| View previous topic :: View next topic |
| Author |
Message |
Huriz
Joined: 31 May 2012
Posts: 13
|
| [SOLVED] Problem 12f629 between internal osc and XT |
 Posted: Thu May 31, 2012 12:34 pm |
 |
|
Hi everybody. I have been working and reading this forum but i didnt found anything about my problem.
I have these fuses
WITH XT CRISTAL
| Code: |
#fuses NOWDT,NOPROTECT,NOCPD,NOMCLR,PUT,XT,NOBROWNOUT
#use delay(clock=4000000)
int16 tx=0;
|
WITH INTERNAL OSCILATOR
| Code: |
#fuses NOWDT,NOPROTECT,NOCPD,NOMCLR,PUT,INTRC_IO,NOBROWNOUT
#use delay(internal=4000000)
int16 tx=0;
|
and the problem is here
| Code: |
tnum = read_EEPROM (0 );
switch (tnum) {
case 1:
tx=150;
break;
case 2:
tx=300;
break;
case 3:
tx=600;
break;
case 4:
break;
}
while(1)
{
output_high(PIN_A0);
delay_us(5);
output_low(PIN_A0);
delay_us(tx-7);
}
|
IM using
| Code: |
PICKIT 2 v2.61 Logic tool
PCWHD v.4.093
33pf capacitors
|
My code is so simple, I take a number from the memory of the pic, I do the switch, and with it I select the "period" time. After that I do tx-7 because I have 2 outputs and a delay, that is I think 7us more.
Ok, till here everything works ok, but using the internal oscillator or using a XT (with 4mhz) my periods are different.
It looks like with crystal the pic works faster or something like that.
Anyone knows what could have happened?
regards
Last edited by Huriz on Fri Jun 01, 2012 6:52 pm; edited 4 times in total |
|
 |
Huriz
Joined: 31 May 2012
Posts: 13
|
|
| Posted: Thu May 31, 2012 12:41 pm |
|
|
if i put this
delay_us(600-7);
instead of
delay_us(tx-7);
works fine but why this? |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Thu May 31, 2012 2:06 pm |
|
|
What is the current problem ?
Is it a different delay when you change oscillator fuses between
INTRC_IO and XT ?
Or, is it a different delay for each of these statements:
delay_us(600-7);
and
delay_us(tx-7);
What is the difference in delay times ? How are you measuring it ?
Are you using an oscilloscope or logic analyzer ?
Also
1. Is this a Proteus project ?
2. What is your CCS compiler version ? |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9163
Location: Greensville,Ontario
|
|
| Posted: Thu May 31, 2012 6:26 pm |
|
|
The internal oscillator will never be the same frequency as an external crystal. Changes in Vss, temperature,etc. will affect it. Check the datasheet on the specs and range the int.osc. will run at.
You should use a 'scope and compare the PIC running using both configurations. |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19338
|
|
| Posted: Fri Jun 01, 2012 1:24 am |
|
|
Simplify first. Just test with a basic pin toggle, no calculated delays etc..
Then post (as PCM programmer asks), your compiler version. We can't test and see if their are any oddities with the compiler in certain circumstances, without knowing this.
Then show the code you are using to setup with the internal oscillator.
Though as Temtronic says, there will be a difference between the internal, and external oscillators (a few percent possibly), it won't be large.
Other possibility is that the crystal is not actually running at 4MHz. What loading capacitors are you using?. What is the actual crystal part number?. Though relatively 'rare', there are occasions, when the oscillator can lock onto an overtone, especially with some crystals designed for low power drive, without a series resistor being fitted.
Best Wishes |
|
|
Huriz
Joined: 31 May 2012
Posts: 13
|
|
| Posted: Fri Jun 01, 2012 4:12 am |
|
|
++++++++++++++
Informatino updated in the main post, with the compiler verison and more things
++++++++++++++
my problem is when i put the variable in the delay, it doesnt work, i dont know why.
If i put a number in the delay it works fine with both systems(internal and external oscilator), but when i put the variable in the delay it doesnt work fine with both system.
right now, i have solved the problem with this "bad" system, but is just what i dont want.
| Code: |
switch (tnum) {
case 1:
loop_1();
break;
case 2:
loop_2();
break;
case 3:
loop_3();
break;
case 4:
break;
}
|
and after i have put some functions like these
| Code: |
void loop_1()
{
while(1)
{
output_high(PIN_A0);
delay_us(3);
output_low(PIN_A0);
delay_us(150-7);
}
}
void loop_2()
{
while(1)
{
output_high(PIN_A0);
delay_us(3);
output_low(PIN_A0);
delay_us(300-7);
}
}
void loop_1()
{
while(1)
{
output_high(PIN_A0);
delay_us(3);
output_low(PIN_A0);
delay_us(600-7);
}
}
|
with this system it works, but...i dont like the space of rom that is using in the pic (is just a example i have 29 fuctions like these, all of them are the same functions with different delay times. |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9163
Location: Greensville,Ontario
|
|
| Posted: Fri Jun 01, 2012 5:23 am |
|
|
generally speaking...
The early versions of the compiler did not allow for a variable to be the delay value....others will know when CCS changed this.
You might not be able to use a 16bit value ?
Also I'd just used fixed numbers(ie. delay_ms(143) not delay_ms(150-7)).
There _might_ be a problem doing the 'math' inside the delay..
Others that use that compiler will know the answers ! |
|
|
Huriz
Joined: 31 May 2012
Posts: 13
|
|
| Posted: Fri Jun 01, 2012 5:31 am |
|
|
what do you mean with
| Quote: | | You might not be able to use a 16bit value ? |
so, you mean that the problems could come from
- the early version of my compiler
- for do a math operation inside the delay |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9163
Location: Greensville,Ontario
|
|
| Posted: Fri Jun 01, 2012 5:59 am |
|
|
You'll have to read the manual for your version of the compiler but ..
in the beginning..
you could only pass an 8 bit value for the delay_xx(nnn) function.Over the years CCS upgraded and allows 16 bit values to be used. I think they always allowed a fixed 16 bit value ie: delay_ms(12345) but NOT nnn=12345; delay_ms(nnn);
To prove is easy.
just try delay_ms(250), delay_ms(1000), nnn=250 delay_ms(nnn), nnn=1000 delay_ms(nnn)
in a small program and see what happens.
I don't have your version of the compiler so I can't test, it should take you 5-10 minutes to see what's happening.
Jay |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19338
|
|
| Posted: Fri Jun 01, 2012 7:35 am |
|
|
As a comment though, there are several things that are rather silly if trying to keep the size small.
Why not:
| Code: |
tnum = read_EEPROM (0 );
do {
output_high(PIN_A0);
delay_us(5);
output_low(PIN_A0);
switch (tnum) {
case 1:
delay_us(143);;
break;
case 2:
delay_us(293);
break;
case 3:
delay_us(593);
break;
case 4:
break;
}
} while (TRUE);
|
Do a search here. I remember there being some problem with the delay code, on PIC12 chips, around your compiler version.
Best Wishes |
|
|
Huriz
Joined: 31 May 2012
Posts: 13
|
|
| Posted: Fri Jun 01, 2012 8:26 am |
|
|
Hi again, I have been doing more test around the int16, and yes, delay_us(int16) doesn't work fine.
I have solved it, thank to all people for help me and for give me new ideas about how solve my problem.
cheers!
 |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
