1) The difference between the AD7781(20bit) and AD 7780(24bit) appears to be
that in the AD7781 (20bit) 4 empty bits are appended such that the
communication length is equivalent. Does this mean that the 24bit is actually
more accurate or does it mean that those least significant bits are in any case
going to be "noise"? For the purposes of choosing an evaluation board does that
mean that the 24bit EVAL-AD7780 is the no-brainer choice?
2) I want to test directly with a Microcontroller and not connect to a PC at
all. What are the requirements for an external power supply? Is it possible to
route it through the ADP3303 or is that component selected only when the
jumpers are set for the USB power source from the PC ? (Part of the purpose of
using the board is to check the value added of that component, so I hope so!)
3) Is 5v excitation for the load cell generally preferable over 3v, further
increasing the resolution?
4) Are there any C language code fragments for the microcontroller SPI
interaction with the AD7780?
1) the AD7780 outputs 24 bits while the AD7781 outputs 20 bits, as you stated.
From the table on the front page of the datasheet, you will see that both parts
have similar p-p resolution. The extra LSBs of the 24-bit part are useful if
you plan on performing some averaging. Averaging by 16, for example, improves
the rms noise by sqrt(16). However, if you do not intend to average, then the
20bit part is fine.
2) The eval board includes the option to connect to an external power supply.
So, an external 3V or 5V supply can be connected to the eval board. Place LK5
in position A. The ADP3303 is only used to regulate the noisy 5V from the USB
to a clean 3.3V.
3) A 5V excitation voltage is generally preferable as it results in a larger
output signal from the loadcell and leads to increased accuracy.
4) I do not have sample C code for the AD7780. However, attached is some sample
c code for the AD7799. This part is similar to the AD7780/81 so the c code