populated board with power and programming cable

Back during the summer 2012 I had some ideas for a short movie clip/mixed performance…thing and used it as an excuse for developing a board that could interface with DMX512 controlled lighting equipment, or perhaps if the excuse was the other way around—it is not very easy to tell, eggs and chickens and all that…
Nevertheless, among the things I wanted were
  • Computer control as well as the possibility of stand-alone operation.
  • Usability as a general I/O card for other projects.
  • Long DMX512 cables in a birds nest together with mains cables made me put some thought in to making sure there is good isolation between the DMX512 I/O and the rest of the card, should anything happen.
  • External power supply (not USB) given that I [a] were unsure about the power requirements, [b] The (imagined) possibility of the USB bus going into power down mode during operation, and [c] aforementioned stand-alone operation.
  • Single power connection even though I wanted isolation.
My familarity with the Atmel avr platform led me to using an atmega164 clocked at 20MHz for the on-board brains. The particular device was selected for its double USARTs, of which I connected one to an FT232RL for computer connectivity, and the other was led through a Si8641 isolator and to a MAX491E handling the conversion between logic and RS422/485 signals.
A Recom 5V DC/DC converter was chosen to make sure there is power on both sides of the isolation gap.

board layout, ground plane not shown, but shaped like the 5V plane
While a lot of things went wrong during the performance (probably due to fragility in the computer interface code and a distinct lack of time to debug the thing before going live), the board has seen a lot of use since then, and I have learned a few things as well.
  • Do not use surface mount electrolytic caps if you intend to hand solder the board, while so and tqfp format ic's as well as passives are easily solderable by a steady hand, pliers and some solder wick—this is not the case for smt electrolytes, their legs/pads are on the underside of the component, and since they are about as big as through hole variants, the smt ones offer no benefit if hand soldering.
  • A two pole screw terminal that is not of the thick-legged high amperage kind needs to be held in place by more than its two connection pins.
  • Given the above, a two pole terminal mounted beside a six pole terminal is pretty stupid, design for, and get an eight pole terminal right away. And even better are the kind that consist of two parts, a pin-header with some polarization, and a matching female header with screw terminals.
When I had my own server up and running I kept an archive with the schematic (gEda gschem format) and board layout (gEda pcb format, as well as gerbers) available, now that I don't, I dunno, you could contact me if you think you'd benefit from them and I'll see if I can dig them up. As for firmware code, that has available on github since more or less the beginning.

No comments:

Post a Comment