I am working on Temprature controller using adc 7109. At present display is showing too much variations so have decided to take avarage of 20 readings.

Peter,

standard temperature sensors have a time constant of many many seconds. Is your resolution 1°C, then any display jitter would be an indication for a design mistake. But if your resolution is 0.01°C, e.g., display noise would be quite normal and an averaging would indeed help. Unfortunately, your posting doesn't tell anything about resolution of display. So, we cannot give you adequate help.

Please tell us: What is resolution of display? Or by other words: How high is variation of display?


Untypical display noise is often observed by people making a design mistake. With ICL7109 and similar products people often forget to connect analog GND and digital GND at some point. Or they use wrong reference voltage. Or their signal leaves common mode input voltage range. Another often made mistake is choose of poor capacitors. Have a look at datasheet and follow recommendations given there.
But the most classical mistake is to omit input RC input filter. They think it's opional ...

There was one situation I remember, when such chip gave wrong reading, although everything was designed correctly. Even when input terminal was grounded, a flickering of last digit of display could be observed. After long testing and checking with adding additional filters, providing even idiotic time constants, it prooved, that a small mains transformer located in too close distance to ICL7109 was the reason for unstable reading!! Unfortunately, a sheet of aluminium foil connected to analog ground directly at ICL7109 and located between chip and transformer could NOT remove the mistake, showing very clearly, that electric field of transformer was NOT the origin of noise. It was the magnetic field!
Only a piece of iron foil, connected to analog ground and providing a magnetic shield could solve the problem.

Interesting to note, that PCB was designed properly, containing a solid groundplane. Magnetic field of mains transformer induced it's disastrous EMF inside chip!

Much better than inserting this unwieldy magnetic shield would be to increase distance to transformer, of course. But in this case this wasn't an option...

Could you publish a schematic of circuit? This would tremendously help.

Kai