I am trying to understand a temp sensor ckt in TI app note AN56.
I have attached the app note ckt and my simplified ckt.
So the sensor has a constant voltage bias. Vbe will be something around 0.65v at zero C. To make the output of the opamp zero, another reference but with opposite polarity is used along with pot R4.
R3,R5,R6 decide gain. I don't understand how R4 will zero the output of the opamp.
7220.opamp temp reduced2.pdf
Hithesh,
I think I agree with your observation. A value of 10K or 15K is probably more appropriate for R4. It appears that there would be about 0.57V at the emmiter of the 2N2222 and this would produce about 190uA into the summing junction. To null this you need to pull 190uA from the summing junction through R4. If R4 is working against -1.22V then the least you can pull out will be -1.22/5K=-244uA which is still too much. 10K is probably the right value for potentiometer R4 and it will likely settle to 6.42K when adjusted to zero.
OK but the real answer to your question is that you should consider much newer devices. The parts in this schematic are from the 1970's (yikes). This app. note was written in 1971! Texas Instruments makes some really great analog and digital temperature sensors that are really easy to use, require no adjustments, and are more accurate than the circuit shown in this venerable (but dated) application note.
For a simple analog temperature sensor with output voltage proportional to temperature, consider LM20. TI also has parts that will work with a remotely located diode wired transistor (as shown in AN-56) as well as 100% digital devices. Many of these parts are in SOT-23 package--way smaller than all the parts shown in AN-56 and cheaper too.
More temperature sensor information here:http://www.ti.com/lsds/ti/analog/temperature_sensor.page?DCMP=TIHomeTracking&HQS=Other+OT+home_p_temp
I hope this is helpful,Tom Mathews
Tom,
Thanks for the reply. I was just trying to understand the ckt.
With regards to temp sensors, is there anything that's +/- 0.1 C accuate?
Check to make sure your op-amp has a negative rail. If it has ground as the negative rail then it probably cannot pull all the way to 0V. I'm not sure which simulator you are running but it may also be simulating the limitation of the rail to rail output. Check also the node voltages around the op-amp to see the op-amp model is faithful to what a real op-amp would do.
I'm not sure what your application is. For high accuracy I would consider digital temperature sensors like the LM92:http://www.ti.com/product/lm92
The LM92 has readout resolution of 0.0625C, however,accuracy for this type of sensor is 0.33C at 30C.