• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 40 subscribers
  • Views 578 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.

TIDA-00489: How fast is the sensor setup?

Slev1n
Genius 5680 points
Part Number: TIDA-00489

Hey guys,

reading the TIDA-00489 PDF I was wondering, if the sensor layout is able to detect to several movements in a few seconds e.g. 1 movement per second. I think in my application 2 movements per second would be sufficient, but still, I'd like to know what the maximum is and if I can alter the electronic layout to be faster. In my later application, the system would sleep until a movement is noticed by the sensor, wake up --> increment a movement counter and immediately go back to sleep. After 10 mins or if 10 movements have been detected, the amount of movements is send to a gateway.

  1. So, how fast is the device in standard configuration?
  2. Can I adapt the filter parameters to get faster?

I hope I was clear about my question, if not, dont hesitate to ask for further information.

Kind regards


Stefan

  • 0
    TI__Mastermind 25311 points
    Hi Stefan,

    Thanks for considering TIDA-00489 for your next design. We are reviewing your questions and will get back to you in a day time frame.

    Gus Martinez
  • 0
    TI__Mastermind 25311 points
    Stefan,

    Both questions are tricky to answer because of variables such as distance from sensor, size of object, speed of movements, etc. However, from a circuit level point of view, the limiting factor will be the time constants of the filter circuit.

    The design guide shows the filter bandwidth is set to ~1Hz to ~10Hz to cover human movement. You could possibly shift this bandwidth to higher frequencies to lower the time constants but you may start missing some motions depending on the variables above. You would also need to keep the bandwidth roughly the same, otherwise the poles and zeroes in the response will get too close and will start to limit the peak gain which will impact sensitivity. Raising the upper cutoff frequency will increase the output noise and make the design susceptible to false triggers.

    One alternative might be to leave the filter alone and check into the possibility of sampling the comparator outputs to establish a count since continuous motion will result in a relatively continuous high on the OR of the two comparator outputs. To do something like this you would still need to set some boundary conditions on the variables above.

    Another alternative would be to sample the output of the amplifiers directly (bypass the comparator) using an ADC such that the sensor signal can be digitized and analyzed more elegantly to reliably count the number of motion events.
  • 0 in reply to Gus Martinez
    Genius 5680 points

    Hallo Guz,

    thank you for your answer. So looking at a bandpassfilter I would calculate the time constant (tau) of the low pass filter part (e.g. 10 Hz). Since I am interested in the rise time (t_r), I use the relation: t_r=2.2 * tau. For 10 Hz as f_-3dB_LPF I come to 0,035s.

    Well, I am not sure how to use the cut of frequency of the high pass filter, but I asume, that I could use it for the settling time (t_s) of the filter response in the same manner as outlined before. Having a f_-3dB_HPF of 1 Hz I come to 0,35s.

    Please correct me, if my humble interpretation of the time domain is wrong, but I think, the sensor should be "settled" after an impulse (or movement) taking t_s + t_r = 0,385s , right?

    If this is right, I would not have to change anything. Please let me know your opinion on this. If necessary I could run some simulations on this, too.

    Kind regards

    Stefan

  • 0 in reply to Slev1n
    Genius 5680 points

    I have made some simulations using "TI WEBENCH". Find them attached. It looks like that the analog signal after the filter system should have settled again after ca. 600ms. Thus, this should not limit me.