Instant Interrupts - Revisited

[sougata]

Hi,

When using multiple interrupt sources you can prioritize which Interrupt flags to check when vectored to ISR, still the results may not be deterministic every time. Say you are already servicing a lower priority interrupt (not Hardware but DT's scheme) a higher priority interrupt like your time tick occurs. Generally adding the rolling timer with the offset value does the trick however if you would like more precision then use the compare module on the PIC which resets the associated timer automatically and gives a rock solid time base.

[mackrackit]

You need to fine tune the timer, not the OSC.
Here is an example
http://www.picbasic.co.uk/forum/show...3259#post23259
And here is another
http://www.picbasic.co.uk/forum/show...5912#post75912
And this will help with the calcs.
http://www.picbasic.co.uk/forum/cont....-PICMultiCalc

Hi, we are pleased to see that you are using our forum. We'd like you to become a member of our community. Membership is FREE, please REGISTER and view our forums without these annoying reminders! Once registered you may post on the forums, download from the file areas and use the WIKI without interruption. REGISTER NOW

[DaTiNs]

Hello good people,


I follow this forum for a long time, it is excellent! I found a lot of good stuff on it.


Now I have a problem with the code from Mr. Darrel Taylor.


Thus, I use pic18f2550 and connect DS1307z and 4 PWM channels to PORTA


SDA Var PORTA.1
SCL Var PORTA.0


ASM
SPWM_LIST macro
SPWM_PIN PORTA, 2, _CH1
SPWM_PIN PORTA, 3, _CH2
SPWM_PIN PORTA, 4, _CH3
SPWM_PIN PORTA, 5, _CH4
endm
SPWM_INIT SPWM_LIST


INT_LIST macro
INT_Handler TMR1_INT, SPWMhandler, ASM, yes
endm
INT_CREATE
ENDASM
@ INT_ENABLE TMR1_INT


Now i got problem with this configuration. When i enable SPWM_LIST then RTC stop working, in fact he working but value he give, its not good. When i disable software pwm, RTC is working good.


I hope that you understande what is my problem and that one of you has a solution for this problem.


Sory for my english


Thanks a lot, Damir

[Darrel Taylor]

Do you have this in your program ... ?

ADCON1 = 15

[DaTiNs]

Thank you for quick reply!

Yes i have, this is configuration for my pic



ADCON1 = $0F
CMCON = 7
UCFG.3 = 1
PORTA = 0
PORTB = 0
PORTC = 0

[DaTiNs]

Just want to add,when I turn off TRM1 before reading i2c, rtc is operating normally, but problem is with LED strips on PWM channel they strart flashing.

T1CON.0=0
I2CRead sda,scl,$D0,$00,[RTCSec,RTCMin,RTCHour,RTCDay,RTCDate,RTCMonth,RTCY ear]
T1CON.0=1

[Darrel Taylor]

Damir,

I can duplicate the issue here, which appears to be the old R-M-W hardware problem.
I moved the PWMs over to PORTC, and everything works fine.

Since I2C uses pull-ups instead of bi-polar driven outputs, it's more susceptable to R-M-W problems.

If you cannot move pins around on your project, you can get away with the following.
In general, I do not recommend doing this type of modification.
But in your specific case, it should work.

  SPWM_LIST macro 
      SPWM_PIN LATA, 2, _CH1 
      SPWM_PIN LATA, 3, _CH2 
      SPWM_PIN LATA, 4, _CH3 
      SPWM_PIN LATA, 5, _CH4 
    endm

[Darrel Taylor]

I should have also mentioned that when using the LATA register, you'll need to set the pins to output manually.

TRISA = %11000011

[DaTiNs]

Thank you very much!

I must add TRISA = 0 to the begin of code, and then start work great.

One more time, thank you, Damir.

[SteveB]

Darrel,
While poking around and getting ideas, I got to looking at ver 1.2 of "Elapsed_INT-18.bas". I was particularly interested in the nice routine to calculate the timer reload constant:

' -------------- calc timer reload Constants -------------------------------
ASM
T1PS = 1                             ; start with 1:1 postscaler
TimerConst = ((OSC*1000000)/4/100)   ; how many timer ticks will it take
  while TimerConst > 65400           ; if it's more than the timer can count
T1PS = T1PS * 2                      ;   double the postscaler
TimerConst = TimerConst / 2          ;   halve the count
  endw
TimerConst = 65536 - TimerConst + 8  ; final reload value

But, I think there is an error in the formula. The absolute time it takes to execute the 8 instruction cycles doesn’t change. But, the time in relation to the TMR1L will be proportional to the prescaler. So, that should be accounted for in the formula, as such:

TimerConst = 65536 - TimerConst + (8/T1PS)  ; final reload value

Since it's not likely the prescaler will need to be above 1:8, the 8/T1PS shouldn’t get too low.

Thanks again for your great work and inspiring examples,
Steve

[SteveB]

The code could also be changed thus, not that it makes a difference:

T1PS = 1                             ; start with 1:1 postscaler
TimerConst = ((OSC*1000000)/4/100)+8 ; how many timer ticks will it take
  while TimerConst > 65400           ; if it's more than the timer can count
T1PS = T1PS * 2                      ;   double the postscaler
TimerConst = TimerConst / 2          ;   halve the count
  endw
TimerConst = 65536 - TimerConst      ; final reload value

[jamie_s]

Hi Darryl,

Currently getting an error with Latest MPLAB IDE:
MPLAB 8.90.00.00

PIC is an 18F45K22 and DT_INTS-18 is latest from your website, both MPLAM and DT_INTS were downloaded from the respective sites on 15/3/2013


ERROR: Unable to execute mpasmwin.Warning[206] M:\<FOLDERPATH><FILE>ASM 2170 : Found call to macro in column 1. (INT_ENABLE)
Error[116] M:\<FOLDERPATH><FILE>.ASM 696 : Address label duplicated or different in second pass (Z00032)
Error[116] M:\<FOLDERPATH><FILE>.ASM 759 : Address label duplicated or different in second pass (Z00033)
Error[116] M:\<FOLDERPATH><FILE>.ASM 812 : Address label duplicated or different in second pass (Z00034)
Error[116] M:\<FOLDERPATH><FILE>.ASM 843 : Address label duplicated or different in second pass (Z00035)

Can you advise if there is any fix for this or do i need to use an older version of MPLAB IDE

[mackrackit]

You might try here
http://support.melabs.com/threads/11...hlight=MPLAB-X

[Christopher4187]

I've used the DT_INTS sucessfully with an 18F4550. I had to move to an 18F87J50 but the DT_INTS would not compile. Not sure if I'm doing something wrong or if the DT_INTS was never tested with J series PIC's. When I compile, I get the following warning message:

Unable to fit variable VP_DelayCount in requested bank 0

In the "VirtualPort.bas" code, I changed this code:

VP_DelayCount  VAR WORD BANK0  ; Used for internal Delay only

to this:

VP_DelayCount  VAR WORD BANK1  ; Used for internal Delay only

and it compiled as expected. I have no clue if this will cause unexpected problems with my program or if this is the correct way to solve it but I haven't noticed any ill effects yet. Seems like this problem could be related to the SFR's but it's way above my level of knowledge.

Just thought I'd pass it along.

[Darrel Taylor]

NO! You can't just change the variable to BANK1!
That variable must be in BANK0.

Both DT_INTS-18 and VirtualPort.bas fit quite comfortably in an 18F87J50, with room to spare. Even when using PBPL.

How many variables have you declared in BANK0 in your program?
Or, what other include files are you using that may have variables placed in BANK0.

[Christopher4187]

NO! You can't just change the variable to BANK1!

Oh. The door was locked and I couldn't find the key so I just picked the lock.

How many variables have you declared in BANK0 in your program?
Or, what other include files are you using that may have variables placed in BANK0.

Good questions and ones that I wish I had answers for. I'm using HIDmaker so that's probably the root cause. Do I need to figure this problem out or is there a workaround?