Need help for i2c communication with FDC2214

[ressas]

Hello to everyone. I am trying to communicate the PIC microcontroller with i2c with FDC2214 integrated. I need to set bit R / W = 1 to read. So I need to increase the address of the address by one: 0x2a + 1 = 0x2B.
However, the address 0x2B belongs to another register, so I can not communicate and constantly comes to 0XFF information. So I can't read, and probably not write. Any idea?
I have adapted the CCS C code from the Arduino code in this link:
https://github.com/zharijs/FDC2214

[temtronic]

comments / questions...

1st. Get rid of use fast_Io and set-tris lines of code. let the compiler handle that ! Too easy to wrongly configure I/O !
2nd what is the VDD voltage of the PIC ?
3rd with all I2C devices, you should copy/compile/run PCM P I2C scanner program located in the 'code library' forum. It'll confirm the PIC CAN communicate with the I2C device AND it's address !
4th 0x2A is the SEVEN bit address, which is 0x54 in 8 bit addressing. PCM P's pgm wil report 0x54. however the ardunio code requires 7 bit which always confuses me....
5th post a 'link' to the FDC22xxx device so we can get the datasheet.
6th if you're using a 'module' not just the 'chip', post a 'link' to that as well.

OK.....
this..

[Ttelmah]

and:

[ressas]

[ressas]

[temtronic]

re: 2. You cannot connect a 5 volt PIC directly to a 3 volt peripheral.You need some kind of 'logic level conversion', or easier, just run the PIC at 3 volts.That PIC( my favourite) will run at 3 volts AND at high speed( not all will..)

re: 3 . In the 'code lbrary' there is a program called 'I2C Scanner'. It will show you the addesses of all I2C perhipherals connected to a PIC ! It should be run to confirm the PIC IS communicating with the I2C device. It will also show you the address. Ardunio's tend to use 7 bit device addressing, whereas CCS uses 8 bit device addresing.

re: 4 'Ardunio ' 7 bit address has to be left shifted 1 bit position to become CCS 8 bit addresses, so 0b00101010 becomes 0b01010100. 0x2A becomes 0x54. What's interesting is that all device registers contain 16 bit ( 2 byte) values and the default 'device address' is 0x3054. the LSB is 54 same as what I calculated.



I prefer to think in 8 bit addressing, the CCS standard, as most( all ?) all data transfers are 'byte' wide for I2C peripherals

[ressas]

[temtronic]

Use 0x54 to write TO the device, and 0x55 to read FROM the device. CCS uses 8 bit combined 7 bit device address and a 1 bit R/*W .

Consider renaming some variables.
Something like 'dev_adrs' for the base address of the I2C device, then perhaps 'reg_adrs' for the 'registers' inside of the device. On pages 22-23 there's a 'register map', which shows what they are for.

These register addresses are only 8 bits( a byte) to taht variable should be an unsigned int8 not a 16 bit.

this line...
unsigned int16 read16FDC(unsigned int16 addresss) {
looks like one too many 's' in addresss.

Also it should be unsigned int8 ( value from 0x00 to 0xFF )


this...
data = i2c_read();
data<<=8;
data |= i2c_read();

could be recoded as

datahigh=i2c_read(); //get high byte of dataword
datalow=i2cr_ead(); //get low byte of dataword
data=make16(datahigh,datalow); //make a 16bit word from 2 bytes

While your code is probably correct, mine is easier to understand and might be faster or more efficent.

You should add //comments at the end of most lines of code. They 'cost' nothing but allow you and others to understand what you're trying to do.

Also I don't like the underscore for a variable name. It's hard for me to see. Maybe it's a proper C convention, but I was never taught C. So far I've spent 1/2 my 66 years cutting C code for PICs, so anything that makes it easier for me I do !

I'm sure others will help if they can

Jay

[PCM programmer]

You need to make the corrections shown in bold below. These items
are shown in the following diagram in the FDC2214 data sheet on pg. 22:
Figure 17. I2C Read Register Sequence
Data sheet:
http://www.ti.com/lit/gpn/fdc2214

[ressas]

Thank you both for your help. The program started running (successfully?). The only problem is that when I run the Arduino code in the Github link, the value I get for 4 channels is as follows;

[temtronic]

Please post your complete program. There are a LOT of 'setup' or 'configuration' bits that need to be seen.

Also,according to teh datasheet, a 12 bit result is stored as 16 bit and it must be rightshifted 4 bits BEFORE you use it.

It could be your print statement, hard to tell without seeing all your code.
So two possible answers as to why the numbers are different.

At least you're making PROGRESS !!!

Jay

[Ttelmah]

Two things:

1) Don't use terms like 'unsigned long'. The problem is thes change with
different processors. To define an Arduino 'unsigned long' on the PIC, you need
to use 'unsigned int32'. To make the variable explicitly 32bit.
Generally it is always better when writing code to use 'explicit' sizes,
not terms that may have different interpretation on different chips.
2) Defining a variable as 'signed' or 'unsigned', does not control how it is
printed. If you send an unsigned int32 to another program, or to 'printf',
it prints this how you tell it. So if using printf for example, it needs to use
%LU to print such a variable as an unsigned value. Sounds to me as if your
output is incorrectly using %LD (which will assume the value to be 'signed').

[ressas]

This code was created with the contributions of "temtronic" and "PCM programmer"

[PCM programmer]

Did you read Ttelmah's post ? You are reading an unsigned int32,
but you're putting it into an unsigned int16 array. See below in bold.

[ressas]

thank you soo much,
data improved ELHAMDÜLİLLAH