The zip file includes a Fritzing (http://www.fritzing.org/) picture on how to connect the chip. It should work also for other chips of the same family (X9Cxxx, xxx=102,103,104,503 according to datasheet).
Had a quick look at the code and it looks nice , still some remarks.
why not mapping the UP and DOWN as follows?
#define DIGIPOT_UP HIGH (or 1) #define DIGIPOT_DOWN LOW (or 0)
would make change() a bit simpler
direction can be of type uint8_t, big enough to hold LOW/HIGH value.
pinnumbers in the constructor can be uint8_t (aka byte)
unsigned prevents negative numbers
and the range is big enough I guess
==> also the private vars to hold the pinnrs can be uint8_t
void increase(int amount);
as amount is a signed int it can accept negative numbers. Note that the for loop inside change() needs a positive amount.
Solution: Use an uint16_t for amount !
what is the max value of amount? you might consider using the constrain function ...
idem for decrease() and change() signatures..
Idea:
It might be interesting to hold the cumulative value of all changes since start in a signed long. MIght be indicative for its value.
Can the value be read from the digipot?
Thanks a lot, Rob, for your comments. It's good to see that there's feedback available for newbies.
I uploaded an updated version of the library with datatype, direction handling and constrain() changes you suggested.
The ability to read back the changes I left for future development. I've also been thinking adding some reset() call, to reset the pot into a known state (like zero), and after that the library could hold the actual state of the potentiometer, and also set it to an absolute value. But, later maybe
v.4, updated on May 26, 2011: modified to be compatible with Arduino 1.0; added set() to set absolute potentiometer values.
typo on the website - hint it is the year
You might add a link in the header of the .cpp /.h file to this thread to discuss the functionality.
Did some refactoring to optimize the code as some checks were made double - just give it a look if you like it
also caught one bug in set() ==> if the _currentValue is unknown you may not do math with it I assume
void DigiPot::reset()
{
change(DIGIPOT_DOWN, DIGIPOT_MAX_AMOUNT); //_currentValue should in fact be enough...?
}
void DigiPot::set(uint8_t value)
{
value = constrain(value, 0, DIGIPOT_MAX_AMOUNT);
if (_currentValue == DIGIPOT_UNKNOWN)
change(DIGIPOT_DOWN, DIGIPOT_MAX_AMOUNT);
else if (_currentValue > value)
change(DIGIPOT_DOWN, _currentValue-value);
else if (_currentValue < value)
change(DIGIPOT_UP, value - _currentValue);
}
uint8_t DigiPot::get()
{
return _currentValue;
}
void DigiPot::increase(uint8_t amount)
{
change(DIGIPOT_UP, amount);
}
void DigiPot::decrease(uint8_t amount)
{
change(DIGIPOT_DOWN, amount);
}
Change() can be optimized by not allways doing amount steps automatically but check if _currentValue == 0 or max too. But it might make the code more complex.
