Re: [avr-gcc-list] ADC anomaly

[Douglas Dotson]

Things just got even stranger. I measured the
1.22v internal reference on channel 31. Measures
low by 22 counts. Reads 477, should be 499. I'm
starting to question the usefullness of this ADC
as 10 bits.

Doug

--- Douglas Dotson <address@hidden> wrote:
> I'll give that a shot. My input is slow moving DC
> so I wasn't thinking about noise being aliased.
> Strange that the other 2 inputs do not show this
> effect though. And that it always settles down to
> the
> same point that is 3 or 4 counts low.
> 
> Doug
> 
> --- Richard Urwin <address@hidden> wrote:
> > On Sunday 15 Feb 2004 4:44 pm, Douglas Dotson
> wrote:
> > > I'll give it a shot. I haven't brought many of
> the
> > > other on-chip peripherals online other than the
> > > USART. I am filtering the readings through a
> > > simple digital filter which is working well.
> This
> > > problem is very slow moving. I have 3 channels
> > > of the ADC operational and the other channels
> > > are not showing any similar behavior. Strange.
> > 
> > Make sure you have hardware filtering on the input
> > at least two octaves 
> > below your sampling period. No amount of software
> > filtering can remove 
> > aliased noise from the signal.
> > 
> > For example, if you have a 1kHz sampling rate, and
> > there is 999Hz noise 
> > being added to that, you will end up with a 1Hz
> > signal on your data. It 
> > would not be unlikely for the signal or a harmonic
> > of it to be even 
> > closer to your sampling frequency. Mains hum is
> very
> > accurate, maybe 
> > only a few micro hertz out; dimmer switches create
> > lots of harmonics; 
> > fluorescent lights often work in the kHz range.
> > 
> > -- 
> > Richard Urwin
> 
> 
> _______________________________________________
> avr-gcc-list mailing list
> address@hidden
> http://www.avr1.org/mailman/listinfo/avr-gcc-list