PCB shrink
Opened this issue · 1 comments
Hi,
just had some time to spend and though why not give something back to the great video?
So I took the layout and tried to shrink it. Applied my design rules for JLC and was able to reduce the PCB size from 28x7.6 mm² to 22x6.3 mm²:
I have to admit that I didn't check for design rules especially for the RF-circuit, so this might not be the best possible approach.
Some points/thoughts for further improvement:
- The connector left doesn't comply with my JLC design rules, might be ok for production but I think it would be better to fix it - or replace it with test points as this might save another 1.5 mm in length.
- To save some vias, switching to bit-bang SPI and rearranging the IOs could save some width of the PCB, but I guess we're talking about fractions of mm here
- I didn't check the balance of the board
- To reduce the total width further, the magnets of the stirbar could be considered into the PCB outline
- Could be worth to look into using PCB traces as a coil (4 layer PCB maybe?) or wire-wound around the PCB like done in commercial tags, though the lack of ferrite surely has impact on the inductance
- A wild thought (having no bigger clue about magnets, ferrites and inductors): with a wire-wound or PCB antenna, could the magnets be used as "kinda substitute" of the ferrite material?
Since I didn't bother to make a pull request, please find the eagle files attached:
The coil could be done on pcb, in 2 layers, but low frequency usually means long distance to get the right inductance, so its best to stick to a coil or other antenna, if a stick on wirewound commercial coil was used, then that could be an option, but that is more difficult than the small inductor
On the software side, I'm less familiar with msp, but if SBWTDIO could use a pin pullup, then R1 may no longer be required
I use primarily Kicad so had another crack at shrinking it, down to 6.75 x 13.9, So I did trade back some width, for a massive drop in length,
Due to the clearance requirements of the antenna, this should be about as close as it can be cut, I even looked up its datasheet to trim the pads slighlyt
due to more traces routing over the top of the antenna, this does leave a slight imbalance in mass, but as its the short axis, I'm not that worried as its moment is tiny,
I stopped before implementing the 4 pads on the left, but at this point that side of the PCB has enough space for some generous solder pads in that area, I don't see much point to throughole pads for a 1 off programming interface, simply make them large enough for a pogo jig to hit reliably,
So here is me passing along the torch for the next person to refine further, this time in Kicad instead of eagle,
