Forums

 Hi, your program has some issue, please get a water glass than a flame ;)

ecstatist

I get all the interrupts/handling/routines doing whatever I need but as mentioned can not "see OUTs" when I try (even if OUTMOD1 ( =SET)  is  tried)

 If you use setmod you enable timer logic to do some action but you need select the related output pin to TAx output funnction and to do that you need an instruction  setting the bit related to that.

ecstatist

SetupP1          bis.b         #001h,&P1DIR                                              ;P1.0 LED

Example

  bis.b BIT1,&P1Sel    ; This select pin p1.x as TA output sconnecting from OUT register

ecstatist

CCR1_ISR       nop
                           mov.w  #0018h,&027Ch                                                ;change stack,RETI SCG0 SCG1 on

  AAAARRRRRGGGGGHHHHHHH WHAT IS THIS ORRIBLE RUBBISH??? DONT DO THAT!!!!!!!!!!!!!!!!!!!!!!!!!!

  Please use SP relative addressing and NEVER use hardwired stack manipulation, you can disrupt another interrupt handle or some other data losing control of program!

  BIS

  BIC  #xxx,&[offset]SP

ecstatist

FAST                  mov.w  #TAIE+TASSEL_2+MC_3,  &TACTL                ;input=fastSMCLK
                           reti

CCR2_ISR       nop
SLOW                mov.w   #TAIE+TASSEL_1+MC_3,&TACTL                 ;input=slowACLK
                            reti

 Again to set reset bit use BIC and BIS otherwise you can set a wrong state and hang subsystem, this is not the case of your simple program but as I see you are using a bad programming style.

 To have OUTMOD visible select a pair of TA1 and TA2 pin and enable corresponding bit on PxSEL and you can watch OUTMOD action on hardware.

Roberto, thank you for your quick reply - I only noticed it now and will try your suggestions tomorrow or at least by Monday.

"To have OUTMOD visible select a pair of TA1 and TA2 pin and enable corresponding bit on PxSEL and you can watch OUTMOD action on hardware."

 I understand this as - connecting an LED or voltmeter to the correct pin out to "watch OUTMOD action" - Do you think that when I have done the correct P1SEL and P1DIR's that I will be able to "watch OUTMOD action" in the REGISTER view of CCS?  I'll check.

"AAAARRRRRGGGGGHHHHHHH WHAT IS THIS ORRIBLE RUBBISH??? DONT DO THAT!!!!!!!!!!!!!!!!!!!!!!!!!!"          I'm still laughing  - I'm so sorry - for the crude programming!

Sometimes when I get rolling along (desperate and crazy), I use quickly "cut and try" rather than "measure and think."  So far, I have not had to write many complex programs so I have got away with this lack of rigour.

Again thank you and I will keep you informed

ecstatist

in continuous, up/down mode and, that "OUT" in TA0CCTLx (x = 1 or 2) does not seem to appear in the CCS  "register display" regardless of which OUTMOD (0-7) I try, stepped or "breakpointed."

Indeed, OUT is not a reflection of the current pin state. It defines the pin state in OUTMOD_0.
OUT only returns what was written to it.
If you switch from any OUTMOD_x to OUTMOD_0, the content of OUT will be written to the internal state of the output unit and to the pin. If you switch back to any toggle mode, the internal state will remain on what OUT was, and then toggle. So you can define a 'start condition' for the toggle mode. Or use this bit as pendant to PxOUT.y in OUTMOD_0 without need to change PxSEL. This is useful if you do PWM, as in PWM mode, you cannot program both, 0% and 100% duty cycle. Since both mark the same place on teh counting circly, only one of the m can be directly progreammed (depends on polarity), while the other one has to be done by switching to OUTMOD_0 with the proper value written to OUT bit.

However, in no case, OUT will change by its own, so you'll never see it changing in the debugger register view, unlexss your code has changed it manually.

As Roberto already pointed out, your trick to exit LPM by stack content manipulation won't work. As soon as your main uses one more local variable, it will access the wrong memory location. Other reasons may shift the stack content too.

Something like
bic #LPM3_BITS, 2(R1)
will do better (in this case accessing the word on top of stack, clearing the LPM bits

Jens-Michael Gross

ecstatist

in continuous, up/down mode and, that "OUT" in TA0CCTLx (x = 1 or 2) does not seem to appear in the CCS  "register display" regardless of which OUTMOD (0-7) I try, stepped or "breakpointed."

Indeed, OUT is not a reflection of the current pin state. It defines the pin state in OUTMOD_0.
OUT only returns what was written to it.
If you switch from any OUTMOD_x to OUTMOD_0, the content of OUT will be written to the internal state of the output unit and to the pin. If you switch back to any toggle mode, the internal state will remain on what OUT was, and then toggle. So you can define a 'start condition' for the toggle mode. Or use this bit as pendant to PxOUT.y in OUTMOD_0 without need to change PxSEL. This is useful if you do PWM, as in PWM mode, you cannot program both, 0% and 100% duty cycle. Since both mark the same place on teh counting circly, only one of the m can be directly progreammed (depends on polarity), while the other one has to be done by switching to OUTMOD_0 with the proper value written to OUT bit.

However, in no case, OUT will change by its own, so you'll never see it changing in the debugger register view, unlexss your code has changed it manually.

This is a great explanation of the timer module OUT bit.

ecstatist

  I understand this as - connecting an LED or voltmeter to the correct pin out to "watch OUTMOD action" - Do you think that when I have done the correct P1SEL and P1DIR's that I will be able to "watch OUTMOD action" in the REGISTER view of CCS?  I'll check.

You can actually use the CCI bit to monitor the current state of the pin where you are currently outputting the timer-generated signal/PWM. So you can look at this CCI bit (instead of the OUT bit) in the TACCTLx register in CCS to see the current state when you hit a breakpoint. Though I do recommend hooking up an oscilloscope if you really want to see how the timing is coming out on the pin.

Regards,

Katie

Roberto Romano

BIC  #xxx,&[offset]SP

OOOOPPS..

 sorry for malformed instruction, the correct stack manipulation for BIC or BIS is

BIC  #xxx,(offset)SP 

 I don't use anymore assembly but MSP has a beautiful clean orthogonal and effective instruction set. It resemble me old good day with motorola 68000 and PDP11 at early uP revolution age.

Katie Enderle

You can actually use the CCI bit to monitor the current state of the pin where you are currently outputting the timer-generated signal/PWM.

Are you sure about this? CCI reflects the state of the selected CCIA/CCIB input in capture mode. Which is not necessarily connected to the output signal in compare mode. And since it is synchronized to TACLK, it isn't necessarily showing the 'current' state.

However, I never used this myself (PxIN does the job fine on most MSPs) and the users guide isn't very "talkative" about this.

Hi Jens-Michael,

I've used the CCI bit in this way on the G2xx parts and on the F5529 when using the timer module in the app note SLAA513. So I've tried it out on these parts. The user's guide indicates that the CCI signals are connected even in compare mode and can be read at any time using the CCI bit, and that the input signal on CCI is latched into SCCI in compare mode when TAR counts to the value in TACCRX (see p. 371 section 12.2.4.2 of the 2xx user's guide) . In the app note (SLAA513) where we're using continuous mode on the timer to generate multiple different PWM frequencies at once. The CCI bit was used to determine if we are in the high or low portion of the PWM to determine the correct offset to add to TACCRx to generate the correct duty cycle.

Regards,

Katie

 Hi Katie, first thank a lot for all information on document, is not so different from what I done with MSP but is appreciated for all hint found.

 I used CCI and SCCI bit to capture input status when I wrote serial communication software, this function is invaluable to reduce external hardware but as I remember,  what I read again on user manual, as Jeans pointed to you  also to me this bit appear as is connected to Input selected by. So our friend has no output pin connected to Timer subsystem and from my knowledge I fear OUTMOD status cannot be read from internal. Did you know more about internal silicon implementation? On weekend I get the hardware and I try do some test on this bit.

I am running late with the results of my "hardware" tests, but I will post them in the next 48 hours.

I did one quick test on the OUT working. The minimalist code below runs on a LaunchPad, flashing

the green LED 50% at approx 1 second. (VLO). 

This uses the OUT of CCR1 (and no peripheral interrupts need to be specifically enabled)

I had a quick look at CCS v5 debug "register display" and it did not show the OUT bit or the CCI bit doing anything.

PIN 1.6 register did show the correct responses. I will double check and test more soon.

Still grateful to you all - james

                    .cdecls C,LIST,  "msp430g2553.h"
                    .text   

RESET              mov.w         #0280h,SP                                                    ;Initialize SP
                           mov.w         #WDTPW+WDTHOLD,&WDTCTL            ;Stop WDT

                           mov.w         #12000,&CCR0                                           ;set period
                           mov.w         #OUTMOD_3,&CCTL1                               ;OUT set/reset
                           mov.w         #00,&CCR1                                                  ;yields 50% duty cycle

                           bis.w          #040h,&P1SEL                                              ;set up P1.6 as second
                           bis.w          #040h,&P1DIR                                               ;output option TA0.1

                           mov.w        #TASSEL_1+MC_3,&TACTL                        ;ACLK,UpDownContinuous
                           bis.w          #LPM3+GIE,SR                                              ;LowPower & IE
                           nop                                                                                       ;For debugger


                         .sect   ".reset"                                                                       ; RESET Vector
                         .short  RESET
                         .end

ecstatist

                           bis.w          #040h,&P1SEL                                              ;set up P1.6 as second
                           bis.w          #040h,&P1DIR                                               ;output option TA0.1
                           bis.w          #LPM3+GIE,SR                                              ;LowPower & IE

All these registers are BYTE registers, so .B and not .W is to be used.
In case of P1xxx you're also programming P2xxx (this combo is identical to PAxxx)
And SR must not be written with a word at all. The behavior is undefined in this case.

ecstatist

This uses the OUT of CCR1 (and no peripheral interrupts need to be specifically enabled)

THis statement is ambiguous. OUT is a bit name. And it isn't used when you use OUTMOD_3. What you mean is TA0.1 output.

As I said, the OUT bit inside TACCTLx only changes its value if you manually change it in software. It is not changing its value based on any hardware action. So that's the expected and intended behavior.

the CCI bit, well does what i expected: nothing. P1.6 is TA0.1 output only and not available for CCI. It is P1.2 which has both, TA0.1 output and CCI1A capture input. It is stil not clear whether CCI is updated when in compare mode, as in compare mode, (obviously) no capture takes place.

Hi guys,

About the CCI bit (and the SCCI bit) - the CCI bit will change IF you are outputting the timer output on a pin - but the pin has to be configured to timer function using the PxSEL bits. When you select that pin to be configured for the timer, the CCI bit is then tied to that pin and will change with that pin (even in compare mode). I've tested this out on a G2553 and a launchpad with just a simple timer code example that toggles (OUTMOD_4)  P1.1 configured to timer output. if you put a breakpoint in the ISR, then you can see that CCI/SCCI bit in the corresponding TACCTLx register toggles every time (output was set to toggle mode). This is the same methodology that is used in my app note SLAA513 that I mentioned before - since we are outputting PWMs on pins, then we can use the CCI bit while we're in the ISR to see the current state of the pin.

See also 2xx user's guide section 12.2.4.2 Compare Mode, which talks about CCI behavior in Compare mode (not just capture mode):

2xx user's guide

12.2.4.2 Compare Mode

The compare mode is selected when CAP = 0. The compare mode is used to generate PWM output signals or interrupts at specific time intervals. When TAR counts to the value in a TACCRx:

• Interrupt flag CCIFG is set
• Internal signal EQUx = 1
• EQUx affects the output according to the output mode
• The input signal CCI is latched into SCCI

Regards,

Katie