Re: [Paparazzi-devel] Critical behaviour of baro_ms5611_i2c

[Felix Ruess]

Hi,

we add bounds like that, but we should probably discard the measurement if this happens and not simply leave the old value...

Could you plz make a pull request for this?

Cheers, Felix

On Mon, Jun 23, 2014 at 5:13 AM, afishindouban <address@hidden> wrote:

Hi Felix

I can not find the reason, I think it is the bug of MS5611, maybe its read
when the data is not ready. Anyway simple add a checker in the  ms5611.c
<https://github.com/paparazzi/paparazzi/blob/master/sw/airborne/peripherals/ms5611.c>
solve the problem.  the checker is reasonable and works in any case, the
range is from  MS5611 document
<http://www.meas-spec.com/downloads/MS5611-01BA03.pdf>  .

So would you like to include it into master?
here is the modified ms5611_calc()

void ms5611_calc(struct Ms5611Data* ms) {
  int64_t dt, tempms, off, sens, t2, off2, sens2, p;

  /* difference between actual and ref temperature */
  dt = ms->d2 - (int64_t)ms->c[5] * (1<<8);
  /* actual temperature */
  tempms = 2000 + ((int64_t)dt * ms->c[6]) / (1<<23);
  /* offset at actual temperature */
  off = ((int64_t)ms->c[2] * (1<<16)) + ((int64_t)ms->c[4] * dt) / (1<<7);
  /* sensitivity at actual temperature */
  sens = ((int64_t)ms->c[1] * (1<<15)) + ((int64_t)ms->c[3] * dt) / (1<<8);
  /* second order temperature compensation */
  if (tempms < 2000) {
    t2 = (dt*dt) / (1<<31);
    off2 = 5 * ((int64_t)(tempms-2000)*(tempms-2000)) / (1<<1);
    sens2 = 5 * ((int64_t)(tempms-2000)*(tempms-2000)) / (1<<2);
    if (tempms < -1500) {
      off2 = off2 + 7 * (int64_t)(tempms+1500)*(tempms+1500);
      sens2 = sens2 + 11 * ((int64_t)(tempms+1500)*(tempms+1500)) / (1<<1);
    }
    tempms = tempms - t2;
    off = off - off2;
    sens = sens - sens2;
  }

  p = (((int64_t)ms->d1 * sens) / (1<<21) - off) / (1<<15);
  if ((tempms > -4000) && (tempms < 8500) && (p > 1000 ) && (p < 120000)) {
    /* temperature in deg Celsius with 0.01 degC resolultion */
    ms->temperature = (int32_t)tempms;
    /* temperature compensated pressure in Pascal (0.01mbar) */
    ms->pressure = (uint32_t)p;
  }
}

And here is the test result.

<http://lists.paparazziuav.org/file/n15720/47.png>



flixr wrote

> Hi,
>
> Would be great if you could dig a bit further! Add
> <define name="DEBUG"
> value="TRUE"/>
>  to your firmware section and have a look at the BARO_MS5611
> message that is sent with every measurement from the baro board wrapper

> &lt;https://github.com/paparazzi/paparazzi/blob/master/sw/airborne/boards/baro_board_ms5611_spi.c&gt;.

> That should help to narrow down the problem, e.g. if D1 or D2 is suddenly
> very large.
>
> If you want to add additional checks to the ms5611 peripheral, add them in
> sw/airborne/peripherals/ms5611_spi.c

> &lt;https://github.com/paparazzi/paparazzi/blob/master/sw/airborne/peripherals/ms5611_spi.c&gt;
>
> Cheers, Felix





--
View this message in context: http://lists.paparazziuav.org/Critical-behaviour-of-baro-ms5611-i2c-tp11592p15720.html

Sent from the paparazzi-devel mailing list archive at Nabble.com.

_______________________________________________
Paparazzi-devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/paparazzi-devel