Unfortunately, I couldn't figure out how to override KiCAD's layer colors for the 3D preview (nor did the internet help), but here's the "final" version I sent to OSHPark shortly before Christmas:



I also ordered a batch of parts from Digikey:
Code:

*bump* The PCBs for my shield version of the TI-DCC board finally arrived earlier this week:


I found minutes here and there for assembly, and here's the current version (with bonus photos, including one where I'm testing the overtemperature LED). I encountered a few interesting issues: I was rusty on soldering tiny SMD discretes, I initially put the ICSP header in upside-down (that is, the normal way up), and I discovered that the pins from the track screw terminal block touch the metal case of the Arduino's USB port, which I plan to solve with an insulator of some sort. I also need to get myself another LMD18200T; the one you can see in the photos elsewhere in this thread had its pins bent in an odd orientation, which meant predictable metal fatigue set in when I tried to bend them back. More updates soon!

Click for full resolution.


Got the last of my parts in today so here it is assembled!

Test by weeks end probably! Nights end


Train on idle, Click for full size.

I made a small mistake and forgot to make a logic ground for the control line coming out of the calculator. Not the end of the world just means I need to remember to hook it up. Though on a real setup the calculator would probably be wired to the 5v rail no the power connector. and there for could easily use the ground pin which would solve the problem.

Some minor changes made and pushed to https://github.com/Cemetech/TI-DCC/ for my smaller board. I forgot some a ground pin as I said above. Fixed that and exposed the copper on the text, And added some silk screening for pinouts...because I kept forgetting Makes it looks nicer :3


Click For full resolution

The assembled version looks great, and the fixed version looks great as well! Did the assembled one pass all of the tests? I see that the headlights on the train are on, but I assume it happily went forward and back as well? I think I shall nag you about DC mode, too, so you can test your new trains and I can test my ACS-64.

It functions exactly as expected All tests passed with flying colors on my setup

DC mode is coming up soon on the robin I feel. I need to get Iambian to whip me up a new interface for it as well. I feel like this controller interface is not meant to manage multiple trains at once. So having a DC only mode makes sense. So Upon hitting the DC button We are going to have it Display DC on the Address box. Do something with the buttons and set it to a step mode. Step count is to be determined still. Have to see what I can realistically squeeze out of the I/O port in terms of PWM resolution.

An update on geekboy's/Iambian's end: DC mode hasn't happened yet, but the interface is pretty much ready for the interrupt. Geekboy and I were tossing thoughts on properly implementing fine-grained PWM back and forth. We both agreed that it should require that separate interrupt to avoid overcomplicating the DCC interrupt, especially since you can't run DC and DCC mode locomotives simultaneously anyway.

A week or so ago, I did some really unfortunate soldering work to temporarily salvage my genuine TI LMD18200T that I got before they stopped sampling that chip, and was able to test it successfully working with my N-scale P42DC.



This weekend, I visited geekboy (and geekgirl), and we did much train debugging. Among the things we explored and figured out:

• Thanks to geekboy getting a number of (relatively) inexpensive "LM18200T"s from China, my shield now is complete! On the next batch, I want to use a slightly bigger drill hit for the LMD18200T's pins.
  
• The address of the new Atlas train that I got at the Amherst Model Railway show this year. We were able to eventually read and program its CVs using my shield, once we finally made sure we weren't trying to drive the 5V line from both my shield (5.13V) and USB (4.30V, because really long USB extension cable).
  
• We (possibly?) found a small issue with the 5V supply in geekboy's board that we want to resolve.
  
• We discovered that the TI-DCC ASM program appears to be producing an unclean PWM signal, more like 75% duty cycle than 50%.

Geekboy, what did I forget here?

A few things that would be good to mention:

1. We brought this project to Maker Faire again this year, in a much more reliable and polished form than last year, and it was a big hit. Brian Benchoff of Hackaday published information about it as part of a larger article about our hardware hacking work with calculators.
   
2. geekboy1011 continues to improve the software, and to my knowledge has been designing the DC mode.
   
3. I finally (over a year late) placed a split AliExpress/eBay order for more LMD18200Ts, those red mushroom emergency stop buttons, more screw terminals, and a couple of LCDs (for the TI-87 project).