And we’re up and running solely on battery! All connections have been made, no more spaghetti! I have two extra leads coming off of the battery that are going to a volt meter, monitoring the battery. There’s nothing else connected. The loose red/black wires at the top are meant to go to a DC barrel jack which I haven’t put on yet; it’s for charging the battery. And you can just barely see the USB connector above the two buttons. It’s on the underside of the PCB. Connecting the USB cable with the main power off will only recharge the battery. Same with DC-IN. Flipping the main switch will turn the unit on and if there’s a USB cable attached, cause it to enumerate on the computer to upload code to it.
A couple (ok several) lessons learned in this process:
-
Stencils, stencils, stencils: I didn’t remake the stencils for this test cycle since I didn’t change a whole lot. I figured I can manually put the paste on. For the most part this worked, but in retrospect it took much longer because of bridges I was causing with too much paste. Next time I will spend the additional $25 for a new set of stencils dammit!
-
Stick with the development parts: I found an SD card connector that is slightly cheaper and with identical footprint. Unfortunately, that came back to bite me big time. I had to rip it off after reflow because one of the pins (MISO) did not connect, and since it’s under the connector itself, there was no way to reheat just that portion. I ended up reheating the top and blowing hot air through the holes. That resulted in the inside melting faster than M&Ms in your mouth and the entire top shield getting discolored. Removing the old one took over 20 minutes where reflowing a new, correct one, took about 3 minutes.
-
Test each module separately: I was eager to pair the two modules and get it working. While the second power module worked right away (the first one didn’t have one of the ICs positioned properly and it didn’t reflow right), on the controller module itself I had several issues: one of the resistors on an array did not reflow properly causing the attached LED not to work; one of the series resistors on the USB data lines didn’t reflow right so it wasn’t enumerating; there were two bridges on the AVR that needed to be cleaned up - lucky for me they were IO pins; the whole snafu with the SD card connector; and lastly the trigger button was defective and needed to be replaced. In the end I ended up separating the two modules twice just to fix those problems.
Other issues that are now on the list to be fixed in the next test cycle:
Better connections: I need to revise how I connect the LED bars to the control/power modules. Soldering 6 wires in a very confined space, even with a thin tip soldering iron, is a pain in the arse! I’m thinking perhaps use regular wire connectors or something else because … damn! I need to keep it very tight though, so it can’t be some large connector either.
Kill the power: If the unit’s running and the main power switch thrown, because of how the LEDs are wired directly to the battery with Sziklai pairs, they never actually get turned off, the controller does! The Sziklai pairs are controlled by the MCU. That’s an oversight on my part. I think I’m going to add a power-off delay so when the switch is thrown, the MCU sends a power-off command first before completely powering down. For now just stopping the unit prior to power off does the trick.
There was something else but I forgot what it was. I’m sure it’ll come to me eventually. At this point, this unit is up and running. Tomorrow I will spend some time getting it inside of the tube, getting the holes cut for the switch and see how that all comes together. I’m guessing I will have more changes to be made based on how that process goes.
In the 30 minutes it took me to write this with the unit cycling through some 65 different images, battery power went from 4.07V down to 3.93V … aaaahhhhh. I’m happy! This means I can run a 5 minute show several time in one night. YES!
