Help needed, CA3162 CA3161 LM35 Thermometer

[goodoltup]

Greetings forum members!
I am a new member, and first time poster. I am working on a thermometer circuit for a university project. I have constructed a thermometer circuit used to read in Celsius to complement a solar cooker project we are undertaking in my engineering class. I built the circuit on strip board, using a published circuit diagram. The thermometer as assembled powers up and reads digits on the LCD display, and I can use the zero adjust pot to change the temperature.
My problem is this; the thermometer can be adjusted to read zero with the probe in ice-water, but then when placed in boiling water it reads over 500 to 600 units. No adjusting with either the zeroing pot nor the gain adjust pot can make it read zero in ice water and 100 in boiling water. It still responds by increasing digits in hot water and decreasing digits in cold water, it just seems to be logarithmically off somehow.
I have attached the schematic that I used. I made a substitution where C2 is now a .33uF cap instead of a .27uF cap. I also included a 5V regulator with a LM7805 to bring the 9V battery down to the correct voltage. Everything else is exactly the same as the diagram.
I'm not sure what could have gone wrong. I am a relatively new at this, and I don't fully understand how the components interact. I have seen the CA3162 and CA3161 used as a digital multimeter, and some of those diagrams use voltage dividers before they input to the CA3162. Could it be that a voltage divider is necessary, or could it be that this thermometer as constructed can still be adjusted to read in Celcius?
Any and all help would be appreciated. All questions will be answered promptly. I can also provide more pictures and diagrams, as well as a parts list, if requested.
Thank you in advance,
Joshua in England

View attachment 60521

View attachment 60520

 

[ericgibbs]

hi and welcome,

The LM35 when using the basic circuit from the datasheet is only specified down to +2C.

 

[goodoltup]

Thank you for your reply,
Does it appear from the circuit diagram that I used that it is in the basic configuration? There is no resistor between output and ground as constructed, so perhaps it is. Even so, the application is a solar cooker, so it only has to measure temperatures between 19 ( room temp ) and 120 or so.
Do you have any ideas about how to adjust it so that it reads in Celsius?
Kind regards,

 

[ericgibbs]

Thank you for your reply,
Does it appear from the circuit diagram that I used that it is in the basic configuration? There is no resistor between output and ground as constructed, so perhaps it is. Even so, the application is a solar cooker, so it only has to measure temperatures between 19 ( room temp ) and 120 or so.
Do you have any ideas about how to adjust it so that it reads in Celsius?
Kind regards,

hi,
With the 0V pin of the LM35 to 0V it is the basic circuit.

As you know the LM35 outputs 10mV change for each 1Cdeg change in temperature.

So for 19C , you should measure with a meter 19 * 0.01 =0.19V and at 120C * 0.01 = 1.2V

So it does read in Celius, how do you plan to display the temperature [ or use the change in LM35 output]?

 

[JimB]

Do you have any ideas about how to adjust it so that it reads in Celsius?

Just expanding a bit on what Eric said:

So for 19C , you should measure with a meter 19 * 0.01 =0.19V and at 120C * 0.01 = 1.2V

If you are using the circuit shown in your original post, disconnect the output of the LM35 from pin 11 of IC1 and connect a "voltage injection box" to pins 10 and 11 of IC1.
1 Inject 0.19v and adjust for a display of 19.
2 Inject 1.2v and adjust for a display of 120.
3 Repeat steps 1 and 2 untill the readings are correct through the range.

Disconnect the injection box and reconnect the LM35.
Dip the LM35 in to boiling water to confirm the 100C point is correct.

The "voltage injection box" could be somethng as simple as a 10k potentiometer and a 1.5 volt battery.
Your university electronics lab may have something a bit more sophisticated.

JimB

 

[Reloadron]

I may be off base with this but reading the CA3162 data sheet I see the max input as 999 mV.

• Overrange Indication
- “EEE” for Reading Greater than +999mV, “-” for
Reading More Negative than -99mV When Used
With CA3161E

And as to calibration:

1. Apply 0V across V11 to V10. Adjust zero potentiometer to give 000mV reading. Apply 900mV to input and adjust gain potentiometer to
give 900mV reading.

So with the LM35 output of 10 mV/C the highest measurement would be 99.9 degrees C, if I read things correctly. It would be a close fit boiling pure distilled water at sea level as I see it.

I may have this wrong?

Ron

 

[cowboybob]

Since your feeding the LM35's output into an A/D converter (usually really high impedance), maybe try a small load resistor on the the sensor's output to calm it down (my term, not a recognized professional term...).

 

[Boncuk]

The temperature of boiling water decreases with increasing elevation.

As a rule of thumb you might use 1deg/C per 1,000 feet of elevation.

So at 2,000 feet water will boil at (roughly) 98deg/C. The correct value is 98.02deg/C.

Taking this into consideration the max input voltage of 999mV for the CA3162 might not be exceeded when measuring the boiling temperature at altitude.

Boncuk

 

[goodoltup]

Thank you all very much for your suggestions,
I do have an update. I replaced both the 10K and 50K variable resistors with 22-turn types to help make the adjustments finer. I also isolated pin 10 on the CA3162E so that it only connects with the LM35 and its own ground ( It was sharing a ground and a connection point with another capacitor. ) I also isolated pin 13 on the CA3162E, so that it only connects to its variable resistor, which is in turn connected to its own ground. I did these things just to make sure nothing was back feeding anything else or making a weird connection.
I can verify .23mv on the LM35, and I can adjust the LCD display to read 23. ( ostensibly in degrees C ) When I put my fingers on the LM35, I can see its voltage increase at a normal rate that you might expect from the heat of my fingers. However, the LCD races up in 2 - 5 step increments until it is well over the actual temperature. No adjusting at the 10K pot will change this. I have tried it at measured 0, 5K, and 10K values, and the screen still just races right up to 50 in a few short seconds. I cannot tell wether it happens slower or quicker with more or less resistance.
Does anyone think that it is down to the gain adjustment? And can anyone ascertain whether I should try a larger or smaller value in the gain resistor if I need to try a new pot?
Could it possibly be the .33uF cap where there was a .27uF cap specified?

 

[Reloadron]

Regarding the cap.

NOTES:
1. The capacitor used here must be a low dielectric absorption type
such as a polyester or polystyrene type.

That cap is your integrating cap and a key part of how well the chip will work.

Also, without the LM35 connected to the circuit have you tried the following per the data sheet:

NOTES:
1. Apply 0V across V11 to V10. Adjust zero potentiometer to give 000mV reading. Apply 900mV to input and adjust gain potentiometer to
give 900mV reading.

Ron

 

[goodoltup]

UPDATE!
I replaced the .33uF cap with a .27, as specified. I also put pin #6 on CA3162 to +5V, so make the chip run in the 96Hz mode. These things made no real difference, other than it updates very quickly now.
I replaced the battery. I had not noticed that it made it down to 4.5v, and that the 7805 regulator was cutting it to 3.2v. That was stupid on my part. Then after hours of adjusting, I realised that the number one digit was not coming on sometimes. I ended up soldering two of my spare meter leads to the board to constantly monitor the LM35, so I could compare it to the screen. After another hour, after checking all my connections again, I pushed the LCD display all the way into the socket. Immediately all three digits lit up, and I could now see that this thermometer reads up to 99 degrees C, to the tenth of a degree. I lit up the number one digit decimal point by hooking it up to ground, and it magically made sense. Now it reads exactly what the LM35 puts out, .214vdc on pin 11 on CA3162, and the display reads 21.4, room temperature.


In summation, it is working perfectly, and much more brightly. I think that I spent over 40 hours on this project, and I am happy that it now works to my satisfaction. This could have been avoided if I had just pushed the LCD all the way in to begin with. I did't want to force it at first, and later, I fell into the trap of not double checking it.
I already used Solidworks and the laser cutter at school to make a plexiglass base for it to sit on, and tomorrow I am vacuum holding a cover for it. I will update again with pictures.
Thank you all for the suggestions, I investigated every single one.
Cheers! Josh

 

[Reloadron]

Glad it worked out for you Josh. So it took longer than expected but the learning curve was at work and you come away from each project with lessoned learned beyond what the book covers. Good going and pictures would be nice.

Ron

 

[goodoltup]

View attachment 60709
Here it is with the LM in some warm water. Celebratory Guiness in the background.