but what I got is pure randomness.

  • This is the first time Im working with these chips that look like a TO-92 transistor.
    I have no idea what to expect from them. Is this a normal behaviour and I dont know something important to add to this circuit or code?
  • These were actually the first thermic sensors I tested. Then I jumped to thermocouple sensors that I discussed here. Also my first time with those too.
  • I looked on some videos on YouTube and it appears that these LM35 sensors are quite linear and they actually output a direct temperature value, without too much conversion, or probably a minimal one. If I understood it right. In its datasheet also shows a direct connection to a LED driver LM3914 a classic logic IC :
    image
    image687×789 58.2 KB

After a couple of days of testing and many-many-MANY hours of trying to understand why I got those successive results, I concluded these LM35 chips are severely damaged. I tested like 5 of them from 10pcs package I receive and they all give me the same result in test.

  • Now, I still dont believe 100% they are damaged. So I turn to you. Im betting on a mistake I made in code or on the circuit breadboard that I dont see it and I am not aware of. You are definetly more experienced with these components and you know exactly what are their behaviour and what to expect from them. Is my result here due to a wrong code or circuit? or they are truly burned/damaged?
  • I also tested them with my osciloscope. If you put a POT on the osciloscope probe, and twist it, you will see a voltage ramping up or down. So that was my reference image to have in my mind while I was testing these LM35's. I did not see any ramping whatsoever on their outputs. I used my fingers to heat them slowly = nothing; then my hot iron at 333*C and again nothing, no ramping, no response. I only seen some voltage spikes but Im uncertain if those spikes are from general interference on the probe or from the actual output.
  • On my arduino program here, it gives signs it is working for a few seconds then dies for another couple of seconds then alive again and so on... very strange !
    To be certain it is not a wire or A0 port creating an erroneous reading,
    I link A0 to a 10k test potentiometer
    image
    image207×244 35.3 KB

    and the values were maintaining constant very close to the dialed resistance division! The change was 400-405 (for ex). So very small reading instability, compared with the LM35 output which is total random.
    Thank you !
2

There have been many reports on the forum about fake LM35s

1 Like
3

I would say the wrong circuit.
The data sheet suggests these two:-

LM35
LM35730×974 40.9 KB

Try this before dismissing them as fake.

2 Likes
4
  • With this exact problem I have here?
    So they are busted then?
    I want a Yes-No confirmation,please. And thank you !
5

I see, good idea !
What is that mechanical dial there? image ; it's an Ampmeter or a Thermometer ?

6

Yes.

Make sure you are using the right pins.

Pinout
Pinout558×432 21 KB

1 Like
7

Not all. Search the forum you will see all types of problems.
If you buy from Digi-Key you will pay $13 (USD) for 10 pieces.
What did you pay?

1 Like
8

I put a text image in my OP that shows the price too, but for convenience, here:

2.92$ for 10pcs

9

Im always careful on every component pins.
Its a life and general rule I respect and Im careful all the time.
So its definitely not that ! (unless they inversed the pins in the factory) but we ca nnot know that. But remains a possibility - haha.

10

Most show as you have in your "copied photo" in original post but........as simple as it it, do you have the device actually connected the same way. Show photo and include any power connections.

11

Is that US currency?
If yes, there is a very good chance they are fake.

12

yes 2.92USD for 10pcs

13

So from Digi-Key I pay $1.30 each and you paid $0.29
Do you think they could be fake?

1 Like
14

I dont know ! They seem legit to me :slight_smile: They have the LM35 laser marking on them !
What Im most interested here is their behaviour to compare it with mine.
But I get the feeling they are busted or completly fake to begin with. Aaah. What a waste of time.
Although its a lot of time passed from then, I will still ask my money back and make noise toward that unscrupulous chinese. I dont expect money back but they will get a piece of my mind for this treachery !
Thank you again @jim-p !

15

Sure - here:

20250307_122504
20250307_1225041920×1080 273 KB

16

I guess I'm spoiled, living here in the US. I only buy from Digi-Key and Mouser and both are authorized distributors for Texas Inst.

1 Like
17

Just to make sure the Breadboard strips on the left half is connected to the R-Side?

Add some smoothing caps.

18

Their laser marking:

20250307_123752
20250307_1237522048×1152 304 KB

Truly looks ok to me !
Left ones are all tested and also the plastic a bit burned from my hot iron tests I mentioned earlier. If you zoom into the image you can see the burn marks.
Right ones are untouched.
20250307_123810
20250307_1238102048×1152 400 KB

19

Follow the wires:

image
image336×458 72.2 KB

Actually in my original tests I did include smoothing caps and I also followed this reference diagram too:

image
image708×219 14.8 KB

It is indeed smoothing and I get a constant reading, apparently !!! - why? because in reality if I waited and look on the serial monitor output, the apparent constant numbering were actually moving down but very-very slowly !!! and after some longer time, in 10's of minutes, like 30, or 40 mins they were starting to raise up, still very slow. So that capacitor dampening it is really working, but the behaviour is the same, only greatly slowed down. I told you, I tested them many-many-MANY hours ! a few days.
My biggest regret now, is that I didnt make these tests sooner when I got them. To fit in return money back time. Aaah...my money! down on the river !

20

That is National Semiconductors logo not TIs, so those parts could have been sitting in someones basement for the past 20 years.

1 Like