noise on an A/D

Andywatson

I'm using a BS2 to communicate with TLC2543 A/D converter (with a 5.0v reference) that reads some voltages from 0 to 5 volts. One voltage I'm monitoring is a 5-volt regulator (a LM340) (I've got a resister-divider on it so the A/D is actually measuring half the voltage), and the other voltages are the analog outputs from a couple of sensors. I'm sampling the data every second and plotting it with the StampDAQ program (which is a wonderful program, by the way). The problem is that I'm getting alot of noise on all the signals. It's approximately 20mv, or about 16 bits, which is too noisy for getting accurate values from the sensors. What can I do to fix this noise? BTW, I know that adding an averaging routine to the code might work, but these sensors will eventually be monitored by something other than a stamp, so I'm looking for a physical way to fix the noise of the signals going into the A/D.

Thanks,
Andy

Newzed

The TLC2543 should have a 1K resisitor in series with the sense input, and the TLC pin you using should have a .1uf bypasscap as close to the TLC as possible.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sid Weaver
New Combo LCD Backpack

http://hometown.aol.com/newzed/index.html
·

Andywatson

The caps seemed to help it. It's down to around 5mv of noise.

I never know when to put caps in circuits. And if I figure one might help, I never know what values to use. Is there a tutorial on capacitors somewhere that might help me decide when, where, and how many caps to use on common circuits?

Philip

Is there any chance that what you are seeing is the harmonic distortion related to quantizing noise? It depends on the signal you are sampling I suppose but that is a possibility and on low level and especially "simple" signals quantizing noise can be significant. Can you dither it?