AVR-Chat

David VanHorn wrote:
> 
> At 12:38 PM 5/11/2004 -0700, Bruce Parham wrote:
> 
> >Craig,
> >
> >I am currently using the code below to run the ADC on an 'M128 with external mux'es to expand
> >the system to 42 channels. The whole mess is scanned once per second (without putting the
> >cpu to sleep) and I'm only seeing +/- one bit of noise. Each mux input has a 10K/.1uF RC filter
> >and all of the filters and mux'es are (SMT) mounted on the top of the board right next to the
> >'M128 and are over and connected only to an isolated ground plane (1st inner layer below the top)
> >that is connected to the AGND pin. To my amazement, this setup worked so well that I didn't need
> >to do noise canceling.
> 
> This is pretty typical in my experience, just beware the "precision" reference on chip.
> 2.7V for a 2.5V reference, into a 10 bit ADC, is pretty funny, in my opinion.

No kidding! This is at the top of the ADC module:


  	// adc "precision" ref correction value in EEPROM

	#pragma data:eeprom

	//double  Ref_Cal_Val = 1.071094;     // Bd #1 = 2.742V
	double  Ref_Cal_Val = 0.9921875;     // Bd #2 = 2.540V

	#pragma data:data


Bruce