ADuC7022 - 13 GPIOs or 15?

When I look at the current data sheet, it says in the selection table
that the 7022 only has 13 GPIOs.  But I count 16 on the pin out drawing
and in the pin list.  Any idea why they say there are only 13?  Is
there some restriction on three of these pins that they can't count
them as GPIO?  I see that one is a JTAG TRST, one is a BM (boot mode)
at power up time and one is a MRST output.  These are the only
functions I can see that might disqualify an IO as being considered
"available", but I the way they are described in the data sheet I
expect all of these will work just fine as GPIO.  Is this why they
count 13 and I count 16?

1 P0.0
2 P0.3
3 P0.4
4 P0.5
5 P0.6
6 P0.7
7 P1.0
8 P1.1
9 P1.2
10 P1.3
11 P1.4
12 P1.5
13 P1.6
14 P1.7
15 P2.0
16 P4.2

Unfortunately, I'm not sure even 16 will be enough IOs.  They sure seem
to burn up a lot of pins on a 40 pin package.  10 are hard dedicated to
the ADC/DAC functions, 4 for JTAG and 10 more for power, reset and
clock.  I guess they wanted to maintain separate IOs for the analog
functions to optimize the performance of the ADC/DACs... that is what
is supposed to set these chips apart from the other ARM MCUs.  :^)

rickman wrote:
.2
>
> Unfortunately, I'm not sure even 16 will be enough IOs.  They sure seem
> to burn up a lot of pins on a 40 pin package.  10 are hard dedicated to
> the ADC/DAC functions, 4 for JTAG and 10 more for power, reset and
> clock.  I guess they wanted to maintain separate IOs for the analog
> functions to optimize the performance of the ADC/DACs... that is what
> is supposed to set these chips apart from the other ARM MCUs.  :^)

Yes multiplexing discrete I/O with A/D inputs pins degrades the
accuracy of the A/D. I have seen chips specify different A/D accuracies
on different pins depending on how much and what was multiplexed on
that pin.

However, you still can use (and I have) A/D inputs as general purpose
discrete inputs (if the reading in above, say,  1.8 volts the discrete
is high, below it is a zero). Same concept with DAC's as a general
purpose discrete outputs. The 7022 has a fast 1 Mhz A/D so it isn't so
bad, time wise, to do this, it's just a couple if statements for each
discrete.

steve

steve wrote:
> Yes multiplexing discrete I/O with A/D inputs pins degrades the
> accuracy of the A/D. I have seen chips specify different A/D accuracies
> on different pins depending on how much and what was multiplexed on
> that pin.
>
> However, you still can use (and I have) A/D inputs as general purpose
> discrete inputs (if the reading in above, say,  1.8 volts the discrete
> is high, below it is a zero). Same concept with DAC's as a general
> purpose discrete outputs. The 7022 has a fast 1 Mhz A/D so it isn't so
> bad, time wise, to do this, it's just a couple if statements for each
> discrete.

Yes, I know.  But it is hard to do fast work with ADC inputs and I
really need outputs and bidirs.  I need about 17 I/Os to implement a
parallel port to boot an FPGA from a NAND flash.  I have not done the
exact pin count, but it will need an 8 bit bidir data bus, about 5
outputs for the NAND control lines and perhaps another 4 or 5 to
control the FPGA.  Unless I can think of a simpler way to do this I'll
look for a processor with more I/Os and still under $5.  The Atmel
parts are pretty good and I think are lower power.  But they don't have
the 125C temperature range of the ADuC parts, so I have to add a $1.50
temperature part to shut down the board when out of temperature range.
Still that is cheaper than the ADuC7025.