On 11/14/05, microsoftwarecontrol wrote: > I don't believe it is worth to add a op to > divide V for ADC acting as a load! > > I think ADC as a fixed input Z load! > Simple 100K 10 turn pot as divider, > has below part parallel with ADC's Z, > it will give accurate enough V dividing! > This is a bad idea, for at least four reasons: 1) It is an unnecessary adjustment. 2) It is an unnecessary expense. 3) If someone turns the pot the wrong way, the full 12V battery will be applied to the ADC input, resulting in destruction of the microcontroller. A good design should not be destroyed by misadjustment. 4) I quote the Complete Mid-Range Reference Manual: "The maximum recommended impedance for analog sources is 10 k." A 100k potentiometer can give up to 25k Thevenin impedance. You can expect some noise if you do this. > > By adding a op, you add error as well! You add the offset voltage of the op-amp. Op amps with less than a millivolt of offset are inexpensive and widely available. > And I believe, this error is not smaller than > above, where simplifying ADC as a pure R. Measurement noise will be less, compared to driving the ADC with a too high impedance. I learned the hard way, too high impedance will lead to several LSBs of noise on the measurement. Here's my 'simplest possible' design: >From 12V to ADC pin, 6k 1% resistor. >From ADC pin to ground, 4k 1% resistor. Bypass the 4k resistor with a 0.1 uF capacitor. This will spill 1.2 mA through the voltage divider. If your reference voltage is Vcc, accuracy will be limited by the 7805's accuracy, which can be as bad as +/- 4%. The ADC voltage will be 4.0 volts at 12 volts input. I like to leave a little headroom like this, because you never know when an overvoltage can happen. If severe overvoltage is possible, a 5.1V zener will protect the microprocessor. Regards, Mark Rages markrages@gmail -- You think that it is a secret, but it never has been one. - fortune cookie -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist