PLA filament + heat = trouble

My card reader had been working pretty well. Then I put it in the back seat of my car to take it to a demo, but before the demo it sat there on a hot day for a couple of hours. As pet owners know, the interior of a car can easily get to 120F or more, and it did. The result was serious deformation of some PLA-printed parts.

Right angles and straight edges under the least bit of tension became terribly distorted, and of course the reader stopped working.

This distortion of PLA-printed parts when exposed to heat is apparently well known to the 3D printing aficionados I sobbed to. who recommend at least switching from PLA to PET-G. Even better would be ASA, but that is more difficult to print. I took the baby step and reprinted using PET-G. At the same time I made several other changes to improve reliability:

• The original design had supports for the control rod assembly bolted to the bottom support plate. That is not sufficiently rigid, particularly for tall parts, and also provides the opportunity for slight misalignment. The new design uses a rigid printed cage for the control rod assembly.
• The alignment pointers on the prism were not entirely inserted into the alignment channel before the no-hole rods contact the cards, which means that the alignment could be slightly off and a rod might hit the circumference of a hole. The new design added 0.2" of travel to make sure the alignment was complete before the card was touched.
• The control rod lock shape wasn't sufficiently wide at the mouth to force alignment for rods that aren't completely pushed in, perhaps because of distortion in the card. The new locks are wider.

• The cards were sometimes hanging up on the alignment pins. So I smoothed the pins -- I otherwise do almost no post-print finishing -- and coated both the pins and the alignment holes of the cards with a teflon lubricant. That helps a lot.
• I made a better alignment jig for sewing the cards, by 3D-printing sections of 6 that can be lapped and glued together. That makes it easier to sew long strings of cards.

• I changed the sewing regimen to knot the upper and lower strings in between the cards. With the strings just interlaced sewing-machine style, it was too easy for the cards to slide along the strings, changing the inter-card distance too much.

Here's a video of it running, again with the cards coded for testing with 2 6-bit characters each. When used in the real Engine they will be coded with instructions and memory addresses instead.