euler angles become NaN after some time

[till]

for testing i leave the UM7 (firmware U71D) running for extended periods of time. during these tests the UM7 is on a flat surface with no movement.

from time to time i receive DREG_HEALTH packets where ACC_N is set. mostly the device seems to continue working correctly after that. this seems strange and i have no idea where it might be coming from.

every morning when i come back to the office though, the euler angles (as reported in DREG_EULER_PHI_THETA, DREG_EULER_PSI) have become 0, DREG_GYRO_PROC_* AND DREG_ACCEL_PROC_* report NaN, while DREG_MAG_PROC_* continues to work as expected. last night’s log shows a health packet change with bad information at device time 32768 (oddly the limit of a signed short int, will check if this is reproducible):

    health   roll   pitch     yaw   xmagraw   ymagraw   zmagraw   yawrate   xaccel   yaccel   zaccel magrawtim
0000010001   0.00    0.00    0.00    -36.00     -2.75     81.25      0.00      nan      nan      nan  32768.28

[till]

last drop to 0 just happened as i moved the UM7, bumping it lightly. it can normally sustain much heavier bumps with no problems:

    health   roll     pitch       yaw   xmagraw   ymagraw   zmagraw   yawrate    xaccel    yaccel    zaccel magrawtim acceltime   heading
0000000001  -6.23     10.42     44.15     -0.47     -0.41      0.53     -3.56      0.18      0.10     -0.94  19357.88  19357.91    137.46
0000000001  -5.97      9.53     44.08     -0.46     -0.40      0.54      0.38      0.14      0.10     -0.96  19357.91  19357.96    137.46
0000000001  -5.55      8.00     43.97     -0.45     -0.40      0.55     -2.75      0.10      0.10     -0.96  19357.96  19358.01    137.46
0000000001  -5.02      5.72     43.74     -0.43     -0.40      0.56     -5.19       nan       nan       nan  19358.04  19358.06    137.46
0000010001   0.00      0.00      0.00     -0.36     -0.38      0.60      0.00       nan       nan       nan  19358.20  19358.21    133.65

was the device time above just a strange coincidence?

[till]

now my first impression looks right again. this is a log from last night:

    health  roll  pitch     yaw xmagraw ymagraw zmagraw yawrate  xaccel  yaccel  zaccel magrawtim acceltime  heading
0000000001 -0.20   0.12   34.07   -1.00  -31.00  -25.00    0.00    0.00    0.01   -0.98  32767.06  32767.10   122.86
0000000001 -0.20   0.12   34.07   -6.00  -30.00  -26.00   -0.06    0.00    0.00   -0.98  32767.11  32767.15   122.86
0000000001 -0.20   0.12   34.07   -6.00  -22.00  -28.00    0.00    0.00    0.00   -0.98  32767.19  32767.20   122.86
0000000001 -0.20   0.13   34.06   -8.00  -30.00  -32.00    0.00    0.00   -0.00   -0.98  32767.23  32767.26   122.86
0000000001 -0.20   0.13   34.06   -4.00  -34.00  -26.00    0.00    0.01    0.00   -0.98  32767.27  32767.31   122.86
0000000001 -0.20   0.13   34.06   -1.00  -27.00  -19.00    0.00   -0.00    0.00   -0.98  32767.31  32767.36   122.86
0000000001 -0.20   0.13   34.07   -4.00  -30.00  -26.00    0.00    0.00    0.00   -0.98  32767.39  32767.41   122.86
0000000001 -0.20   0.12   34.07   -4.00  -34.00  -22.00    0.00    0.00    0.01   -0.98  32767.43  32767.46   122.86
0000000001 -0.20   0.12   34.07    1.00  -31.00  -23.00    0.00    0.00   -0.00   -0.98  32767.47  32767.51   122.86
0000000001 -0.20   0.12   34.07   -4.00  -34.00  -22.00    0.06    0.00    0.00   -0.98  32767.55  32767.57   122.86
0000000001 -0.20   0.12   34.07   -3.00  -35.00  -23.00    0.00    0.00    0.00   -0.98  32767.59  32767.62   122.86
0000000001 -0.20   0.12   34.07   -6.00  -30.00  -24.00    0.00    0.00    0.00   -0.98  32767.63  32767.68   122.86
0000000001 -0.20   0.12   34.07    4.00  -34.00  -26.00   -0.06    0.00    0.00   -0.98  32767.67  32767.72   122.86
0000000001 -0.20   0.12   34.07   -3.00  -31.00  -25.00    0.00    0.00    0.00   -0.98  32767.75  32767.78   122.86
0000000001 -0.20   0.12   34.07    0.00  -26.00  -18.00    0.00    0.00    0.00   -0.98  32767.79  32767.83   122.86
0000000001 -0.20   0.12   34.07   -7.00  -27.00  -25.00    0.00    0.00    0.00   -0.98  32767.83  32767.88   122.86
0000000001 -0.20   0.12   34.07   -3.00  -27.00  -27.00    0.00   -0.00    0.00   -0.98  32767.92  32767.93   122.86
0000000001 -0.20   0.12   34.07   -3.00  -31.00  -21.00    0.00    0.00    0.00   -0.97  32767.95  32767.99   122.86
0000000001 -0.20   0.12   34.07   -6.00  -32.00  -20.00    0.00    0.00    0.01   -0.97  32768.00  32768.04   122.86
0000000001 -0.20   0.12   34.07   -1.00  -33.00  -21.00    0.06     nan     nan     nan  32768.04  32768.09   122.86
0000010001  0.00   0.00    0.00    8.00  -30.00  -24.00    0.00     nan     nan     nan  32768.36  32768.39   123.22

note how health shows ACC_N bit set.
regards

-till

[till]

i’m very confident now there must be a bug with short int overflow …

    health   roll   pitch      yaw xmagraw ymagraw zmagraw yawrate xaccel yaccel zaccel magrawtim acceltime  heading
0000000001  -0.59   -0.86  -311.30   -4.00  -20.00  -26.00    0.00  -0.02   0.01  -0.98  32767.63  32767.67   136.63
0000000001  -0.59   -0.86  -311.30   -7.00  -19.00  -27.00    0.06  -0.02   0.01  -0.98  32767.67  32767.72   136.63
0000000001  -0.59   -0.86  -311.30  -10.00  -16.00  -26.00    0.00  -0.01   0.01  -0.98  32767.75  32767.78   136.63
0000000001  -0.60   -0.86  -311.30  -10.00  -14.00  -22.00    0.06  -0.02   0.01  -0.98  32767.79  32767.83   136.63
0000000001  -0.60   -0.86  -311.30   -8.00  -20.00  -22.00    0.00  -0.01   0.01  -0.98  32767.83  32767.88   136.63
0000000001  -0.60   -0.86  -311.30  -10.00  -16.00  -24.00    0.00  -0.02   0.01  -0.98  32767.87  32767.93   136.63
0000000001  -0.59   -0.86  -311.30  -12.00  -20.00  -28.00    0.00  -0.02   0.01  -0.98  32767.96  32767.98   136.63
0000000001  -0.59   -0.86  -311.30   -4.00  -24.00  -22.00    0.00  -0.02   0.02  -0.98  32767.99  32768.03   136.63
0000000001  -0.59   -0.86  -311.30  -11.00  -15.00  -27.00    0.00  -0.01   0.01  -0.98  32768.04  32768.08   136.63
0000000001  -0.59   -0.86  -311.30   -7.00  -17.00  -27.00    0.00  -0.01   0.01  -0.99  32768.08  32768.13   136.63
0000000001  -0.59   -0.86  -311.30  -11.00  -15.00  -23.00    0.00  -0.02   0.01  -0.98  32768.16  32768.18   136.63
0000000001  -0.59   -0.86  -311.30  -10.00  -20.00  -28.00    0.00  -0.01   0.01  -0.98  32768.20  32768.23   136.63
0000000001  -0.59   -0.86  -311.30  -15.00  -15.00  -25.00    0.00  -0.02   0.01  -0.98  32768.24  32768.29   136.63
0000000001  -0.59   -0.86  -311.30   -9.00  -15.00  -23.00    0.00  -0.01   0.01  -0.99  32768.28  32768.34   136.63
0000000001  -0.59   -0.86  -311.30  -10.00  -22.00  -20.00    0.00  -0.02   0.01  -0.98  32768.36  32768.39   136.63
0000000001  -0.60   -0.86  -311.30  -10.00  -22.00  -20.00    0.00  -0.02   0.01  -0.97  32768.40  32768.44   136.63
0000000001  -0.59   -0.86  -311.30  -10.00  -18.00  -24.00    0.00  -0.02   0.01  -0.98  32768.44  32768.49   136.63
0000000001  -0.59   -0.86  -311.30  -11.00  -17.00  -27.00    0.00  -0.02   0.01  -0.98  32768.48  32768.54   136.63
0000000001  -0.59   -0.86  -311.30  -12.00  -18.00  -24.00    0.00  -0.02   0.01  -0.98  32768.56  32768.59   136.63
0000000001  -0.59   -0.86  -311.30   -8.00  -22.00  -22.00    0.00    nan    nan    nan  32768.61  32768.64   136.63
0000010001   0.00    0.00     0.00   -6.00  -18.00  -22.00    0.00    nan    nan    nan  32769.01  32769.05   135.38

[Michael]

Hi Till,
Thanks for this update it’s greatly appreciated.

Can you please contact the support team at Redshift labs.
They would like to discuss this further and possibly gather some more information from you.

Regards,
Michael Malloy

[Author]

Posts