TheBackShed.com

Forum Index : Microcontroller and PC projects : RP2040 program for Keypad with more than 4x4 keys.

Page 2 of 3

Author

Message

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 04:30am 17 Jan 2026

Copy link to clipboard

Print this post

There appears to be some interest in a MMBasic driven custom keyboard so here is an expanded version of the Keypad program.
It uses one or more 74HC595 serial-to-parallel chips to drive the rows, freeing up Pico pins for a bigger input Port for more columns.
12 columns x 8 rows would give you 96 keys.
Still not enough? Add another 74HC595 and go for 16 x 16 = 256 keys!

The next firmware release looks like it will have a two pin version of Device Bitstream (perhaps only on the RP2350?). If so the HC595 Sub can be simplified.

' KeyPad_RP2040_74HC595_v03.bas - Program for more than 4x4 Key Pad
' Rows scanned by a 74HC595 serial-to-parallel chip
' rows = number of rows. For more than 8 rows daisychain another 74HC595 chip.
' cols = number of columns
' RowPinGP = row GP pin number for the 74HC595
' ColPinGP = first of the column GP pin numbers
' Key = number of the key that has been pressed.
' eg. 0 to 63 for 8 columns x 8 rows. 1st row 0 to 7, last row 55 to 63.
' X x Y Keypad with buttons bridging column x to row y

Option base 0
Dim integer n, Key=-1, RowPinGP=0, Rows=8, ColPinGP=1, Cols=8
' Dim KeyMap$(rows * cols - 1)=("Q","W","E","R","T","Y", etc) 'map keys to characters with KeyMap$(Key) if req.

Dim integer RowPinNo = MM.Info(pinno "GP"+Str$(RowPinGP))

SetPin RowPinNo, DOUT 'set the output pin to the 74HC595 and set its outputs to 0
HC595 RowPinNo, Rows+7, 0 'Call Sub with DataPin, max. number of Rows, Word to send, 0=all low
'Rows+7 to ensure any unused pins are also set to 0

For n=ColPinGP To ColPinGP+Cols-1
SetPin MM.Info(pinno "GP"+Str$(n)), INTL, Key.Pad, PULLUP 'set the input Port pins and ISR
Next

Do
' Your main processing loop
If Key+1 Then Print "Key";Key, "&H";Hex$(Key,2), Timer-t;"mS" : key=-1
'to retain the value of Key set a flag in the sub and reset it here
Loop

Sub Key.Pad
t=Timer
Local integer y, x, x1, n
Pause 10 'contact bounce delay, set to minimum value that is reliable

For y = 0 To rows 'scan the rows
HC595 RowPinNo, Rows, 1<<y Xor (1<<Rows+1)-1 'Call Sub with DataPin, no. of Rows, Row to read, inverted
Execute "x1 = (1<<cols)-1 XOR Port(GP"+Str$(ColPinGP)+",cols)" 'read the columns

If x1 Then x=Log(x1)/Log(2)+1

If x>cols Or y>rows Then 'check for values out of range and abort this read
x=0 :y=0 : Key=-1
HC595 RowPinNo, Rows+7, 0 'prepare for next key input, 0=all low
Exit Sub
EndIf
Key = y * cols + x - 1 'get key no. from row no. and col. no., starting from 0
If x1 Then Exit For
Next

HC595 RowPinNo, Rows+1, (1 << Rows+1)-1 'all high, block further input to prevent double press
Pause 150 'set to minimum value that is reliable, or remove it and/or the HC595 Sub call above
HC595 RowPinNo, Rows+7, 0 'prepare for next key input, 0=all low
'Set a flag here if main prog. needs it
End Sub

Sub HC595 OutPin As integer, R As integer, Dat As integer 'Data output Pin, number of Rows, Word to output
Pin(OutPin)=1 'ensure correct polarity for Bitstream
Local integer n, el=R*2-1, B(el)
For n=0 To el Step 2 'prepare Bitstream array
B(n)=3-(Dat>>(el-n)\2 And 1)*2 'low data pulses, 3uS=0 ,1uS=1
B(n+1)=1 '1uS high spaces
Next
Device BITSTREAM OutPin, R*2, B() 'send it
Pin(OutPin)=1 : Pause 0.04 ' >20uS high = latch clock
End Sub
End
' 74HC595 connections
' 3V3 to pins 16, 10
' Gnd to pins 8, 13
' BitStream input to pin 11 (clock)
' pin 11 - 1k8 - pin 14 (data in) - 1n5 - Gnd
' pin 11 - A-diode-K - pin 14 (ser. data in) eg 1N914 or 1N4148
' pin 11 - 4k7 - pin 12 (latch) - 1n5 - Gnd
' pin 11 - K-diode-A - pin 14 (latch)
' pin 9 (ser. data out) to next 74HC595 (if used) pin 14 (ser. data in),
' parallel pin 11 to pin 11 and pin 12 to pin 12
' Row output pins:- 15, 1, 2, 3, 4, 5, 6, 7

Edited 2026-01-17 15:46 by phil99

Mixtel90

Guru

Joined: 05/10/2019
Location: United Kingdom
Posts: 8868

Posted: 08:11am 17 Jan 2026

Copy link to clipboard

Print this post

pack of 3 74HC595 is £3.90 + £3.50 delivery from The Pi Hut

1x RP2040-Zero is £1.30 from AliExpress (bundle deal for 3x is £3.90 with free shipping).

They now have a RP2350-Zero for £2.90. :) That would make a nice I2C+IRQ X-Y keyboard scanner. 72 keys on 3 GPIO pins. :)

Mick

Zilog Inside! nascom.info for Nascom & Gemini
Preliminary MMBasic docs & my PCB designs

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 09:14am 17 Jan 2026

Copy link to clipboard

Print this post

Quote pack of 3 74HC595 is £3.90 + £3.50 delivery from The Pi Hut

Some years ago I wanted 4 to drive a display. Too expensive in the local shops but just a couple of $ for a tube of 25 from AE, postage included.
Now have to find excuses to use them. ;-)

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 02:50am 18 Feb 2026

Copy link to clipboard

Print this post

Using the MMBasic Keypad command requires 8 pins for 4 x 4 keys.
A 74HC595 and a diode/resistor matrix can read the keypad using just 2 MCU pins.
A DOUT scans the keys via Bitstream and a Interrupt Sub reads the pressed key.
As with the previous version the rows are driven by idle low outputs.
Inverted logic is used to read the columns via diode-resistor logic. Switching the columns is done by more idle low outputs.
For more than 16 keys more 74HC595 chips can be daisy-chained. 2 give up to 64 keys and 3 up to 144.

Testing was done on a PicoMite MMBasic RP2040 V6.02.00 @ 200MHz and 378MHz.
Time taken to read a key (@ 378MHz) and return to the main loop is about 5mS for Key 0, varying up to 10mS for Key 15.
It varies because it stops scanning when it finds the key.

' KeyPad_RP2040_74HC595_2_Pin_v01.bas - Program 4x4 Key Pad
'
' Keypad reader using a 74HC595 and 2 MCU I/O pins.
' For more than 8 rows plus columns daisy-chain another 74HC595 chip or two.
' then adjust the Rows and Cols limits for the For / Next loops.
' 1 chip is up to 16 Keys, 2 is up to 64 (8 x 8), 3 is up to 144 (12 x 12)
' keypad does not need to be square. 2 chips = 16 outputs so could do 12 x 4 (12 + 4 = 16)
'
Dim Integer Rows = 4, Cols = 4, Key = -1, OutPin=MM.Info(pinno GP0), InPin=MM.Info(pinno GP1)
SetPin OutPin,dout : Pin(OutPin)=1 'initialize keypad row output
HC595 OutPin,Rows+Cols,0 'set all pins low
SetPin InPin,intl,KP_int,pullup 'initialize keypad column input

Sub KP_Int
Pause 2 'key de-bounce time, adjust to suit your keypad
Local integer R, C

For R=0 To Rows-1 'Scan the rows (low) one at a time, the rest high
HC595 OutPin,Rows+Cols,15-(1<<R)
If 1-Pin(InPin) Then Exit For 'exit when the active row is found
Next

For C=0 To Cols-1 'Scan the columns (low) one at a time, the rest high
HC595 1,Rows+Cols, (15-(1<<C))<<4
If 1-Pin(InPin) Then Exit For 'exit when the active column is found
Next

If Rows-R Then Key = R*Rows+C 'Evaluate the key unless the row count is too high
'Set a flag here if the main loop needs it.
HC595 OutPin,Rows+Cols,0 'Prepare for next press, set all pins low
End Sub

Sub HC595 Out.Pin As integer, RC As integer, Dat As integer
'Data output Pin, number of Rows+Cols, Word to output
Pin(Out.Pin)=1 'ensure correct polarity for Bitstream
Local integer n, el=RC*2-1, B(el)
Array Set 1,B() '1uS low data and 1uS high spaces = all 1s
For n=0 To el Step 2 'prepare Bitstream array
B(n)=3-(Dat>>(el-n)\2 And 1)*2 'low data pulses, 0 = 3uS, 1 = 1uS
Next
Device BITSTREAM Out.Pin, el+1, B() 'send it
Pin(Out.Pin)=1 : Pause 0.025 ' >20uS high = latch clock
End Sub

Do
If Key+1 Then Print "Key";Key : Key = -1
'If you need to preserve the value of "Key" replace "Key+1" with a flag
' that was set in the ISR and reset the flag instead of "Key = -1"
Loop

End
' 74HC595 connections
' 3V3 to pins 16, 10
' Gnd to pins 8, 13
' BitStream input to pin 11 (clock)
' pin 11 - 1k8 - pin 14 (data in) - 1n5 - Gnd
' pin 11 - A-diode-K - pin 14 (ser. data in) eg 1N914 or 1N4148
' pin 11 - 4k7 - pin 12 (latch) - 1n5 - Gnd
' pin 11 - K-diode-A - pin 14 (latch)
' pin 9 (ser. data out) to next 74HC595 (if used) pin 14 (ser. data in),
' parallel pin 11 to pin 11 and pin 12 to pin 12
' 4x4 Row output pins:- 15, 1, 2, 3,
' 4x4 Column output pins:- 4, 5, 6, 7

And the circuit diagram. The Bitstream pin goes to the input and the output goes to the INTL pin.

Keypress response times:-

Key 0 5.003000006 mS
Key 1 4.600000001 mS
Key 2 6.101999998 mS
Key 3 6.671000004 mS
Key 4 4.600999996 mS
Key 5 6.107000001 mS
Key 6 5.691 mS
Key 7 6.275000006 mS
Key 8 5.148000002 mS
Key 9 6.662 mS
Key 10 7.221000001 mS
Key 11 7.800999999 mS
Key 12 6.663000003 mS
Key 13 7.230999999 mS
Key 14 6.831 mS
Key 15 7.395999998 mS

Edited 2026-02-18 16:49 by phil99

Footnote added 2026-02-26 12:34 by phil99
A version for the MicroMite Mk II.
The changes account for the lower CPU speed, lower logic threshold voltage and lack of Bitstream command. Tested at 40 and 48MHz. If there are too many button errors replace the 4 output diodes with Schottky types as the forward voltage of regular diodes isn't much less than the logic threshold voltage. Added to that is the voltage drop on the 1k5 resistors, leaving little headroom.

' Keypad reader using a 74HC595 and 2 MCU I/O pins for MicroMite Mk2
' For MicroMite Mk II v5.05.05, also PicoMite
' Can be expanded to larger keypads with more 74HC595 chips.
' 1 chip is up to 16 Keys, 2 is up to 64, 3 is up to 144
Dim Integer Rows = 4, Cols = 4, Key = -1, OutPin = 4, InPin = 5
Dim map$(15)=("1","2","3","A","4","5","6","B","7","8","9","C","*","0","#","D")
SetPin OutPin,dout : Pin(OutPin)=1 'initialize keypad rows
HC595 OutPin,Rows+Cols,0 'set all outputs to 0
SetPin InPin,intl,KP_int,pullup 'initialize keypad columns

Sub KP_Int
SetPin InPin,off :SetPin InPin,Din,pullup 'disable inerrupt while reading
t=Timer
Pause 2 'key de-bounce time, adjust to suit your keypad
Local integer R, C

For R=0 To Rows-1 'find the active row
HC595 OutPin,Rows+Cols,15-(1<<R)
If Pin(InPin)=0 Then Exit For
Next

For C=0 To Cols-1 'find the active column on the active row
HC595 OutPin,Rows+Cols, 255-(1<<C+4)-(1<<R)
If Pin(InPin)=0 Then Exit For
Next

If R<Rows Then Key = R*4+C
HC595 OutPin,Rows+Cols,0 'Prepare for next press, set all pins low
SetPin InPin,intl,KP_int,pullup 'restore interrupt
End Sub

' CLS 0
Do
If Key>-1 And Key<Rows*Cols Then
Print "Key";Key, map$(key), Timer-t;" mS"
Key = -1
EndIf
Loop

Sub HC595 O.Pin As integer, N.out As integer, Dat As integer
' Data output Pin, number of Rows+Cols, Word to output
Pin(O.Pin)=1 'ensure correct polarity for pulses
N.out = N.out - 1 'array B() uses Base 0
Local integer n
Local float B(N.out)
For n=0 To N.out 'prepare data timing array
B(n)=(11-(Dat>>(N.out-n) And 1)*6)/1000 'low data pulses, 20uS=0 ,10uS=1
Next
For n=0 To N.out :Pulse O.Pin,B(n) :Next 'Send the data (a substitute for Bitstream)
Pin(O.Pin)=1 ': Pause 0.02 ' >20uS high = latch clock 'time to get back to main loop is more than this.
End Sub

End

It should also work on the MicroMite Plus and other 'Mites but not tested.

Footnote added 2026-02-26 21:34 by phil99
The intention was to have the code cater for more than 4 x 4 keys but one line in the column search loop limits it to 8 outputs. Here is the corrected version, 255 (2^8-1) is replaced by 2^(Rows+Cols)-1.

For C=0 To Cols-1 'find the active column on the active row
HC595 OutPin,Rows+Cols, 2^(Rows+Cols)-1-(1<<C+4)-(1<<R)
If Pin(InPin)=0 Then Exit For
Next

And here is an example of the output.

Key 0 1 11 mS
Key 1 2 15 mS
Key 2 3 18 mS
Key 4 4 15 mS
Key 5 5 18 mS
Key 6 6 22 mS
Key 8 7 18 mS
Key 9 8 21 mS
Key 10 9 25 mS
Key 13 0 25 mS
Key 12 * 22 mS
Key 14 # 28 mS
Key 3 A 22 mS
Key 7 B 26 mS
Key 11 C 28 mS
Key 15 D 31 mS

For the 160MHz MicroMite Plus the times could be reduced to something close to the PicoMite by reducing the data and latch time-constants and with corresponding adjustments to the pulse times in the HC595 Sub.

For n=0 To N.out 'prepare data timing array
B(n)=(11-(Dat>>(N.out-n) And 1)*6)/1000 'low data pulses, 11uS=0 ,5uS=1 (=11-6)
Next

Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5898

Posted: 09:16am 18 Feb 2026

Copy link to clipboard

Print this post

Hi Phil,

When you have only 2 pins you can also use a PCF8574, that can be had for few cents on the internet. The PCF8574 is perfect for this in that it has weak pullups, on bi-directional pins. A 4x4 keyboard requires only that chip, nothing else. Not resistors, caps, diodes. But it needs I2C bus. In this test the PCF8574 was connected to system I2C bus on pins GP14 and GP15.

This is possible code. Requires between 0.5ms and 1.5ms on a 200MHz RP2040. Dual key presses are reported as key numbers above 32.

'PCF8574 4x4 keyboard
'connect keyboard rows to pins 9,10,11,12, and columns to pin 4,5,6,7

pcf%=&h20 'I2C address when A0/A1/A2 are GND
pinrow%=&b00001111 'setup rows and columns, lsb in, msb out

'for decoding
Dim pin2num%(15)=(8,8,8,8,8,8,8,3,8,8,8,2,8,1,0,8)
Dim rowpat%(3)=(&b01111111,&b10111111,&b11011111,&b11101111)

'---------------------------- main program -----------------------

'init default for rows and columns
I2C2 write pcf%,0,1,pinrow%

'main loop
Do
I2C2 read pcf%,0,1,d% 'read keyboard
If d%<>pinrow% Then 'check if key pressed
Print whatkey%() 'show the key after scanning
End If

Pause 100
Loop

'------------------------------- subs -----------------------------

'scan keys
Function whatkey%()
Local i%,d%
For i% = 0 To 3 'scan rows
I2C2 write pcf%,0,1,rowpat%(i%) 'push pattern
I2C2 read pcf%,0,1,d% 'read response
If d%<>rowpat%(i%) Then 'different then key pressed
whatkey%=i%+4*pin2num%(d% And 15)'decode
I2C2 write pcf%,0,1,pinrow% 'restore default rows
Exit Function 'and quit
End If
Next
End Function

Tested on V6.02.01b3 on RP2040

In case you have 3 pins available you can also connect the INTn pin on the PCF8574 and do not need to poll at all.

For 64 keys (8x8) use 2x PCF8574 on different I2C addresses. Scan time will double. Routing/coding is simplest if you have one PCF8574 do columns, the other Rows.

Volhout
Edited 2026-02-18 19:46 by Volhout

PicomiteVGA PETSCII ROBOTS

PhenixRising
Guru

Joined: 07/11/2023
Location: United Kingdom
Posts: 1935

Posted: 11:02am 18 Feb 2026

Copy link to clipboard

Print this post

Quote

4 x 4 keypad reader using a single I/O pin, four 7.5K resistors, four 30K resistors, and a 1000pF capacitor

Start(pin) : I/O pin connected to the keypad circuit
Poll : Returns the key pressed in ASCII when key is pressed initially
Returns false if no key is pressed or if a key is held down
Check_Keypad : Returns key even if held down

┌────────────┐ The keypad that I tested with has an 8-pin header:
│ 1 2 3 A │─┳──────┐ 1 - row 2 (4,5,6,B)
│ │ 30K │ 2 - row 3 (7,8,9,C)
│ 4 5 6 B │─┫  3 - column 1 (1,4,7,*)
│ │ 30K 4 - row 4 (*,0,#,D)
│ 7 8 9 C │─┫ 5 - column 2 (2,5,8,0)
│ │ 30K 6 - column 3 (3,6,9,#)
│ * 0 # D │─┫ 7 - column 4 (A,B,C,D)
└────────────┘ 30K 8 - row 1 (1,2,3,A)
    │
┣┻┻┻────┘
 7.5K
┣─────────────────── I/O pin
1000pF

Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5898

Posted: 11:22am 18 Feb 2026

Copy link to clipboard

Print this post

Yeah, I did that one on an Arduino once.

Volhout

PicomiteVGA PETSCII ROBOTS

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 08:44pm 18 Feb 2026

Copy link to clipboard

Print this post

The fixed width font of a code box untangles the mess.
That looks interesting so will have a go at reading it with MMBasic.
It looks like you set the pin as an output to charge the capacitor then switch the pin to input and measure how long it takes to discharge.

┌────────────┐ The keypad that I tested with has an 8-pin header:
│ 1 2 3 A │─┳──────┐ 1 - row 2 (4,5,6,B)
│ │ 30K │ 2 - row 3 (7,8,9,C)
│ 4 5 6 B │─┫ ⏚ 3 - column 1 (1,4,7,*)
│ │ 30K 4 - row 4 (*,0,#,D)
│ 7 8 9 C │─┫ 5 - column 2 (2,5,8,0)
│ │ 30K 6 - column 3 (3,6,9,#)
│ * 0 # D │─┫ 7 - column 4 (A,B,C,D)
└────────────┘ 30K 8 - row 1 (1,2,3,A)
    │
┣┻┻┻────┘
 7.5K
┣─────────────────── I/O pin
1000pF

As I mentioned to Mick I have a heap of 74HC595s left over from a project years ago so they are effectively free. They must be good for something! :)
Edited 2026-02-19 07:05 by phil99

Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5898

Posted: 06:11am 19 Feb 2026

Copy link to clipboard

Print this post

Hi Phil,

I used it opposite. Discharge cap, let resistors charge it again.
Don’t use a 2350 A2.
Mmbasic may be slow to defect some keus. 1000pf/7.5k

Volhout
Edited 2026-02-19 16:12 by Volhout

PicomiteVGA PETSCII ROBOTS

dddns
Guru

Joined: 20/09/2024
Location: Germany
Posts: 835

Posted: 08:40am 19 Feb 2026

Copy link to clipboard

Print this post

Many thanks for sharing all this useful code and different methods..appreciate it!

phil99 said
It looks like you set the pin as an output to charge the capacitor then switch the pin to input and measure how long it takes to discharge.

Another idea is to use an analog input. It might be faster not to wait for the discharge.

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 08:41am 19 Feb 2026

Copy link to clipboard

Print this post

Quote Discharge cap, let resistors charge it again.

I will give that a try too, when I get back to it.
As the end of the resistor chain (near the top) points downward I assumed it was going to ground, but it could equally go to Vcc.

Quote Another idea is to use an analog input.

At first I thought that was how it was meant to work then saw the capacitor, though I guess is could be a noise filter.
.
Edited 2026-02-19 18:44 by phil99

PhenixRising
Guru

Joined: 07/11/2023
Location: United Kingdom
Posts: 1935

Posted: 10:01am 19 Feb 2026

Copy link to clipboard

Print this post

Take a look at this 3 X 5 decoder

bfwolf
Senior Member

Joined: 03/01/2025
Location: Germany
Posts: 246

Posted: 12:35pm 20 Feb 2026

Copy link to clipboard

Print this post

PhenixRising said Take a look at this 3 X 5 decoder

Ingenious circuit!

How does it work? A square wave oscillator whose frequency is controlled by the keys?

I looked at the example program for the PIC microcontroller, but I didn't really understand it.

Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5898

Posted: 12:43pm 20 Feb 2026

Copy link to clipboard

Print this post

All,

Frequency changes depending what key pressed. Measure frequency in the arduino/mite.

All these single wire keyboards have the same problem. In practise you can make it work for 4, maybe 9 keys.

Frequency changes with temperature (logic gates threshold and capacitor drift), so you could get wrong key's identified.

More reliable is to feed the resistor divider voltages to an ADC. Resistors typically do not drift much, and ADC's are developed to be accurate. Avoid capacitors and non controlled logic gate thresholds.

Another problem with all POLLED systems is that you need to actively check continuously, not to miss a key. The PCF8574 with it's interrupt free's you from that task. But yes... it uses 3 wires (I2C + interrupt).

Volhout
Edited 2026-02-20 22:45 by Volhout

PicomiteVGA PETSCII ROBOTS

bfwolf
Senior Member

Joined: 03/01/2025
Location: Germany
Posts: 246

Posted: 02:02pm 20 Feb 2026

Copy link to clipboard

Print this post

I assume the oscillator also oscillates when no key is pressed? This "fundamental frequency" could be used for regular calibration. This would significantly reduce temperature effects. And if the frequency evaluation is done via a timer input on the microcontroller, and IRQs are used, "software interrupts" could also be triggered when a frequency change is detected.

This circuit has another major advantage: only two wires between the keyboard and the microcontroller – and depending on the design, a considerable distance can be bridged. Another advantage: it's difficult to hack – for example, if used to create a combination lock.

bfwolf

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 11:15am 24 Apr 2026

Copy link to clipboard

Print this post

Had a think about the one pin keypad circuit and came up with this interrupt triggered analogue input version.

The 2 33nF capacitors pull the input low enough to trigger the interrupt when any key is pressed.
The transistor, pot, diode and 100k form a crude constant current source for the resistive divider chain.
This produces linear voltage steps for the keys.
The reference voltage is the difference between 5V (VSYS) and 3.3V (Vdd) so ideally requires them to be stable and low noise.
The diode provides some temperature compensation for the transistor and ensures the output voltage can't go above Vdd.
On the Pico this is just adequate for 16 steps with only occasional errors.

For larger keypads a better constant current source will be needed. Also the resistors will need to be precision types.
The value of the row resistors should be:-

R(row) = R(column) x N columns.

For this 4 x 4 version 3.3K is close enough to 820 x 4 but bigger keypads will need to be more accurate.
The current source should be adjusted to produce about 3V with the maximum resistance in circuit (bottom right key).
Also more 33nF capacitors will need to be connected to the rows to ensure triggering the interrupt.
Every third row should be ok. The columns should be ok with just one.

5V________/\/\/\/\___ ____________________________________________AIN GP28
10k pot | | _____________________________
| |_________| |
B / E | \ | 1 2 3 A |
3V3____|\|_______|/ PNP | / | |
|/| | |\ PN200 | 3k3 \ | |
1N4148 | \ C | / | |
| | | |__| 4 5 6 B |
| | | \ | |
| | | 3k3 / | |
| | | \ | |
| | | /__| 7 8 9 C |
| | | \ | |
| | | 3k3 / | |
/ | | \ | |
\ |_________|__| |_/__| * 0 # D |
100k / | | |____________________________|
\ 33nF | | | |
| |_/\/\/\|_/\/\/\|_/\/\/\|
| | 820R 820R 820R |
| | _|_
| | 33nF ___
0V___________|_____________________________|_______________________|

'Interrupt triggered analogue keypad driver using 1 pin.
Dim integer key1=-1, key2=-1, n, Rows=4, Cols=4, Key.pin=MM.Info(pinno GP28)
Dim Float Vcal=3.09 'calibration voltage - while highest key pressed
Dim Kmap$(15)=("1","2","3","A","4","5","6","B","7","8","9","C","*","0","#","D")

SetPin Key.pin, intl, Key.int 'set key interrupt

Sub Key.int
SetPin Key.pin, ain
Local key = (Rows * Cols), KeyV
If Pin(Key.pin) < 3.2 Then
Pause 50 'de-bounce and capacitor charging time
If Pin(Key.pin) < 3.2 Then KeyV = Pin(Key.pin)
If KeyV >= 3.2 Or KeyV < 0.0001 Then
SetPin Key.pin, intl, Key.int 'restore interrupt
Exit Sub
EndIf
key1 = Cint(KeyV * (Rows * Cols -1) / Vcal)
Print Format$(KeyV,"% 5.3g"); " V ", 'Comment this out after adjusting voltage
' for highest key (3.0 to 3.1V) then setting the value of Vcal (eg Dim Float Vcal=3.09)
EndIf

Pause 50
SetPin Key.pin, intl, Key.int 'restore interrupt
End Sub

Do
If key2 <> key1 And Key1 < Rows * Cols Then
Print " Key no."; key1, "Key code "; Kmap$(key1)
Pause 150 'adjust repeat rate
key1 = -1 : key2 = key1
EndIf
Loop

End
' https://www.thebackshed.com/forum/ViewTopic.php?PID=253854#253854#253854

> RUN
0.0176 V Key no. 0 Key code 1
0.237 V Key no. 1 Key code 2
0.426 V Key no. 2 Key code 3
0.647 V Key no. 3 Key code A
0.863 V Key no. 4 Key code 4
1.08 V Key no. 5 Key code 5
1.27 V Key no. 6 Key code 6
1.48 V Key no. 7 Key code B
1.68 V Key no. 8 Key code 7
1.89 V Key no. 9 Key code 8
2.1 V Key no. 10 Key code 9
2.28 V Key no. 11 Key code C
2.47 V Key no. 12 Key code *
2.71 V Key no. 13 Key code 0
2.87 V Key no. 14 Key code #
3.07 V Key no. 15 Key code D

Edited 2026-04-24 22:22 by phil99

robert.rozee
Guru

Joined: 31/12/2012
Location: New Zealand
Posts: 2526

Posted: 03:07pm 24 Apr 2026

Copy link to clipboard

Print this post

phil99 said Had a think about the one pin keypad circuit and came up with this interrupt triggered analogue input version [...]

there were similar ideas discussed way back in 2014:
https://www.thebackshed.com/forum/ViewTopic.php?PID=66729#66729#66729

cheers,
rob :-)
Edited 2026-04-25 01:08 by robert.rozee

phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3257

Posted: 04:30am 25 Apr 2026

Copy link to clipboard

Print this post

Rob, that is an excellent solution. Especially the self calibration when no key is pressed.

Spent some time trying to graft that on to mine but as the extra resistor bridges the rows and columns it defeats the interrupt key detection for most of the keys.
Working on an alternative method.

robert.rozee
Guru

Joined: 31/12/2012
Location: New Zealand
Posts: 2526

Posted: 05:04am 25 Apr 2026

Copy link to clipboard

Print this post

phil99 said Rob, that is an excellent solution. Especially the self calibration when no key is pressed.

Spent some time trying to graft that on to mine but as the extra resistor bridges the rows and columns it defeats the interrupt key detection for most of the keys.
Working on an alternative method.

you could possibly provide an interrupt signal using a comparator:

- connect one input pin of the comparator to the junction of the 470r and 180r resistors. this point should always sit at about 64mV (varying slightly with temperature, but this doesn't matter).

- connect a high-value resistive divider between the collector of the PNP transistor and ground. the sum of the divider resistors should be 500k or more. you are wanting to divide the voltage down to around 1/20th, so high value to collector of the PNP transistor, low value to ground.

- connect the junction of the comparator to the other input pin of the comparator.

- connect the output of the comparator to your interrupt pin.

with no button pressed, the voltage at the collector of the PNP transistor will be sitting at a little under 2 volts, so the voltage at the junction of the divider should be at about 100mV. when a key is pressed, this voltage should fall below 64mV. i think!

but you'll still need the timer event firing at some slow rate just to do the auto-calibration.

cheers,
rob :-)
Edited 2026-04-25 15:04 by robert.rozee

Footnote added 2026-04-25 22:49 by robert.rozee
actually, the "high-value resistive divider between the collector of the PNP transistor and ground" does not need to be excessively high-value, as the current it draws flows only through the PNP transistor, bypassing the LM334. so 15k at the top end and 820r at the bottom end would suffice, contributing just another 150uA of current draw to the whole micromite/picomite setup.

robert.rozee
Guru

Joined: 31/12/2012
Location: New Zealand
Posts: 2526

Posted: 05:14am 25 Apr 2026

Copy link to clipboard

Print this post

just had another thought - if you increase the value of the 2k7 resistor that is connected across the "f4" key, you can substantially increase the 'no key pressed' voltage without affecting anything else. with a high enough value, you should be able to push this voltage up to close to 3v3. BUT, if you do this be sure to add a base resistor between the PNP transistor and the LM334, as the LM334 will try to pull 10's of mA to push the voltage across the 470r resistor up to 64mV.

with careful selection of the base resistor, you may then be able to hook an interrupt pin directly to the V+ pin of the LM334, no comparator needed; with no key pressed the V+ pin will be LOW, whereas with any key pressed the V+ pin will sit at around 0.6v below 3v3.

cheers,
rob :-)
Edited 2026-04-25 15:19 by robert.rozee

Page 2 of 3

Print this page