>> Why don't you connect 4 PIC16Fs in an I2C arrangement?
>> Each one would be a UART port?
>> The PIC16F87 comes to mind...One as I2C master, 3 slaves...
>
>Some points against this idea:
>
>1) You would need to write two different firmwares (if you use
> the same parts for master and slave, take care not to mix them);

It is not that hard to write the firmware so that all 4 copies are =

identical, and the master/slave relationship is determined by the address.

>2) Program the three slaves with different addresses (or waste
>at least two pins for address);

But with the extra chips, there are probably going to be sufficient extra =

pins anyway.

>3) Deal with an extra protocol (I2C) and bus (don=B4t forget
>that some PIC=B4s I2C modules don=B4t work well in slave mode);

I2C as a protocol is pretty straight forward. Microchip has some example =

code that seems to work well, it certainly did for me.

>4) I2C is slow, its bandwidth may be insufficient for
>three 38400bps channels;

Given the constraints he mentioned, I think a 400kbps I2C channel will be =

sufficient.

>5) Use more board area than one single bigger chip;

I reckon I could get 4 SOIC 18 pin chips in a similar area to an 80/100 pin =

TQFP. A 64 pin TQFP would be pushing it.

>6) May cost more than a PIC32.

But would it cost more than a Maxim UART? =


-- =

http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist