I recently was fortunate enough to receive two MSP432 Launchpad development boards, and I've been hoping to use them for a few calculator-related projects. The most obvious task is to try using Energia, TI's Arduino IDE fork for their Launchpads, to port ArTICL programs to the MSP432. A slightly more complex goal that I've set myself is to try to control DCC-enabled model trains with the MSP432 and a graphing calculator. I'll be documenting my adventure in this topic.

First, here's what I've found so far in futile efforts to get a simple PWM interrupt working within an Energia sketch:
Timer_A "Documentation":
https://github.com/machinaut/msp432-driverlib/blob/master/driverlib/MSP432P4xx/timer_a.h
https://github.com/machinaut/msp432-driverlib/blob/master/driverlib/MSP432P4xx/timer_a.c
Information about AVR timers to compare:
http://maxembedded.com/2011/07/avr-timers-ctc-mode/
TI Forum help about MSP432 interrupts (doesn't work with Energia?):
https://e2e.ti.com/support/microcontrollers/msp430/f/166/p/428511/1531341

Because it seems to be documented in shockingly few places, here's how you translate bank+pin-style pin names (ie, P1.0 or P2.4) to Energia numeric pin numbers (for example, the LED pin is not 2, as some documentation claims; it's 7:
http://energia.nu/wordpress/wp-content/uploads/2015/03/MSP432-LaunchPad-%E2%80%94-Pins-Maps.jpg

Edit: I'm currently working on getting a simple ArTICL sketch working first, and not having much luck:
- Timeouts have been temporarily lengthened; because they're based on numbers of loop iterations, rather than the correct solution (using timers), they need to be longer on faster uCs.
- Sends appear to work properly.
- Gets do not work. Using the existing code largely yields no bits received at all. Switching to a scheme where pin reads involve a unanimous voting scheme over 16 samples helps a little, in that a single byte is received successfully, except that the byte received is a nonsensical (but consistent) value.

The current plan is to load a known-working ArTICL build onto an Arduino and try to get it to talk to ArTICL on the MSP432 to debug what the non-MSP432 endpoint is receiving (and sending).

I have had my first successes talking to a graphing calculator with an MSP432, at last. Here's the state of the project:
- I can talk to a TI-83 Plus properly, in both directions.
- I cannot talk to either of two TI-84+CSEs, in either direction.
- When I disconnect the calculator, the MSP432 thinks it's still getting acks for bits, which means the lines are floating around, even though I am writing HIGH to pins set to INPUT (which is supposed to set the pull-up resistors on). I'll try adding explicit pull-up resistors to 3.3VCC and see if that helps. The problem disappears if I disconnect my leads as well, which supports the hypothesis that it's about floating/noise.

Edit: For those not familiar with ArTICL, the following is a demo of the TypeLetter example program. It repeatedly issues the silent linking command to type an M.

*bump* Apparently the MSP432 Launchpad and I have different ideas about what pull-up resistors are. When I added external pullup resistors to 3.3V, I stopped getting spurious sends to a phantom calculator (ie, a link cable not plugged into anything). However, the Energia code uses the exact same commands as on the Arduino to set the pull-up resistors (ie, set the state of an input pin to HIGH), so I'm not sure why those pull-up resistors aren't coming on. Two 10K resistors is not a terrible addition to a BOM, though, so I can survive. The remaining problem is that I still can't talk to a CSE, and I don't know if it's a timing issue or a voltage issue with the 3.3V GPIOs.

R_Pull: 20k min, 35k typ, 50k max. They're pretty weak. You also don't mention the PxREN bits, which must be set to enable pull resistors (PxDIR = input, PxREN = 1, PxOUT=1 for pull-up).

Let's see what their Wiring-like library does... from scanning msp430/wiring_digital.c, their pinMode() command takes OUTPUT, INPUT, or INPUT_PULLUP, which touches REN, DIR, and OUT, but their digitalWrite() doesn't seem to account for HIGH = pullup. Confusingly, there seems to be no corresponding MSP432 code.

Please note that I've brought this project into a bit of a released state, so I likely won't be adding further to this topic unless I start a significant new calculator + MSP432 Launchpad project not related to ArTICL. You can find the topic about the MSP432 Launchpad ArTICL library here:
https://dev.cemetech.net/forum/viewtopic.php?t=11863