Quick rs232 question

dyeatman · Joined: 06 Sep 2003 Posts: 1934 Location: Norman, OK

I have used the NXP SC16IS75x serial UARTS with good success. I put a driver for these in the code library.

dave10 · Joined: 09 Apr 2013 Posts: 14 Location: UK

Ttelmah · Joined: 11 Mar 2010 Posts: 19539

So much depends on what the UART has to do.
I did post here some time ago, a crude software receive using a timer interrupt, as an example, to show how this can be done. For baud rates that are no more than perhaps 1/1000th the CPU clock, this can be simple, cheap and pretty reliable. A google search should find the post. Had one mistake, and that was fixed by another poster a moment later in the thread.
The higher the baud rate the harder this is.TX & RX can be done in the same interrupt. Other interrupts can interfere.
You can cheat a lot of ways if you can change the baud rates involved. For instance I recently needed to get data 'out' of a chip without any spare interrupts. The software UART was perfectly acceptable, if I turned the baud rate up, so that the entire byte is sent in only a few uSec, allowing then interrupts to be disable on a 'per character' basic, and still not interfere with the other peripherals.
The Phillips UART will work OK with a 5v device.
It has 5v tolerant inputs (so you can use pull-ups to 5v on the I2C bus), and the output is genuinely 'open collector', so can also drive a signal that is pulled up to 5v. They designed it to be 'reverse compatible' with the older 5v version, except for the supply.
There is a big 'caveat' waiting to catch you though. The 5v tolerant inputs, are only 5v tolerant, once it's 3.3v supply is established. You need to ensure it gets it's supply at least as early as the 5v pull-up resistors....

Best Wishes

dave10 · Joined: 09 Apr 2013 Posts: 14 Location: UK

Thankyou Ttelmah for the prompt reply really appreciated.

I think I'll try using the NXP device to start with and play around with the test code from the code lib for starters ??
Regarding the potential hazard you mention. Could I not declare the I2C pins within the pic as inputs intially and upon power-up of the 3.3V rail change the pin configuration to I2C.

Problem being the 3.3V Regulator is powered from 5V supply

What do you think is this a good way to go ?

dave10 · Joined: 09 Apr 2013 Posts: 14 Location: UK

Thankyou Ttelmah for the prompt reply really appreciated.

I think I'll try using the NXP device to start with and play around with the test code from the code lib for starters ??
Regarding the potential hazard you mention. Could I not declare the I2C pins within the pic as inputs intially and upon power-up of the 3.3V rail change the pin configuration to I2C.

Problem being the 3.3V Regulator is powered from 5V supply

What do you think is this a good way to go ?

temtronic · Posted: Tue Apr 09, 2013 10:04 am

Why not just use the PIC18F46K22 or similar device? It has 2 hardware UARTS as well as a ton of memory and other peripherals.Has to be simpler to wireup, no additional chips,simpler coding
I 'migrated' from the 4550 when I found out I could buy ttl-usb modules for $2...
Just an idea....