I have come across the following SPI slave code at this site:

#include <SPI.h>

char buf [100];
volatile byte pos;
volatile bool process_it;

void setup (void)
  Serial.begin (115200);   // debugging

  // turn on SPI in slave mode
  SPCR |= bit (SPE);

  // have to send on master in, *slave out*
  pinMode (MISO, OUTPUT);

  // get ready for an interrupt
  pos = 0;   // buffer empty
  process_it = false;

  // now turn on interrupts

}  // end of setup

// SPI interrupt routine
ISR (SPI_STC_vect)
byte c = SPDR;  // grab byte from SPI Data Register

  // add to buffer if room
  if (pos < sizeof buf)
    buf [pos++] = c;

    // example: newline means time to process buffer
    if (c == '\n')
      process_it = true;

    }  // end of room available
}  // end of interrupt routine SPI_STC_vect

// main loop - wait for flag set in interrupt routine
void loop (void)
  if (process_it)
    buf [pos] = 0;
    Serial.println (buf);
    pos = 0;
    process_it = false;
    }  // end of flag set

}  // end of loop

How to modify the above code so that it responds to input clock signal just like in this SSI protocol? Uno or Nano can be used.

enter image description here

Basically I want to let an arduino act as an absolute encoder so that it will transmit 13-bit binary data(it can be any random fixed binary data) as shown in the above SSI protocol. So the Arduino will output 13-bit binary data along with with start bit and tm taken into account as shown in the document above without parity.

The master's code I plan to use:

const int CLOCK_PIN = 2;
const int DATA_PIN = 3;
const int BIT_COUNT = 16;  
void setup() {
    pinMode(DATA_PIN, INPUT);
    pinMode(CLOCK_PIN, OUTPUT);
    digitalWrite(CLOCK_PIN, HIGH);
void loop() {
    float reading = readPosition();

//read the current angular position
float readPosition() {
 unsigned long graysample = shiftIn(DATA_PIN, CLOCK_PIN, BIT_COUNT);
 delayMicroseconds(100); // Clock must be high for 20 microseconds before a new sample can be taken
unsigned long binarysample = grayToBinary32(graysample);
 return ((binarysample * 360UL) / 65536.0); // ouptut value from 0 to 360 with two point percision
//read in a byte of data from the digital input of the board.
unsigned long shiftIn(const int data_pin, const int clock_pin, const int bit_count) {
 unsigned long data = 0;
 for (int i=0; i<bit_count; i++) {
 data <<= 1; // shift all read data left one bit.

 PORTD &= ~(1 << 5); // clock pin goes low
 PORTD |= (1 << 5); // lock pin goes high
 data |= digitalRead(data_pin); // cat the new read bit to the whole read data.
 return data;
unsigned int grayToBinary32(unsigned int num)
 num = num ^ (num >> 16);
 num = num ^ (num >> 8);
 num = num ^ (num >> 4);
 num = num ^ (num >> 2);
 num = num ^ (num >> 1);
 return num;
  • What prevents you from doing this with digitalWrite() and delayMicroseconds()? Oct 21, 2019 at 12:49
  • @EdgarBonet The SSI requires 100kHz minimum. Also Im not into the subject, I need this because I want to check a master library which can read incoming gray code from an encoder. But the encoder is too expensive so I wanted to mimic the slave
    – floppy380
    Oct 21, 2019 at 12:54
  • @EdgarBonet I would appreciate your answer but if dont have time or interest, I want to ask you if I do this with DigitalWrite() and delayMicroseconds(); I also need the slave to act with a clock as shown in SSI document. So it seems a bit more tricky than I was thinking.
    – floppy380
    Oct 21, 2019 at 13:09
  • That's a hard protocol to bitbang. You won't do it with SPI since it has things in there that SPI can't do. So forget SPI. You'd only do it with an interrupt for the clock and some very fast and tight code (which probably means assembly language). Personally I'd implement it in a CPLD rather than an MCU.
    – Majenko
    Oct 21, 2019 at 14:33
  • @Majenko How about Arduino as master and using SPI? I saw some using in some forums such as forum.arduino.cc/index.php?topic=156812.0 and jgelectronics.nl/art-net-dmx/arduino-project-4.html I can at least buy a cheap encoder as slave but I need the master to work at least
    – floppy380
    Oct 21, 2019 at 14:48

1 Answer 1


You may try to bit-bang the protocol using direct port access. This is less convenient, but way faster than digitalWrite(). Below is an example, untested program. I am assuming the pinout is

  • data = PB0 = digital 8
  • clock = PB1 = digital 9
  • pins PB2 – PB5 (10 – 13) are unused

Edit: This updated program uses the clock as an input, and the data as an output. The previous version used both lines as outputs, which was due to a misunderstanding of mine.

#include <avr/io.h>

int main(void)
    uint16_t data = 0x155;

    DDRB  |= _BV(PB0);
    PORTB |= _BV(PB0);
    TCCR1B = _BV(CS10);  // start timer 1 @ F_CPU
    loop_until_bit_is_clear(PINB, PB1);

    for (;;) {
        loop_until_bit_is_set(PINB, PB1);
        PORTB = data & 1;
        data >>= 1;
        TCNT1 = -304;        // 20 us * 16 MHz - 16 cycles
        TIFR1 |= _BV(TOV1);  // clear overflow flag
        while (bit_is_set(PINB, PB1)) {
            if (bit_is_set(TIFR1, TOV1)) {  // timer overflowed
                PORTB = 1;  // data high
                data = 0x155;
                loop_until_bit_is_clear(PINB, PB1);

This should be able to operate at frequencies up to about 200 – 250 kHz. There is a delay of about 0.5 µs between the rising edges of the clock and the data been sent out, with a jitter of 3 clock cycles.

The data word is sent least significant bit first, contrary to what the protocol diagram shows. However, if you send arbitrary data, this should not be an issue.

Note that this is a pure avr-libc program, which does not use the Arduino core. It uses polling to check the inputs and the timer. The interrupts are never enabled.

  • Thank you very much! I will try this with the master code I have. One arduino will be slave and the other will be master. Your code simulates the slave if Im not wrong but theres one thing Im a bit confused in your code. In this protocol slave supposed to receive clock signal. But in your code when I check digital 9 pin output by a scope, it outputs a very similar bit stream as digital 8. I thought slave would only receive clock signals as inputs from the master yet in your case the slave seems outputting bit stream. Am I getting something wrong here?
    – floppy380
    Oct 22, 2019 at 9:49
  • Re “In this protocol slave supposed to receive clock signal”: Oh, I see, I wasn't aware of this. Oct 22, 2019 at 10:03
  • Please see that at the end of my question I now added the master code I found for a 16 bit encoder. The code is written for master and converts the gray code to binary. (Usually abs encoders output gray code for better efficenicny.) But for the slave side supposed the receive the clock in a way that each clock it outputs one bit. More info about this protocol here: posital.com/media/posital_media/documents/… My aim is to try this master code with a simulated slave. Slave can send any angle data. Better be constant data to check out
    – floppy380
    Oct 22, 2019 at 10:19
  • This works great! I just need to add in master code after return //digitalWrite(clock_pin,LOW); PORTD &= ~(1 << 5); // clock pin goes low delayMicroseconds(1); //digitalWrite(clock_pin,HIGH); PORTD |= (1 << 5); // lock pin goes high delayMicroseconds(1); to exactly mimic the required pattern. I checked it works great now thank you,
    – floppy380
    Oct 22, 2019 at 14:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.