Understanding timer issues [Archive]

View Full Version : Understanding timer issues

boroko

- 28th March 2009, 22:59

Hi all,
trying to understand why this isn't working as it seems it should.

The Transmit: TMR1 Int handler is ~30mS and works fine. At the end of the TMR1 int, TMRO is given settings so that the first cycle through, there should be a ~190uS int cycle. After the first cycle, the TMR is reloaded so that the next passes are ~2.74 mS. For some reason, the two timer loads affect each other and changing one affects both areas in ways that I'm not understanding. I can't get any single point to adjust the two timing areas.

DEFINE OSC 8
DEFINE DEBUG_REG PORTA
DEFINE DEBUG_BIT 0
DEFINE DEBUG_BAUD 2400
DEFINE DEBUG_MODE 0
CLEAR
TRISA = %00100000 ' PortA : All outputs except RA.1 for INT_INT
TRISB = %01000100 ' PWM1,3,5 outputs,RB2=INT, RB0,2,3: DIO
ADCON1 = %00001111 '
T1CON = %10000001 ' 16bit, 1:1 ps, 30 mS INT
RCON = %00011111 ' NO Priority for interrupts
OSCCON = %11111110 ' INTOSC, 8MHz
DEBUG "CSS1330-2inc",13,10
'::::::::::::
INCLUDE "DT_INTS-18.bas" '
INCLUDE "ReEnterPBP-18.bas" '
'::::::::::::

EchoTime VAR WORD
OutPulseCnt VAR BYTE
Gain VAR BYTE
GCA VAR PORTA.1
GCB VAR PORTA.2
GCC VAR PORTA.3
GCD VAR PORTA.4
PulseOut VAR PORTB.0
EchoIn VAR PORTB.2
ledR VAR PORTB.3 '
ledG VAR PORTB.4 '
Prog VAR PORTB.6 '
CntSw VAR WORD
flag VAR BYTE
IncCnt VAR BYTE
MdCnt VAR BYTE
TimOut VAR BYTE
Lock VAR flag.0 ' 1= locked, 0 = unlocked

ASM
INT_LIST macro ; IntSource, Label, Type, ResetFlag?
INT_Handler INT2_INT, _GetEcho, ASM, yes ;straight I/P
INT_Handler TMR0_INT, _BlankGain, ASM, yes
INT_Handler TMR1_INT, _Transmit, ASM, yes
endm
INT_CREATE ;
ENDASM
@ INT_ENABLE TMR1_INT ; enable Timer 1 interrupts
'************************************************* **************************
GOTO Start

'*** Sub to Select MODE *************************************
ModeSelect:
' first time through, need to unlock. Then can pass to others
IF MdCnt = 2 THEN '1st check for unlock
Lock = 0
ENDIF
IF Lock = 0 THEN
ledG = 1 '
ELSE
ledG = 0
GOTO DoneCk
ENDIF
SELECT CASE MdCnt
CASE 1
Lock = 1
DEBUG "Case 1",13,10
GOTO DoneCk
CASE 2
Lock = 0
GOTO DoneCk
DEBUG "Case 2",13,10
CASE 3
DEBUG "Case 3",13,10
CASE IS >3
DEBUG "Case >3",13,10
Lock = 1

END SELECT
DoneCk:
IncCnt = 0 ' reset MODE counter
RETURN
'**********************
'* END Subs
'**********************
Start:
Lock = 1
Mainloop:
DEBUG #EchoTime,13,10

'code for menu selection here
GOTO Mainloop
'*********** Int handlers ***********
'---[TMR1 - interrupt handler]-------
' TMR1 = 16 bit, 1:1, 65535 - 5543(preload) ~30 mS cycle
Transmit:
@ INT_DISABLE INT_INT ; Disable external ints to stop Echo RX
TMR1H = $15 '
TMR1L = $A7 '
OutPulseCnt = 0 ' start new cycle
Gain = 0 '
PORTA = Gain ' turn off the gain
ASM
BSF PORTB,0
NOP
NOP
NOP
NOP
NOP
NOP
NOP
BCF PORTB,0
NOP
NOP
NOP
NOP
NOP
NOP
NOP
BSF PORTB,0
NOP
NOP
NOP
NOP
NOP
NOP
NOP
BCF PORTB,0
ENDASM
T0CON = %11001000 ' On, 8b, PSon, 1:1, For Blanking
TMR0L = 70 ' Blanking inhibit 190uS ???
@ INT_ENABLE TMR0_INT ; Enable TMR0 to start Blanking/Gain block
@ INT_RETURN
'---[TMR0 - interrupt handler]-------------------------------
' inhibit GetEcho for 1 cycle through TMR1 with value loaded at end of transmit
' code then change TMR1 to ramp up gain in 2.74mS steps
BlankGain:
OutPulseCnt = OutPulseCnt +1 '1st pass, TMR0 Blanking set @ end of "TX"
SELECT CASE OutPulseCnt
CASE 1
T0CON = %11000101 ' On, 8b, PSon, 1:64, for Gain step
TMR0L = 170 ' Preload to 170 for 2.74mS INT
@ INT_ENABLE INT2_INT ; enable INT2 to catch Echo
@ INT_RETURN ; go back for next INT on TMR0
CASE 2
GOTO MoreGain ' 2nd pass: Gain inc 1
CASE 3
GOTO MoreGain ' gain inc'd to 2
CASE 4
GOTO MoreGain ' gain inc'd to 3
CASE 5
GOTO MoreGain ' gain inc'd to 4
CASE 6
GOTO MoreGain ' gain inc'd to 5
CASE IS > 6
GOTO EndGain
END SELECT
MoreGain:
Gain = Gain + 2 ' RA.0 is I/P, so count x2
PORTA = Gain 'sent out RA.1-RA.4, set gain amp
SkipGain:
@ INT_RETURN

EndGain:
@ INT_DISABLE TMR0_INT ; Shut-off Blanking & Gain incr until next cycle
@ INT_RETURN

'---[INT - interrupt handler]-------------------------
GetEcho:
EchoTime.LowByte=TMR1L
EchoTime.HighByte=TMR1H ' capture the counts
@ INT_RETURN
END



The DEBUG quits after the splash. I suspect it is getting blown up by the interrupt.

Thanks for looking at it.
Bo

boroko

- 30th March 2009, 13:57

Hi all,
Still working on a timer issue and hoping that someone might see what I'm not.
Here is a little better explanation and some cleaned up code if you don't mind looking at it again.

The problem is with the TMR0. It seems to not recognize the preloaded value. I can't seem to make any changes that are predictable. The function is this:

1) TMR1 starts a ~30mS cycle with a pulse. (This timer works well.)
2) At the end of the pulse, it loads values for, and enables TMR0
3) First time through TMR0, it should wait ~.9mS for the interrupt and then
enable INT2_INT to capture the count on TMR1 when a pulse comes in
4) At the end of the first TMR0_INT, a new value gets loaded for an ~2.74mS delay,
5) each of the next cycles use the 2.74mS delay until they go through a certain number and get shut off.
6) The next TMR1_INT starts the whole thing again.

DEFINE OSC 8
TRISA = %00100000 ' PortA : All outputs except RA.1 for INT_INT
TRISB = %01000100 ' PWM1,3,5 outputs,RB2=INT, RB0,2,3: DIO
ADCON1 = %00001111 ' Set up ADCON1
T1CON = %10000001 ' 16bit, 1:1 ps, 30 mS INT
'T0CON = %11000101 ' On, 8b, PSon, 1:64 : DIDN'T HELP
RCON = %00011111 ' NO Priority for interrupts
OSCCON = %11111110 ' INTOSC, 8MHz
'::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::
INCLUDE "DT_INTS-18.bas" ' Base Interrupt System
INCLUDE "ReEnterPBP-18.bas" ' Include if using PBP interrupts
'::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::
' Variable definition
'::::::::::::::::::::::::::::::::::::::::::::::::: :::::::::::::::::::::::::::::
EchoTime VAR WORD
OutPulseCnt VAR BYTE
Gain VAR BYTE
PL1lo CON $A7 ' Preload for TMR1 LO byte, Cycle time
PL1hi CON $15 ' Preload TMR1 HI byte ,PL =5543/ 1:1
'************************************************* **************************
ASM
INT_LIST macro ; IntSource, Label, Type, ResetFlag?
INT_Handler INT2_INT, _GetEcho, ASM, yes ;straight I/P
INT_Handler TMR0_INT, _BlankGain, ASM, yes
INT_Handler TMR1_INT, _Transmit, ASM, yes
endm
INT_CREATE ; Creates the interrupt processor
ENDASM
@ INT_ENABLE TMR1_INT ; enable Timer 1 to start sonic cycle
'************************************************* **************************
Mainloop:
' code here
GOTO Mainloop
'*********** Int handlers **************************************************
'---[TMR1 - interrupt handler]-------------------------------
' TMR1 = 16 bit, 1:1, 65535 - 5543(preload) =30 mS cycle, (WORKS WELL)
Transmit:
@ INT_DISABLE INT_INT ; Disable external ints to stop Echo RX
TMR1H = PL1hi 'Preload HI byte
TMR1L = PL1lo 'Preload LO byte =
OutPulseCnt = 0 'clear everything & start new cycle
Gain = 0 ' zero the gain counter
PORTA = Gain ' turn off the gain
ASM
BSF PORTB,0 ; pulse out
NOP
NOP
BCF PORTB,0
NOP
NOP
ENDASM

'+++++++++++ HERE IS THE ISSUE and different things tried ++++++++++++++++++++

' T0CON = %11000100 ' On, 8b, PSon, 1:32, For Blanking
' TMR0L = 199 ' Blanking delay: .9mS : NOPE
@ INT_ENABLE TMR0_INT ; Enable TMR0 to start Blanking/Gain block
' TMR0L = 199 ' Blanking delay: .9mS : WAY OFF!
' T0CON = %11000100 ' On, 8b, PSon, 1:32
TMR0L = 40 ' Blanking delay: .9mS : CLOSEST!? WHY???
T0CON = %11000010 ' On, 8b, PSon, 1:8
' TMR0L = 40 ' tried loading after T0CON, no dif
@ INT_RETURN
'---[TMR0 - interrupt handler]-------------------------------
' 1 cycle through TMR0 with value loaded at end of transmit (TMR1), (~.9mS)
' then change TMR0 to ~2.74mS increments
BlankGain:
OutPulseCnt = OutPulseCnt +1 '1st pass, TMR0 Blanking set @ end of "TX"
SELECT CASE OutPulseCnt
CASE 1
' TMR0L = 84 ' PL 2.74mS INT : THIS MESSES IT ALL UP
' T0CON = %11000100 ' On, 8b, PSon, 1:32, for Gain step
@ INT_ENABLE INT2_INT ; enable INT2 to catch Echo
@ INT_RETURN ; go back for next INT on TMR0

'+++++++++++++END OF SUSPECTED THINGS THAT CAN AFFECT THE ISSUE+++++++++++

CASE 2
GOTO MoreGain ' 2nd pass: Gain inc 1
CASE 3
GOTO MoreGain ' gain inc'd to 2
CASE 4
GOTO MoreGain ' gain inc'd to 3
CASE 5
GOTO MoreGain ' gain inc'd to 4
CASE 6 '.......
'........
CASE IS > 14
GOTO EndGain '
END SELECT
MoreGain:
Gain = Gain + 2 ' increment gain: RA.0 is I/P, so count x2
PORTA = Gain 'sent out RA.1-RA.4 to control gain on amp
EndGain:
@ INT_DISABLE TMR0_INT ; Shut-off Blanking & Gain incr until next cycle
@ INT_RETURN

'---[INT - interrupt handler]-----------------------------------------------
GetEcho:
'@ INT_DISABLE TMR0_INT ; Haven't seen a need for this yet
EchoTime.LowByte=TMR1L ' capture the counts
EchoTime.HighByte=TMR1H
'@ INT_ENABLE TMR0_INT ; Re-enable TMR0
@ INT_RETURN
END



Again, thank you for looking at it.

Bo

Darrel Taylor

- 30th March 2009, 22:04

Transmit:
@ INT_DISABLE INT_INT ; Disable external ints to stop Echo RX

INT_INT is the same as INT0_INT.
But you are using INT2.

In the BlankGain: handler, you have ...

CASE IS > 14
GOTO EndGain '

That will use PBP's system variables. Which means the handler can not be an ASM type anymore.

A CASE ELSE might fix that problem.

There may be other issues, but that's a start.

boroko

- 1st April 2009, 02:32

Thanks for the observations. The INT_INT was a leftover from a different chip and I didn't catch it there.

I'm afraid the problem still exists. I have been trying to get a pattern so that I can at least identify why the timer seems to be out of control. I'm sure it is going to be something stupid, but last nights session didn't reveal it to me. I'm going to devise a few more tests tonight to try again. Decided to take a break from it and work on other areas for a while for a break.

I will say that my understanding of PBP and PIC's in general has come a long way by working through recent projects. That has only been possible because of all of the help that this forum and it's members. I know that it can be frustrating, but your efforts are very appreciated.

Thanks
Bo

boroko

- 7th April 2009, 02:56

A few more questions on this mystery...

I'm using a 18F1330 and TMR0 is a selectable width timer (8/16bit).
My understanding is that because it has TMR0L and TMR0H, in 8 bit mode, you preload TMR0L only. There is no TMR0 register specifically. Do I have that right? Are there any other issues with "changeable" timers

I know that I have read threads concerning changing timer values on the fly in the program, but at the time, I was not facing those problems and wasn't smart enough to leave a trail to find my way back now that I need the insight. I have tried searching, but I haven't hit the right terms yet. Anybody remember a link to their favorite discussion on changing timer values?

If anything jumps to mind, I would appreciate it. In the mean time, I'll see if I can cajole an answer from some testing.

Thanks, I learn a great deal from you all. (sometimes, what questions NOT to ask!...) ;-)

Bo