16F688 INTERNAL CLOCK UART PROBLEM

[dodang8811]

This is my code but MCU sends wrong message.

[temtronic]

It could be a 'timing' problem. The internal osc (currently 4MHz) could be off a bit, also since the internal osc can be selected as 8MHz, I'd use that.
2 options to try...

1) use a real 4MHz xtal (and 2 caps), alter code, recompile and test.

2) use your current setup but try 1200 Baud. Slower speeds tend to be more 'forgiving'.

You do have a 'max232' type interface chip between the PIC and PC ?

Another test is to remove PIC, jumper the TTL rx to the TTL tx pins and 'loopback' the PC terminal hardware to prove there's nothing wrong on that 'side' of the project.

Please post back what happens and we can help you from there.

hth
jay

[Ttelmah]

What voltage are you running the chip?. What temperature?.

The internal oscillator, on this chip, is guaranteed to be accurate enough for RS232, _if you are running at 3.5v_. Otherwise it 'may be accurate enough' at 5v and room temperature (but is getting borderline), but outside this, it is well outside the required accuracy.....

Best Wishes

[dodang8811]

First, I have 3 16f688 chips and the results are all same.

//temtronic

I'm using max3232 chip between PC and PIC.
All frequencies(1M, 2M, 4M, 8M...) sends wrong data.

1) I can't use xtal because the chip is in a product. It's smd type circuit and I can't change it.

2) 1200 Baud send worse data like "fe fe fe fe fe ...."
I can't understand why slower speed results are worse.

3) Loopback test is OK.

//Ttelmah

When I used 3.5V, I got worse data.


Please give me more advice~

[temtronic]

hmm..
well this is curious !!
I don't have a 16F688 so I 'white breadboarded' a 16F887 which has same/similar internal oscillator section, grabbed your code, made appropriate changes for the UART pins., and used the 887 device header file.

It works fine for me. I'm using Realterm for the PC terminal program and a TTL<>USB module for the interface.

Though it's a long shot, perhaps the fail-safe oscillator is enabled?
I would select the fuse 'NOFCMEN' though I didn't in my test.

Another question. Did you compile under MPLAB in 'debug' or 'release' mode?
If you compiled under 'debug'( the supplied default) , your code will NOT run correctly,usually slow.
I ran into this 'debug/release' issue a few years ago and my default is 'release' as I debug code in the real world not in simulations.

I'll keep the board up and running, but 'on the surface' I can't see anything wrong in your code or hardware setup.

hth
jay

[dodang8811]

Is smd type chip more accurate?

Actually, yesterday I tested on DIP type chip, breadboard and max3232.

And I tested on my product few minutes ago.

In the product, PIC sends data to CPU.

And I checked data from CPU. It's working well now.

"123456789", "hello"

I don't know why, but thank you very much for your concerning.

[temtronic]

Glad it's up and running.
I'm always happy to help.

Jay

[ezflyr]

[dodang8811]

//ezflyr

Thanks. I'll test on a new breadboard and jumper later.

Have a good day~!