leaflabsandy, good work. I'm glad you're digging deeper into the issue. It's both troublesome and perhaps the epitome of interesting when things change right under our purposefully observant nose, eyes, wave sensors, and we're left with an incomplete story, even if the change that occurs is central to the reason for investigation.
I'm not too impressed with the PCB assembly of Maple. We have a test setup to test every board before they go out, and we're interested in finding more weirdos to test for. I still cannot think of a reasonable explanation of how a Maple might work with USB power but not external power and then after washing it with alcohol to maybe remove some excess flux, the problem goes away. None of our hypotheses are compelling. I have some more questions about what you mean by "when it wasn't working with external power". Correct me if I'm inaccurate. I'm assuming you checked that the voltage at VCC was correct. Did you check anything else? Was the clock running? Did you check power consumption? You mentioned that the chip "wouldn't reset". Was the chip starting up...did the blue LED blink, like normal, at startup if you power cycled? what was the state of the reset pin...did it actually get pulled up and down by the button?
I just tried a 9V external supply with USB and both USB and EXT power jumpers connected...seemed to work fine. Don't try to shoot me. I probably have more rounds (17) than you:
http://dl.dropbox.com/u/42394/okieglock.jpg
I can't think of any short or capacitance due to flux causing this problem unless something was shorted to the power trace connecting the external power jack to the power selection header or to VIN. I'm basing this on what I know about the whole system. I know the circuit and layout well and have a very general idea of how the USB bootloader and serial stuff works. My main angle of attack with this problem assumes the legitimacy of the following scenario if our Maple boards and softwares are the same versions (I'm using Maple r3 and :
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If we're both able to compile and upload the same sketch that you posted, then if we then unplug the USB cable, plug in 9VDC to EXT power, switch the power jumper from USB to EXT, and press the reset button, then our boards should behave the same regardless of what operating system we're using.
It's kind of fun to try random things and look for "stuff", so i just did it for a couple minutes. i started connecting the reset pin to the resistors, caps, and transistors behind the USB port and to VIN and the protection diode terminals and any metal exposed in the vicinity through a 3.3k resistor and then pressing the reset button and watching the blinking LED start-up sequence. /Sometimes/, or about one in every 5 times i press the reset button when i'm connecting the resistor from the reset pin to the right pad of one of the four lower resistors behind the USB connector (R8,R9,R10,R12), the first blink of the blue LED is several times longer than usual. </garbage>
<garbage>
"You need TWO JUMPERS INSTALLED TO MAKE IT WORK WITH EXTERNAL POWER!!!!"
</garbage>
I wonder if someone is going to come across this forum thread one day and think this is "a regular ole' general fact of life" when it really only used to happen sometimes on leaflabsandy's bench.
Your data volume is very much appreciated. Unfortunately, you seem to be the only Maple with this power supply problem that magically disappeared. I say "magically" because we have neither a reasonable testable hypothesis on what was/is causing the problem nor enough data or a Maple that fails in this manner to explore it. Did you capture any images with your USB microscope that show the flux you removed? I have not seen a problem with excess flux on other boards. Believe it or not, I am going to apply a hefty amount of flux to the parts of the circuit you mentioned and others, but I think there's a very small chance of reproducing these things because I haven't seen this problem on any boards and no one can think of a way this kind of behavior could be caused by flux...or anything for that matter.
Happy 4th of July!
