Forums

LPC2124: A/D Input impedance

Started by fl429 May 27, 2005
I did not see any specification in the Datasheet on the input
impedance of the Analog inputs. When I switch A/D channels, this
impedance will determine how long the sampling time should be before
the conversion can start. When in Burst mode, there is no mention on
the sampling time at all.

Can I assume that the 11 clocks already take care of this ?

Any ideas ?

Thanks

Lei


An Engineer's Guide to the LPC2100 Series

At 09:41 PM 5/27/05 +0000, fl429 wrote:
>I did not see any specification in the Datasheet on the input
>impedance of the Analog inputs. When I switch A/D channels, this
>impedance will determine how long the sampling time should be before
>the conversion can start. When in Burst mode, there is no mention on
>the sampling time at all.

Since this is really a single A/D prefixed by a sample and hold and they
don't mention any input buffering it is VERY unlikely that the inputs are
high impedance. The usual configuration would effectively simply be a
sampling cap preceded by a couple of switches with a finite
resistance. Unfortunately, I don't think either of those values are
documented. Hint to Philips those are useful figures to have, even if only
a range is provided. In particular it would be really nice to know what
the sample capacitance upper limit was.

>Can I assume that the 11 clocks already take care of this ?

Not w/o a little bit of care. What I usually do is use an RC filter as the
last, or only, element of my input Nyquist filter. If the cap on that is
next to the input and reasonably large it provides a large enough charge
storage that it effectively nulls out the source impedance. IE the voltage
change from a 0.1uF cap charging or discharging the sample cap, which is
usually only a few pF, is less than a single bit. At that point the only
element that limits the input sample time is the RC of the input switches
and the sample cap. The response time of the signal itself will be limited
by the RC of the input filter.

One of the clearest outlines I've seen of the typical input interface is
actually in an Intel app note:
http://www.intel.com/design/mcs96/datashts/27211501.pdf They don't,
however, mention the need for Nyquist filtering to reduce sampling artifacts.

Robert

" 'Freedom' has no meaning of itself. There are always restrictions, be
they legal, genetic, or physical. If you don't believe me, try to chew a
radio signal. " -- Kelvin Throop, III
http://www.aeolusdevelopment.com/