If not needed, add a 10k pullup resistor to +V

If they're still going low, might be a typo in the code.  If writing directly to registers using hex values, he could have had the bits off a bit on hex digit.

Fundamental issue:  If he loads the old code back onto it, does it work?

If so,

diff -C2 oldcoce.c newcode.c

Look for lines, included libraries, "forgot abut that' commands, etc.

Remember when I was talking about long wires and interference issues a few years ago?

Put a 0.1 uF cap between power and ground, and another across each wire leading to the motor, and another across the wires at the motor.   Link is what I use, I tend to buy lots since they're used everywhere, but I don't think you need 1,000 of them, but having them in surface mount as well ad a ton with leads makes filtering a lot faster and easier (compared to SMD).  ~3 days from digikey to you.

Here's a kit that could come in handy by the time you add shipping..  Sadly.   MLCC/multilayer ceramic is what you look for, preferably Kemet, but that's sort of wanting best there is, most will work fine if they are the rated value, should say "104" on it (10, 4 zeroes = 100000 picofarad = 100 nanofarad = 0.1uF)  I generally don't trust unknown vendors, but these aren't critical.   This also includes the 10uF I mentioned above, so good to go and  your kid has enough caps to play with for maybe a year of projects.


The motors will work as generators when power is cut since that's what inductors do, a motor is a big inductor.  There's also the kinetic aspect, where power is cut, motor is spun from mass of system, it generates an opposite polarity voltage back to board which can mess with the controller.  Have a diode in parallel with the cap, cathode to + for each motor.

Also, add a 10-100uF cap to the V+ line to ground on both arduino board and motor controller, both 3.3v and 5v (I don't know if Arduino has 4.4, but if it does, it's just 2 extra capacitors.    You'll want the 0.1 in there even if you add a bulk cap of >10uF due to the way capacitors are constructed.  A ceramic 0.1uF cap is better at filtering high frequencies than a surge storage cap, there's enough on this topic to fill a few libraries, so have him check into that.

Long term interference is my guess if the old software doesn't work.   Possibly bad driver board, (now that it's been 'fixed') who knows.

Is the code for the motor driving library open source?  There could be something in there pulling the pins 2/3 low, which would overcome a pullup resistor.  Again, what changed, when, etc.

You said it was working until he fixed it.  That's a GOOD sign.   He's showing traits of being an engineer!   You Never learn anything or how stuff works if everything you do works right the first time.  

Maybe @GoatBoy can shoehorn that into your user title.

Make sure there's not something else in the code either polling or writing to the pins the interrupt is on for now.  Doing both can be done, just need to dance a bit.  Are the masks for which Interrupts and which type of Interrupt set correctly?  Picking edge or level, in/out, etc.  Is it tri-state high impedance?   If so, it requires a pullup or pulldown resistor to make the line impedance low enough to not be triggered by interference, controller interrupts are extremely sensitive to interference.

Proof:  Turn on a multi-meter, set to voltage mode, grab one lead, put other in air.  I can get up to 0.14 V from just all the RF around (which some call 'free energy'), from all the cell towers, power supplies, power lines, and radio stations around providing signals.

Now put a 100k resistor between the two leads and look at voltage when disconnected.   Same thing, a cap removes anything remaining if it was tens/hundreds of millivolts, as a resistor's function doesn't change with a sudden signal change while capacitors are frequency sensitive.  Combine them and it's a low pass filter, only 'slow' signals allowed.  Slow defined by values of cap and resistor, still in kHz for this example.

That inteference can be enough to trip an interrupt.  If you add a 100k impedance (instead of the 10MOhm of a tri-state input), the interference doesn't have enough "current supply" (source impedance)  to 'push the voltage' around more than a millivolt or two, which then goes to zero if the cap is added to ground on the interrupt line, response to interrupt delayed a microsecond, but better than false triggering.   More of a trouble shooting step.

Leave the caps across motor on both ends, one on motor end to stop interference before it goes down the wire creating RF, and one at the control board, to stop any residual from the filter on the far end, wire in between is a long inductor (aka Antenna).  With the two caps and wire, a lowpass pi filter is created.

On power leads, it keeps the supply voltage steady, so a motor suddenly starting doesn't trip it up, like how the house lights sometimes flicker when the compressor starts running.  The bulk and bypass caps on all supply lines eliminate that 'flicker' of power change, since a fast power change (step response) creating ringing in a square wave.

This is why even with digital electronics, analog and RF are still in effect and cannot be ignored by thinking "it's either 0 or 1".   If high impedance input is floating, a motor start/stop would bump that over the Interrupt trigger level, creating an interrupt, which could then keep interrupting itself if interrupts aren't disabled immediately in interrupt routine, and then enabled right before exit, I'm not sure if Arduino does that automatically or not.

The pullup or pulldown resistor will keep voltage as close to one rail as possible so stray interference can't bump it over the interrupt threshold value, and the caps prevent the step spike of a switch closing a long wire from becoming an oscillator.

Did you measure and test fit before going too far with things like glue and lengths?

--ETA:  Hello, TopMen, glad to see you not utterly failing.  

Looks like a flat end mill they tacked a point on.   Do they rip the outer circumference for first pass like a Forstner wood bit?   I've always used split point and a punch if it walks at all, didn't ever try those "gimmick" ones that can't be sharpened with a grinder and dremel.

LOTS.

Cockpit glow (greenish seems Appropriate, cabin lighting in red on BH side and "warm white" dimmed to nearly yellow on M1 side.   Nav lights, landing light done the same way you did for the train, but with a cold blue LED much brighter.

What's going to control the LEDs flashing/dimming/sequencing?   Designed for bright/daylight only, or will night tours be a thing?

--ETA:  Will there be pilots to bounce the light off what's visible on head?  With helmet and "driving helmet", covers most, so wouldn't need much - easy to sculpt in a week.  Otherwise, maybe have it in a different mode for no ppilots...  on ground?

No tint that I'm aware of.

Just thought of something.  Since it's going to be viewed at an upward angle from the floor, put a red light on the M1 side/center of the cabin "ceiling" facing the BH side window, and  a few yellowish LEDs on the BH side, angled toward the M1 side.  Looking from the ground I don't know the angles, but you can guesstimate, put a shield on the backs so penumbra isn't too harsh, but it could be seen a light is on.

The flight lights just do as a "helicopter" scheme is, No need to get fancy if BH and M1 have slightly different timing/patterns for fast night ID.

I didn't see your post about successful Window created until just now.  I thought I was at end of thread.  Looking good, but I'd suggest adding a larger smooth layer on tail football, or maybe 'spherical triangle panels' as on half a basketball (or two lines of latitude form equator to pole, with equator as bottom of triangle, 3D-2D transform) to make the edge/ridge textures look more.. intended?

--ETA:  Pictures at level like this see through other window, so no tint unless it's recent.

If wrinkles end up there no matter what, just pinch them so they "point" toward the center, or perpendicular to it, like lat/long again.  Little thing like that makes it seem more "oh, it's supposed to be like that" instead of "couldn't quite match the shape there".