+
indicates connected pino
indicates unconnected pin=
is used to list other things the pin is connected to-
s inserted between some of the pin functions for readabilityOC**
pins enclosed in parenthesis had lines over them in the pinout
GND +---.....---+ Vcc
SS PB0 + + PF0 ADC0
SCLK PB1 + + PF1 ADC1
MOSI PB2 + + PF4 ADC4
MISO PB3 + + + + PF5 ADC5
RTS OC1C OC0A --------- PB7 + PE6 AREF + PF6 ADC6
OC0B INT0 SCL PD0 + AIN0 + PF7 ADC7
INT1 SDA PD1 + INT6 + PB6 ADC13 OC1B OC4B
RXD1 ----------- INT2 --- PD2 + + PB5 ADC12 OC1A (OC4B)
TXD1 ----------- INT3 --- PD3 + + PB4 ADC11
OC3A (OC4A) -------- PC6 + + PD7 ADC10 T0 -- OC4D
ICP3 ----- OC4A --------- PC7 +-+-+-+-+-+-+ PD6 ADC9 T1 - (OC4D) onboardLED
CTS XCK1 PD5 --/ | | | \-- PD4 ADC8 ------------ ICP1
Vcc ------------------/ | \-------------- RST
GND --------------------/
power_negative GND +---.....---+ Vcc power_positive
column_7 PB0 + + PF0 row_5
column_8 PB1 + + PF1 row_4
column_9 PB2 + + PF4 row_3
column_A PB3 + o o + PF5 row_2
(OC1C) LED_3 PB7 + PE6 AREF + PF6 row_1
(SCL) I2C PD0 + + PF7 row_0
(SDA) I2C PD1 + + PB6 LED_2 (OC1B)
column_B PD2 + + PB5 LED_1 (OC1A)
column_C PD3 + + PB4 = Vcc
column_D PC6 + o PD7
PC7 o-o-o-o-o-o-+ PD6 onboardLED = GND
PD5 --/ | | | \-- PD4
Vcc ----/ | \---- RST
GND-------/
- notes:
- Row and column assignments are to matrix positions, which may or may
or may not correspond to the physical position of the key: e.g. the key
where
row_4
andcolumn_2
cross will be scanned into the matrix at[4][2]
, wherever it happens to be located on the keyboard. Mapping from one to the other (which only matters for defining layouts) is handled elsewhere. - LEDs are labeled using numbers (starting with '1') instead of letters (starting with 'A') as on the PCB.
- SCL and SDA: Need external pull-up resistors. Sometimes the Teensy internal pull-ups are enough (see datasheet section 20.5.1), but i think for this project we'll want external ones. The general recommendation for 400kHz I²C seems to be 2.2kΩ.
- Row and column assignments are to matrix positions, which may or may
or may not correspond to the physical position of the key: e.g. the key
where
DDRxn function PORTxn function
1 output 1 drive high
0 drive low
0 input 1 internal pull-up on
0 internal pull-up off
PINxn action function
write 1 toggles the value of PORTxn
read returns the logical value (1|0) of the pin
- notes:
- Unused pins should be set as input, with internal pullup enabled in order to give them a defined level (see datasheet section 10.2.6).
- PD6 (the onboard LED) already has a defined level (low), so there's no reason to set internal pull-up enabled on it. If we do, it will source current to the LED, which is fine, but unnecessary.
- Initially, we want either columns or rows (see <../options.h>) set as
hi-Z without pull-ups, and the other set of pins set as input with
pull-ups. During the update function, we'll cycle through setting the
first set low and checking each pin in the second set.
- To set a pin hi-Z on this board, set it as input with pull-up disabled.
- Switching the driving pins (the first set of pins) between hi-Z and drive low (treating them as if they were open drain) seems just as good as, and a little safer than, driving them high when they're not active.
- We need to delay for at least 1 μs between changing the column pins and
reading the row pins. I would assume this is to allow the pins time to
stabalize.
- Thanks to [hasu] (http://geekhack.org/member.php?3412-hasu) for the suggestion [here] (http://geekhack.org/showthread.php?22780-Interest-Check-Custom-split-ergo-keyboard&p=606415&viewfull=1#post606415), and [PrinsValium] (http://geekhack.org/member.php?6408-PrinsValium) for noting that his firmware had erratic behavior without the delays [here] (http://geekhack.org/showthread.php?22780-Interest-Check-Custom-split-ergo-keyboard&p=606426&viewfull=1#post606426). DOX tried it and confirmed that it worked for his protoype PCB (as of 3 June 2012) [here] (http://geekhack.org/showthread.php?22780-Interest-Check-Custom-split-ergo-keyboard&p=606865&viewfull=1#post606865). Before adding a delay we were having [strange problems with ghosting] (http://geekhack.org/showthread.php?22780-Interest-Check-Custom-split-ergo-keyboard&p=605857&viewfull=1#post605857).
-
notes: settings:
- PWM pins should be set as outputs.
- we want Waveform Generation Mode 5
(fast PWM, 8-bit)
(see table 14-5)- set
TCCRB[4,3],TCCRA[1,0]
to0,1,0,1
- set
- we want "Compare Output Mode, Fast PWM" to be
0b10
"Clear OCnA/OCnB/OCnC on compare match, set OCnA/OCnB/OCnC at TOP"
(see table 14-3)
this way higher values ofOCR1(A|B|C)
will mean longer 'on' times for the LEDs (provided they're hooked up to GND; other way around if they're hooked up to Vcc)- when in a fast PWM mode, set
TCCR1A[7,6,5,4,3,2]
to1,0,1,0,1,0
- when in a fast PWM mode, set
- we want "Clock Select Bit Description" to be
0b001
"clkI/O/1 (No prescaling)"
(see table 14-6)- set
TCCR1B[2,1,0]
to0,0,1
- LEDs will be at minimum brightness until OCR1(A|B|C) are changed (since the default value of all the bits in those registers is 0)
- set
-
notes: behavior:
- The pins source current when on, and sink current when off. They aren't set to high impediance for either.
- In Fast PWM mode setting
OCR1(A|B|C)
to0
does not make the output onOC1(A|B|C)
constant low; just close. Per the datasheet, this isn't true for every PWM mode.
-
abbreviations:
- OCR = Output Compare Register
- TCCR = Timer/Counter Control Register
Copyright © 2012 Ben Blazak [email protected]
Released under The MIT License (MIT) (see "license.md")
Project located at https://github.com/benblazak/ergodox-firmware