i2c_write() is not allowed

mire · Joined: 05 Jan 2019 Posts: 5

Trying to get htis up for hte first time.

When I try to compile with i2c included I get
Function used but not defined: :: i2c_start() not supported when using HW I2C on this device, see i2c_transfer()

Setup is

Ttelmah · Joined: 11 Mar 2010 Posts: 19546

What chip?.

Vital data.

mire · Joined: 05 Jan 2019 Posts: 5

PIC18F45K42.

Can use the I2C HW ok from MPLab.

Found that using
#use i2c(Master,slow=100000)
and the code compiles and downloads.
But it does nothing on the port (see with oscilloscope).
_________________
: In the

Ttelmah · Joined: 11 Mar 2010 Posts: 19546

OK. Your chip does not have the MSSP. Instead it has a dedicated I2C
module. This supports multi-byte buffered I2C, and can automatically
generate a start when data is sent. This is why CCS are suggesting
to use the I2C_transfer function. This can automatically do the entire
transfer.
I must admit I would have expected the older mode to still be supported
Will have a play when I get back to my development system.
Why not just use I2C_transfer?.

mire · Joined: 05 Jan 2019 Posts: 5

By all means have a play.
But if it won't work I'll figure it out.
Really was more interested in seeing CCS working.
Do you know if CCS supports a Dallas style 1 wire interface?

Regards.
_________________
: In the

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

temtronic · Posted: Sat Jan 05, 2019 1:59 pm

I don't know if CCS has a driver in their library but there are several postings of 'one wire' code in the code library and I've used DS18B20 temp sensors with PICs for years.
'drivers' can be quite a challenge. Often you need to make them 'generic', applicable to several PICs or 1-wire devices and not everyone needs every function of every device connected to every PIC.
In my case I 'cheated', only used enough code to read the sensor as I didn't need anything else. I also 'cheated' by using 1 I/O pin per device ( 4 in total). While that made the code larger, no need to read device IDs, etc. PICs have LOTS of codespace these days. Also by having 1 device per pin a failure in one wire doesn't kill the rest of them, though I have a way around that...

Jay

Ttelmah · Joined: 11 Mar 2010 Posts: 19546

There are several one wire device drivers in the CCS code.

DS18B20 (thermometer).
DHT11 (humidity).
DS2432 (EEPROM).

There may well be more. As Jay has said there are also drivers in the
code library.

Looking at the code being generated for I2C, the reason they are not supporting I2C start is that they are setting the peripheral up to support
the automatic generation of start, when data is loaded. So targeted at
using the I2C_transfer format. If you must use start separately it is only
a matter of defining a couple of bits to change the mode, and then just
setting the control bit for this manually.

mire · Joined: 05 Jan 2019 Posts: 5

Thanks all.
Understood.
However if I 'FORCE_SW' should I see a Bit Bash version?
Still no output

thanks
(How do you close a thread)
_________________
: In the

temtronic · Posted: Sun Jan 06, 2019 8:24 am

from my foggy memory..

FORCE_SW will create a bit bashed driver, for any 2 I/O pins, providing they
are bidirectional, however you will not get interrupt capability.

Jay

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

Ttelmah · Joined: 11 Mar 2010 Posts: 19546

Also have you turned off analog I/O. This has priority over digital.

mire · Joined: 05 Jan 2019 Posts: 5

Where do I find the code for the DHT11 say.
_________________
: In the

temtronic · Posted: Thu Jan 10, 2019 6:30 pm

CCS puts the drivers for devices like the DHT11 in the 'drivers' folder.