Interesting idea. Maybe more like a debugger. The higher-end ides like visual studio do some of this.

I can recommend brey terminal as a more capable monitor but it is still a stream of text.

Wow!

Could you reduce the number of pins dramatically by using something like spi?

Post a schematic and the scope trace.  What output do you get from your code and from the mini-tach.

I run mine at 3.3 using a small red led to drop the voltage. The current draw is very small so the led's barely visibly lit.

There's an excellent writeup linked to by a thread On this site - search for interrupt latency maybe. I believe the basic answer is several microseconds but there are ways to improve it.

I still don't quite believe this but the documentation for 1.0 sounds like the pullups are *not* active for pinMode(INPUT).

They are not active by default. They can be activated, using digitalWrite(). Or, the call to digitalWrite() is made for you if you use the type INPUT_PULLUP.

Whew, thanks.

I think that the "disabled" in that sentence refers to the fact that there is now a call to digitalWrite() in the pinMode() function. It is called with HIGH or LOW based on the use of INPUT_PULLUP vs. INPUT.

A peak at the code confirms this. The digitalWrite() function isn't actually called, but the equivalent register manipulations occur.

Where can I find info about "Hardware features" like the internal resistor, i guess it is not the only hardware feature on the board?

Start here: http://arduino.cc/en/Main/Hardware
Pick whichever board you have, and read all the details, even if you don't think you will use them (yet).

I still don't quite believe this but the documentation for 1.0 sounds like the pullups are *not* active for pinMode(INPUT). This would break a lot of my http://arduino.cc/en/Reference/PinMode
.

Actually, just a 1K or 2.2K series-R will probably work.

thanks - i bet it would, certainly my circuit is not much different from a 22k.

i'm using a parallax ping distance sensor which has a single signal pin.  You pulse it high for 5-15 us then turn and listen for IT to pulse the same line high for a length of time corresponding to the distance.  http://profesores.fi-b.unam.mx/m3615m/datasheet-sen136b5b.pdf

I'm using an mcu with dedicated input and output pins so i can't quite do that.  I'm planning to compensate with the circuit above, sending the pulse out on one pin and reading the response on another.  The trigger output from the MCU is sent high for 15 us then the MCU lowers it to 0.  The resistor pulls the pin down to 0 but the the ping sensor should be able to overcome it and the diode should stop the MCU's 0 from interfering.

This looked ok on the scope except the high pulse from the MCU was down to about 3.3v by the time it got to the sensor - the response pulse looked fine and fat at near 5v.

I thought all was fine til the sensor stopped working.  Trying a spare, it's fine so it looks like I borked the first one.

Does this circuit seem safe?

 Is that a RS232 chip for serial conversion?  If so, usually there are some capacitors involved for it to work properly.

The soic pattern in the bottom left is a spare spot for experimenting. 

Add the top & bottom ground polygons. If you have to cut a trace and add a wire, the ground plane will not make a difference. You could lay the 7805 flat on the board than and use the ground planes to help cool it.
The design needs some more caps - on the VCC, AVCC, AREF, and 22 pf caps on the XTAL1 & XTAL2 pins.
I see an ICSP connector for downloading a bootloader, or a sketch that starts in place of a bootloader.
I see a serial interface if you want to download a sketch that way via the bootloader. The pinout used is not standard. You might want to revisit that so you plug on a standard part like an FTDI cable or an FTDI Basic. If you're planning on making your own adapter cable you're all set.
A Diode across the pullup resistor will make for more reliable downloads,

Thanks very much for looking.  I'll add the polygons.  
I thought I had a cap on vcc but i'll add it.  The resonator has the caps built in I thought.
Holy cow if that serial cable layout is wrong that's a big oops - it's meant to be the same as my RBBB - i'll get right on that!

Thanks again

whew: I'm pretty sure the serial connector is ok - it matches my prototype which works with the same cable as my RBBB.  I guess it's not well labelled though - I'll add some clues for when I forget>
from top to bottom goes ground, skip, 5v, PC TX - AVR RX, PC RX - AVR TX, and RTS/RESET.

I've been putzing at this thing forever and I think I'm ready to send it off to Seeed for fabrication.  It's an arduino clone with ttl serial interface along with a 4021 shift register and a 74c02 nor gate connected to a custom 30 pin interface to another microprocessor board.  The routing is a combination of auto-router and my efforts where I thought I could make a difference.  The ground and +V traces are 24 thou, the rest are 16. 

One specific question is whether I need ground polygons.  I'm tempted to leave them off because I figure I will have done *something* wrong and it will be easier to fix with just the working traces on the board.  Aside from the crystal there wouldn't be any very high frequencies on the board.  The  other processor is running at a few hundred khz.

The AVR is a dip but i used soic for the other chips because they're small and easy to solder.  The area at the left will be covered with a small chunk of proto board for connecting to the other micro at the 30 pin connector.  I stuck a small grid of pads and an extra soic pattern in there because there was room.

If anybody has any suggestions or observations I'd be grateful.  I'm going to print the gerber image off at actual size and lay out parts on it as another check before i commit.

I think your problem is at the arduino end rather than the pc software. Try reading/printing your data as char's rather than byte or int.

Presumably there's a motor driver between the arduino and the motor. I believe mine has specific settings for braking although i've never used them.

Worst case, just reverse the current.