Quick Microchip I/O Extender question

[GDSever]

I've been playing around in with a 28x1 and the MCP23008 and MCP23016 chips (in VSM for now), and have been successful getting the I2C bus to read inputs and activate the _INT_ pin when they change... but have been wholly unsuccessful getting the outputs (writes) to work.

Are the outputs on the MCP23016 like a darlington driver IC where it sinks current from the 5V rail, or like an output from a PICAXE where the load is tied between the IC output and the ground rail?

I've tried it both ways, and neither worked - so it is entirely possible I'm missing something in my code too.

Thanks in advance!

 

[westaust55]

Suggest that you post your programs so we can see what you are doing.

Presume that you are setting up the direciton registers for outputs.

From the datasheet:

The IODIR register controls the direction of the pins
(input or output). More specifically, the IODIR registers
simply enables/disables the output driver. When the
driver is activated (IODIR = 0), the pad is driven to the
state in the latch register (OLAT). When deactivated
(IODIR = 1), the driver is high impedance.

I have not used these chips but from the datasheets, but looking at the MCP23008:

IODIR is register $0 controlling direction of each of the 8 bits.
To set bit 0 as an output, you need to write $01 to register $0.

then write to GPIO (register $09) which internally transfers the data to OLAT (register $10)
so writing $01 to Register $09 should bring pin 0 high.



Also have you had a look at this thread by BCJKiwi.
Has a code sample for the MCP230016

http://www.picaxeforum.co.uk/showthread.php?t=8098

 

[GDSever]

Ack. Sounds like there might be more register config I need to do. I have read the datasheet more than once, but I'm obviously missing something... Right now I am just trying to read several inputs of GP1 and then set corresponding LEDs on GP0 - a very simple I/O test to make sure I understand the I2C commands.

I had looked at the code in that other thread, however BCJKiwi is using the HI2C commands which I am not - I potentially have a DS1307 on the same I2C bus, so I was trying to stick with the lower speed I2C commands. I'll look thru them again and see if I can pull out an extra register config command I missed along the way...

Here is the code I am working with now, which will change if I figure out what I'm doing wrong....

' MCP23016 registers
Symbol GP0 = $00
Symbol GP1 = $01
Symbol OLAT0 = $02
Symbol OLAT1 = $03
Symbol IPOL0 = $04
Symbol IPOL1 = $05
Symbol IODIR0 = $06
Symbol IODIR1 = $07
Symbol INTCAP0 = $08
Symbol INTCAP1 = $09
Symbol IOCON0 = $0a
Symbol IOCON1 = $0b

' I2C Pins and MCP23016 interrupt trigger
Symbol SDA = pin4
Symbol SCL = pin3
Symbol MCPInt = pin5
Symbol MCPAddr = %01000010

setint %00000000,%00100000  ' Interrupt when input pin 5 goes low

pause 1000        ' wait for MCP23016 to boot

GoSub ConfigureMCP23016

Main:

        sertxd (#b0)
        pause 100
        '
        GoTo Main
        

Interrupt:

    setint %00000000,%00000000  ' Disable our interrupt while we process things
        
    ' Select the MCP23016 chip on the I2C bus
    i2cslave MCPAddr, i2cslow, i2cbyte
    
    ' Read our states from GP1
        i2cread GP1, (b0)
        pause 10
        
        ' Output the state to the first bus
        i2cwrite OLAT0, (b0)
        pause 10
        
        ' Re-enable our interrupt.
        setint %00000000,%00100000  ' Interrupt when input pin 5 goes low

        Return

ConfigureMCP23016:

     ' Select the MCP23016 chip on the I2C bus
    i2cslave MCPAddr, i2cslow, i2cbyte
    
    ' Set GP0 to all writes / outputs ($00), GP1 to all reads / inputs   ($FF)
    i2cwrite IOCON0, (%00000000, %11111111)
    pause 10
    
    ' set a "off" state on everything
    i2cwrite OLAT0, (%00000000, %00000000)
    pause 10
    
    ' Set the interrupt bit to occur when any of the GP1 states change
    i2cwrite INTCAP1, (%11111111)
    pause 10
    
    i2cread GP1, (b0)
    pause 10
        
    Return

 

[GDSever]

well, one of my problems is definitely using IOCON0 instead of IODIR0... But that wasn't the whole problem.

 

[westaust55]

Have not got much time (and untested) but here is a start:

headers with some comments (in red) to make easier to follow:

' MCP23016 registers
Symbol GP0 = $00       [COLOR="Red"]; read write here for port 0[/COLOR]
Symbol GP1 = $01       [COLOR="red"]; read write here for port 1[/COLOR]
Symbol OLAT0 = $02
Symbol OLAT1 = $03
Symbol IPOL0 = $04     [COLOR="red"]; polarity for input on GP0 invert if 1[/COLOR]
Symbol IPOL1 = $05     [COLOR="red"]; polarity for input on GP1 invert if 1[/COLOR]
Symbol IODIR0 = $06    [COLOR="red"]; direction register 0[/COLOR]
Symbol IODIR1 = $07    [COLOR="red"]; direction register 1[/COLOR]
Symbol INTCAP0 = $08   [COLOR="red"]; interrupt data capture register for GP0 – read only[/COLOR]
Symbol INTCAP1 = $09   [COLOR="red"]; interrupt data capture register for GP1 – read only[/COLOR]
Symbol IOCON0 = $0a    [COLOR="red"];  Interupt action (bit0=1 for fast sample)[/COLOR]
Symbol IOCON1 = $0b    [COLOR="red"];  just a shadow of IOCON0[/COLOR]

initialise routine: (mods in red)

ConfigureMCP23016:

     ' Select the MCP23016 chip on the I2C bus
    i2cslave MCPAddr, i2cslow, i2cbyte
    
    ' Set GP0 to all writes / outputs ($00), GP1 to all reads / inputs   ($FF)
    i2cwrite [COLOR="Red"]IODIR0[/COLOR], (%00000000, %11111111) [COLOR="red"]; GP0 =Out GP1=In[/COLOR]
    pause 10
    
    [COLOR="red"]i2cwrite IOCON0, (%00000000)   ; normal sample rate for interrupt[/COLOR]
     pause 10

    ' set a "off" state on outputs
    i2cwrite [COLOR="red"]GP0, (%00000000)[/COLOR]
    pause 10
    
    ' Set the interrupt bit to occur when any of the GP1 states change
    [COLOR="red"][B]; [/B][/COLOR] i2cwrite INTCAP1, (%11111111) [COLOR="red"][B]; cannot write – this is read only[/B][/COLOR]
    pause 10
    
    i2cread GP1, (b0)
    pause 10
    return

 

[GDSever]

After doing some more reading on the HI2C commands, I realize that they aren't just limited to high speed. I've rewritten the program to use hi2cout and hi2cin, and I like it better - but my PicaxeVSM program is still not setting the outputs on the MCP23016... So I wonder if its a VSM problem.

I'll post something over in that forum and see if anyone has insight.