09/29/2016 at 00:22 •
New version designed with KiCad. Yet one more connector orientation. I might have been distracted and I left a hazard for SAM D21 compatibility. So I will respin this.
This revision has a more suitable RTC w/ a 6.8mm backup battery retainer on the bottom. I swapped out the hinged SD socket for a side-entry push-pull.
I'll release the schematic, layout, BOM, etc. after I've verified everything.
07/28/2016 at 07:19 •
Short Stack is suiting my purposes well, but I've decided to rev the board to bring the cost down a bit (the lipoly charger and power path have much more configurability and control than needed). I'm also looking at using the SAM L21 which would cut runtime current in nearly half (theoretically). The L21 also adds a few interesting features (op amps). I'll also max out the flash as things get really tight if you want to, say, use both FAT on the microSD and another heavy stack like BLE (There's plenty of room if you use the SD raw without a file system). Additionally, I want to source different FETs (decrease footprint, lower on resistance, lower leakage) and inductors (better availability).
After I verify the next revision, I could release all of the design files and libraries if folks are interested. I'm also open to making boards to order.
07/19/2016 at 22:00 •
No updates on this project. Office flooded in April. Luckily my gear was spared, but I temporarily lost my work space.
02/09/2016 at 01:23 •
I finally got around to writing a driver for the LS013 LCD screen. I made an ASCII font table and rudimentary "terminal" API. I wanted something really small, so the font only has numerics, punctuation, and capital letters. Lower-case letters would nearly all look the same. ;) I have the draw code sideways with respect to its mounting position, but it was the quickest way to get something on the screen. The terminal draw code requires zero pixel memory on the MCU. Easy enough to modify which direction it "scans" from.
Sorry, a bit of a boring log in this picture:
Side view showing the side spacers (there's a bit of slack in the M2.5 holes in the acrylic, I should have aligned the pieces a bit better before snapping a pic)
...also, I forgot to make a plain acrylic pane with no "branding" on it, oops.
02/08/2016 at 08:14 •
A bit of a diversion, but I wanted to see what some laser-cut acrylic shields would look like (first time I've ever made anything with acrylic).
Small top and bottom guards on the main board + 16-chan servo controller configuration (there is enough space between the bottom acrylic and the main board to host a LiPoly):
Also, outer rings that can be stacked to make a solid case:
Here I have the small board-sized top and the rings together (they mount into a full pane on the bottom with appropriately spaced holes). Alternatively, one could just mount the boards on the standoffs without the board-sized top and use a full-sized pane on top (made with this batch, includes the engraving as well, but not shown)
You cannot see it well in this photo but there are also outer rings with cuts for a USB plug to fit in. They can also be oriented to provide access to the two user buttons, but was not done for this picture.
01/21/2016 at 18:23 •
After a bit of a break with the holidays, got back to putting some of the various boards together. Reflowed the simple PWM board and hand-soldered some right angle headers to it. Wrote up a quick servo demo (activated when pressing switch 2), seen below.
I admit to debugging this for _several_ hours (it's always the simplest of boards!). It would work fine and then stop ACK'ing it's I2C address. Eventually I found that I had a dry joint on one of the I2C address-select pulldown resistors that seemed fine under test because the pressure of the continuity probe closed the circuit. I have a big fat note about avoiding this issue from another project and yet I managed to fall victim to it again. Anyhow, servos! LEDs! Backlights! Gate drives!
12/12/2015 at 08:19 •
Real small update. While everything is functional with the current board, I still need to rev the board due to the poor land pattern I had for the JEDEC MO-203 (aka SOT-353-5 aka SC-88A aka SC-70-5), so I decided to do a little component placement clean up. While I was moving things about, I decided to change the battery connector from a PicoLock to a JST-PH. The PicoLock is nice as it keeps the overall profile very thin, but the availability of LiPoly batteries prefitted with JST-PH connectors (Adafruit!) makes having a JST much more convenient. With its relatively large lands it was rather hard to find a place for it, but after moving some things around, I managed to squeeze it in.
(bottom side shown (from top))
12/03/2015 at 18:23 •
Assembled the first of the QFN version w/ 0402. Did a visual inspection...all joints look good however one of the footprints for a JEDEC MO-203 part looks functional but a little suspect, so I'll be double checking that footprint. Will do a continuity check and power it up soon.
Reflow looks good. So far, been very pleased with OSHStencils...
The land pattern for the IC marked VJ is rather questionable... :/
11/30/2015 at 23:49 •
Quick fit test. Soldered up the LCD board and mounted the screen. Perfect fit on the flex to molex ZIF on the backside. Huzzah trigonometry!
11/22/2015 at 08:51 •
FCI makes a rectangular connector with dimensions that, when fitted male to female, fit the 6mm between boards perfectly. The Harwins were 1/2 a millimeter off which works but a gap is visible.
FCI on first (top) layer, Harwin on second...
Will be looking at the LC situation of lines running through the connectors soon. 100khz I2C was functional with 8.2k pullups (the top board has a bunch of environmental sensors), but I haven't hooked anything up to the scope yet.