• State TI Thinks Resolved
  • Locked Locked
  • Replies 2 replies
  • Answers 1 answer
  • Subscribers 40 subscribers
  • Views 622 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: PIR Motion Sensor Flutter

Jay Yamaha
Prodigy 10 points
Part Number: TIDA-00489

I am using the TIDA-00489 reference circuit as the basis for my design using the same BOM.   The Murata PIR detector part does have a lens installed.  Unlike another user who did not see any activity without the lens installed, I do see activity regardless of the lens installation.  "Activity" means that I see transitions on the PIR_OUT_HI and PIR_OUT_LO.   When I walk back, I see the last stage of analog (  R19) is moving between the voltage extremes.   The comparator stage looks like it is working fine and outputting a valid signal for what's fed to it.

I have used a "fresh" PIR sensor that has always been covered with a lens - so it has no scratches or fingerprints - with similar results.  Power integrity is key - this design runs off of 2 AA batteries (3V) which also powers other circuitry.  I have ensured that the other circuitry is OFF to avoid any crosstalk issues or power dips (power integrity) due to "high" current demands of other external circuitry.

The purpose of the PIR is to "wake up" the rest of the circuitry - so none of that other circuitry is active when the PIR is operational.  The problem is the PIR is constantly "firing off" - so I  synthetically keep the rest of the circuitry from waking up when PIR_OUT_[HI,LO] transition (ignore transitions).

I see some periodic (not at any given regular period) high frequency noise on the PIR Vin side.  I am going to attempt to correlate this with the PIR's amplified output (low frequency swings back and forth) to see if these short bursts of noise correlate with the movement on PIR or not.  In order to see these, I set the vertical (Voltage) offset to center on Vin of the PIR and then turn the scale way up to make the noise more evident.

I am wondering if there were ever issues with the "high impedance" Vin to the PIR.  A series 619k as part of an RC filter is different from Murata's reference ( ) which uses a 1ohm series off their regulator (w/22uF cap).  Experimentally, I could shotgun a lower R value for the power series ("pull up" of the RC filter) - but I was wondering if this has been seen as a general problematic area.  My assumption is that the circuit is low current enough that the R value scales according to the current consumption of the TI parts vs. the Murata design (NJR parts).

  • 0
    TI__Prodigy 600 points
    Jay,

    We are looking into your questions from your post. I should have a reply for you within the next 12 hours.
  • 0
    TI__Prodigy 600 points
    Sorry for the delay in replying to your post.

    I agree with your debug steps. We have seen higher noise on some sensors than others with the high gain in the amplifier stages, due to the low sensor bias current, which leads to false triggers. However it seems like what you're seeing is noise being coupled in to the supply pin of the sensor. Because the sensor output is a JFET transistor configured as a voltage follower, any noise present at the supply pin of the sensor will show up at the sensor output. The question then becomes, what is the spectral content of this noise and how much of it lies within the bandwidth of the following filter stages? You mention that you are using 2 AA batteries as the power source. How long are the wires connecting the battery holder to the PIR sensor board? These could be picking up noise from fluorescent lighting or nearby cellphones or PC's if the batteries are in fact wired to the PCB as opposed to being part of the PCB assembly as in the case of the TIDA-00489 reference design.

    We have not seen any issues with the 619k resistor as it relates to what you're seeing. In the design, R4 sets the bias current for the sensor. This current will flow through R2 causing a voltage drop which ultimately limits the minimum battery voltage since the sensor needs 2v minimum to operate. The voltage drop across R2 sets a maximum value for its resistance available for filtering based on the minimum battery voltage for the application. Because we biased the sensor at such a low current to achieve long battery life, we were able to increase the value of R2 to 619k as a fair tradeoff. In fact, if you notice in Murata's reference document you attached, the ref circuit at the bottom of page 4 shows a 33k resistor in series with the supply pin with a 47k bias resistor to ground.

    Using an LDO for supply regulation will generally reduce the filtering requirements into the sensor pin, especially if the circuit noise on the supply rail falls within the bandwidth of the regulator. This is why you see 1 ohm instead of 33k when the regulator is included in Murata's reference document.