Skip to content

Latest commit

 

History

History
69 lines (51 loc) · 4.85 KB

Capacitive.md

File metadata and controls

69 lines (51 loc) · 4.85 KB
Raw

Capacitive stuff, IBM or otherwise

Only switch-specific things are here. Most of the doc is in Common.md!

Planning: pinouts

  • It is STRONGLY RECOMMENDED to set matrix size to your physical matrix size.
  • There are some ready-made configurations for the default pinout in config.h.
  • YOU MUST USE CONTIGUOUS ROWS/COLUMNS, STARTING FROM ZERO.
  • Pins on the board are marked 0.0, on envelope - P0_0, and in PSoC Creator - P0[0]. All 3 refer to the same pin. 2.3-2.7 means "2.3 to 2.7, including".
  • Effort is made to make pins contiguous, but that's not always possible, so pay attention.

Default pinout: Project "Capsense"

default pinout WARNING to use column 24, R5 (between 0.1 and 12.6) MUST be removed.

If you don't - column will always read zeroes. Which is, for a NORMALLY_HIGH switch like Beamspring, means they will appear PRESSED ALL THE TIME.

  • D0: 0.2. Must be connected to the nearest GND.
  • Guard: 0.3. Must be connected to the nearest GND. Separate wire is HIGHLY RECOMMENDED
  • ADC Vref: 0.4. Must be connected to the nearest VDD.
  • Rows: 2.0, 2.3-2.7, 12.7, 12.6
  • Cols: 1.0-1.7, 3.0, 3.1, 3.3-3.7, 15.0, 15.1, 15.5, 0.0, 0.1, 0.5-0.7, 15.4.

If you're changing the pinout

  • You'll need 35 pins. That's not much more than the kit has - so choose the layout wisely.
  • rows are top(Rows0[0]) to bottom(Rows0[7]), columns are left(Cols[0]) to right(Cols[23]).
  • Keep in mind that in FlightController everything is 1-based so normies can use it, so rows will be 1 to 8.
  • P12(all pins of it) cannot be used for analog connections. That means you can only assign Rows to those pins.
  • It is better to separate Rows and Cols by a pin electrically connected to the ground, but not necessary. If you use adjacent pins - one matrix cell will have higher readings, that's all.

Additional Hardware configuration hints

  • If key monitor only shows zeroes, no matter which keys you press - set ADC resolution higher.
  • If even at 12 bits it still doesn't work - increase charge delay. Recommended setting is 18.
  • If several keys fire at once - set longer discharge delay. I found that 180 (10 microseconds) works pretty well, but you can go even higher for increased EMI rejection.

Configuring thresholds

NOTE for beamspring, invert direction, so max->min and adjustments are negative.

!!!NOTE!!! If you see double presses on some keys, set threshold higher - just under the steady pressed state. Some keys have physical bounce and there are two peaks. Controller is too fast and sees this as 2 keypresses.

Short version ():

  • Click "Key Monitor" button.
  • Click "Start!". Get the idea of levels that should be there - press keys, observe readings going up and down. Small numbers below 7-segment indicators are min/avg/max.
  • Click "Stop!". Select "Max" into dropdown near the "reset" button, click "Reset", "Start!".
  • Wait 15 seconds or longer, while readings stabilize.
  • click "Stop!". Click "Set thresholds". Close window.
  • Click "Thresholds" in the main window. put a small positive value (see below) into adjuster spinbox, click "Adjust".
  • Click "Apply".
  • Close threshold editor. Select "Config -> Upload" in menu. Test. Once you're satisfied with results, "Command -> Commit".

Thresholds should be set ~2x higher than most of the matrix settles on. For beamspring - probably 75% of the highest reading. TEST SETTINGS BEFORE COMMITTING. If you get thresholds wrong - there will be a red "UNSAFE" light in the status bar and keyboard will refuse to produce output. It's supposed to retest settings after you upload. If it doesn't - click "Scan", it will try to restart scanning.

Debugging the matrix

If something is wrong - you can still use matrix monitor using the following trick:

  • open matrix monitor,
  • make sure the mode is "Now"
  • click "Start!"
  • go to main window
  • click "Scan"

The numbers under the "LCD" numbers are min/avg/max since last reset. Looking at those will tell you A LOT about what controller sees - but not everything, because in normal mode controller processes about 30 thousand rows per second and there is no way to transmit that amount of data over USB 2.0.

WARNING exit this mode in the reverse order (or by pulling out the USB plug), otherwise you'll likely have to reboot.

This is mostly useful for beamsprings - because all keys are down in beamspring by default. Model F is much easier - everything is up and you can just press a key to observe the effect. So, recommended debugging strategy - pull out the PCB, lay it down over grounded conductive surface, start matrix monitor, put one flipper (or a small coin) over the PCB and observe the changes. If several rows/columns change at the same time - you have a short between those, find it. If readings are normal but assembled keyboard is unstable - it's likely a ground problem.