PIC 16F1788 hardware i2c issue (with oled ssd1306) SOLVED

superfranz · Joined: 05 Sep 2014 Posts: 10 Location: Trieste, Italy

The code posted works fine.
But only with soft i2c. if i force hardware doesn't work.
why? I checked the pins and are right, maybe something wrong in the device (correct with device editor)?

temtronic · Posted: Mon Sep 08, 2014 6:58 am

Could be a compiler bug, so please post your compiler's version....

..it doesn't work....

hmm does it work at all ?
what 'doesn't work' ??

jay

Ttelmah · Joined: 11 Mar 2010 Posts: 19512

What are you using as pull up resistors?.

The logic high level required by the hardware I2C, is 0.7*the supply. The software I2C, will only require about 2v. For a 3.3v bus at 400KHz, assuming it is reasonably short, you need about 1K2R. The typical value used for a 5v bus is _not_ low enough. Also using software the bus is unlikely to actually be getting to 400KHz (even with a 32MHz CPU).

You could specify 'SMBUS' in the #use I2C, which will then use 2.1v buffers, but I'd get the pull up working properly.

superfranz · Joined: 05 Sep 2014 Posts: 10 Location: Trieste, Italy

I use 2,2k pullups.
I power the display and the pic with lipo 3,7v battery directly. (3 to 4.2 V?).
The pic in datasheet operating 1.8 to 5.5v. the display's seller say from 2.5 to 5v.

I thought this was ok...

..the speed problem is the reason to use hardware (and little less rom)
_________________
SuperFranz

superfranz · Joined: 05 Sep 2014 Posts: 10 Location: Trieste, Italy

for me "doesn't work" means that with force_sw the display do what i want, with forcce_hw it remains black.

sorry for my english, I swear I have more words in Italian to express... Smile

i'll try with smaller resistor

thanks so much
_________________
SuperFranz

Ttelmah · Joined: 11 Mar 2010 Posts: 19512

There is a problem here.

The Vdd specified for the display, is 3.3v _maximum_. If the board you have can operate to 5v, then it must have a voltage regulator dropping the voltage to the display chip. The output 'high' from a pin on the chip (pulled up by an I2C resistor), cannot be more than about 0.4v above the core logic voltage of the display. As such, probably clamped below the voltage that the PIC I2C input requires. Drop your pull-up resistors to 1.5K (this is safe for your entire supply range - your current value is a little large), and select the SMBUS option on the PIC.

This then has a chance, but you really need some more data about the display module. Does it have a regulator?. What is the minimum voltage this can use?. The display logic can operate down to 1.65v, but the lower this is, the lower the I2C output voltages will be....

superfranz · Joined: 05 Sep 2014 Posts: 10 Location: Trieste, Italy

There is only a generic description, isn't an original "adafruit". from 3 to 5 Vdc.
I think should have a regulator and some type of clamps on i2c line.
I power both pic and display with same rail.

new tests:
5v vdd and 2,2k pullups: nothing
4v vdd and 1,5k pullups: nothing
3.3 vss and 1,5k pullups: nothing

with force_sw: works with all pullups and all vdds.

Tested with my friend's arduino, works at 5 and 3.3v supplies.

I haven't a probe to check the i2c lines, but with voltmeter i can see a sort of carrier only when "force_sw" both scl and sda.

I think something wrong on compiler or device configuration.
Compiler version is 5.008
Now i'll study well the pic datasheet and check the registers configurations.
_________________
SuperFranz

superfranz · Joined: 05 Sep 2014 Posts: 10 Location: Trieste, Italy

ok, ok.

SOLVED!

there is an error in registers configuration at 0x11D and 0x11C APFCON1 and APFCON2.

overwritten just only 0x11d with zero's ("default" pin configuration) in the start of main loop and all is ok, even with 2,2k pullups!

i think the error is diffused to other registers, i have to check, but is a lot of work!
and, how to correct them? if I edit the register with editor, how can I save?
_________________
SuperFranz