• State Answer Suggested
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 78 subscribers
  • Views 4439 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

MSP430g2553- use of Timer-Capture and Port interrupt (Nested interrupts)

Vasuki Shankar
Prodigy 250 points
Other Parts Discussed in Thread: MSP430G2553, TDC7200

Hi,

I am building an Ultrasonic fluid flow meter and am using MSP430G2553 for the velocity measurements. The outputs from pair the sensors is digitized and given to the msp. I need to find the Time Interval between the rising edge of first sensor (Sensor A) to the rising edge of the second Sensor B  (during Along and Against the liquid flow).

I have used PORT 1.1 for the input of first pulse and the capture mode for detecting the Second pulse, given to PORT 1.2.

NOTE: the timing difference is in micro-seconds 

I have written this code, please provide insight about it

-------------------------------------------------------------------------


#include <msp430g2553.h>
int i,j=0,k=0;
int T[100]=0;

void main(void)
{
int delta_t, velocity;

WDTCTL = WDTPW + WDTHOLD;                                      // Stop watchdog timer
P1DIR = BIT0+BIT6;                                                              // green and red as outputs
P1IE |= BIT1;                                                                          // P1.1 interrupt enabled
P1IES &= ~ BIT1;                                                                   //lo to hi
P1IFG &= ~BIT1;

_BIS_SR(GIE);                                                                        //enabling interrupts

while(1)
{

WDTCTL = WDTPW + WDTHOLD;                                               // resetting it to be on the safer side
 

for(i=0;i<099;i=i+2)
   for(j=i+1;j<100; j= j+2)
{
 delta_t= T[i]-T[j];                                                                    // difference between Along and Against the flow
 velocity= delta_t * 1482*1482/ (2*0.03*1000000);               // to calc velocity
display(velocity);
}
}

}

// Port 1 ISR to detect the first rising pulse from Sensor-A

#pragma vector=PORT1_VECTOR
__interrupt void Port_1(void)
{


TACTL=TASSEL_2 + MC_2 + TAIE_0;                                   //SMCLK, continuous mode, disable timer interrupt  
TACCTL0= CM_1 + CCIS_0 + SCS_1 + CAP_1 + CCIE_1 + COV_0;       // Capture to be taken at BIT2 (Sensor-B), rising edge
_BIS_SR(GIE);

P1SEL |= BIT2; //making bit 2 capture input
P1IFG &= ~BIT1;

}

// ISR when CAPTURE occurs

#pragma vector=TIMER0_A0_VECTOR
__interrupt void Timer_A (void)
{
k++;
while(k<100)                                                              // to update the timer values in an array 
{
T[k]= TA0CCR0;                                                        // time diff between pulses for 1 direction of fluid flow
TACTL=MC_0+ TACLR_1 ;                                      //clear timer, for the concern of timer interrupts

}

--------------------------------------------------------------------------------------

Questions

1. Should I reset the watchdog timer in the infinite while loop also?

2. What is the clock frequency of SMCLK? 

3. Is ok to raise another interrupt inside the port interrupt?

4. Most importantly, WILL THE PROGRAMME WORK?

  • 0
    Guru 74080 points

    Vasuki,

    you do not reset your watchdog, you are disabling it. When you have disabled it right after startup, it remains disabled. No need for a second time. Same for GIE. The clock frequency after startup is something around 1MHz. If you want a more precise calibrated frequency, you should use the factory programmed frequency settings. Insert these two lines after disabling the watchdog to set the DCO to the desired frequency:

    BCSCTL1 = CALBC1_1MHZ;
DCOCTL  = CALDCO_1MHZ;

    This will give you 1MHz clock speed. You can also choose 8, 12 or 16MHz - just replace the '1' in _1MHZ (both lines) by the value you want. Concerning interrupts in interrupts...at least I think you can enable the GIE bit in a running interrupt again. But there is no need for that at all. If you think you could "loose" interrupts, first speed up your processor. And look if your ISRs are short and contain only the most important. Don't do complex calculations in there.

    Does your program work? I don't know. Test it :-) I haven't looked at your program flow now. Next time when posting code you should use the syntax higlighter. It is easier to read then.

    Dennis

  • 0
    Guru 27285 points

    Vasuki Shankar said:
    I am building an Ultrasonic fluid flow meter and am using MSP430G2553 for the velocity measurements. The outputs from pair the sensors is digitized and given to the msp. I need to find the Time Interval between the rising edge of first sensor (Sensor A) to the rising edge of the second Sensor B  (during Along and Against the liquid flow).

     I think this processor is powerful enough but why not use specialized one for Flow metering like 430FG series or a faster of series 55?

     Are you using a frontend to flow  meter? Is same are you asking for 550KHz phase detection generally used in flow metering?

  • 0 in reply to Roberto Romano
    Prodigy 250 points
    Thanks Roberto for the support,

    I am not aware of the 430FG series. Yes the 550kHz phase detection is for flow measurement
  • 0 in reply to Vasuki Shankar
    Guru 27285 points

    Vasuki Shankar said:
    Yes the 550kHz phase detection is for flow measurement

     Are you using a front end to measure fly time? MSP timer resolution is not enough, all depend of precision you need and what is your application. Generally Flow is also related to temperature to measure heat energy flow..

     This type of applycation in a short chamber with sound speed near of 1Km/s require resolution in range of nS so have you seen some front end? TI announced but I am not sure it is available.. A better have it integrated on MSP die as for AFE series and all specialized. For now just flow for simple wheel detection seems available, ultrasound you have browse competitor.

    www.ti.com/.../application_and_implementation;tisearch=Search-EN-Everything

    and also tdc7200 was released so

    www.ti.com/.../TDC7200;tisearch=Search-EN-Everything

    with these two chip you can measure pS resolution necessary to evaluate flow. Another way is to use C2000 or DSP series processor to estimate fly time and also phase if necessary.

  • 0 in reply to Vasuki Shankar
    Guru 27285 points

    Vasuki Shankar said:
    Yes the 550kHz phase detection is for flow measurement

    www.ti.com/.../tdc7200evm

     This is better than waste time to community for a out of order measurement, better measurement can be done with Timer D MSP430 is near 4nS you need pS resolution, so switch to appropriate hardware...

  • 0 in reply to Vasuki Shankar
    Guru 27285 points

    Vasuki Shankar said:
    for flow measurement

     This is competitor chips, I fear this has no ready software and I don't know about price, both TI evaluation board are really affordable but you need 5529 launchpad, G series is not appropriate or better it is but you need interfaces to transfer data and also some math power to do calculation involved with flow integration.

    www.maximintegrated.com/.../tb_tab0

  • 0
    Prodigy 250 points
    Thank you for the reply,

    I could implement the transit time measurement using msp430g2553 itself (by configuring the sensor outputs as port interrupts and finding the time between them). I did changes to the flow meter setup and increased the transit time to around 1.3 micro secs. I also ran the msp at 23MHz (DCO_5 and BSCTL_15).
    But the results are not very accurate.
    As far as I know theres no MSP device with a higher clock rate right? (greater than 25MHz), also how better will it be to use Assembly instead of CC Studio in terms of machine cycle optimization, so that I can increase the resolution upto the machine cycle interval of 40ns.

    I started working on TDC-GP22, which is dedicated for flow measurements. That will provide very high accuracy and resolution.
  • 0 in reply to Vasuki Shankar
    Guru 227245 points
    Using a port interrupt to measure a timing isn't the best idea. Due to interrupt latency, you likely won't get stable results.
    Better let the timer run as fast as possible, then program a separate capture interrupt for both signals. Only trigger an interrupt at the second signal's capture and then simply subtract the two captured values to get the difference. As long as executing your ISR doesn't take longer than the distance to the next edge pair, this should give you the maximum possible precision or ~1 timer tick. But probably a resolution of 40 ns (25MHz) isn't good enough.
    There are MSPs with a high-speed TimerD module which go a bit higher, but probably an external counter/latch hardware or a dedicated chip is better. Or some circuitry that turns the delay into an analog voltage which can be measured ab an ADC then. Like a current source that starts charging a capacitor at the first edge and stops with the second, resulting in a voltage the is directly proportional to the delay between the two signals.

**Attention** This is a public forum