Author Topic: PIR teardown  (Read 2786 times)


  • Administrator
  • Hero Member
  • *****
  • Posts: 1015
PIR teardown
« on: September 26, 2013, 04:07:10 PM »
Here is an old alarm passive sensor.

IMG_1296 by SeanB_ZA, on Flickr

IMG_1297 by SeanB_ZA, on Flickr

Removed from the case and the unit is fully shieled in a steel plate housing.

IMG_1298 by SeanB_ZA, on Flickr

Back of the unit, made in 1990 by Raca Guardall in Britain. I had already unsoldered the case at this point, had to pull out the big soldering gun to get enough heat in to melt the solder on those pins there.

IMG_1300 by SeanB_ZA, on Flickr

Bottom of the board, look at the solder mask complete with wrinkes, and the layout with all leaded parts.

IMG_1303 by SeanB_ZA, on Flickr

Top view, reed relay in the metal case, PIR sensor in the middle and the custom controller chip on the top of the view.

IMG_1306 by SeanB_ZA, on Flickr

Good capacitors used there, still good after 2 decades plus of use.

This was replaced as part of fault finding, as the alarm was false triggering intermittently. As the zone was a shared zone I initially changed the sensor as they are known to fail after a long time. This did not solve the issue, so I next exonerated the control panel by swapping the 2 zones with the next zone, and the intermittent moved with the swap. Next I swapped the high zones back, and this moved the fault to the other circuit, proving it was not the PIR but the lower zone wiring, or the door reed switch attached to it. Disconnected the switch and shorted the wiring at the door side, and then it was not false tripping, so the switch was going intermittently high resistance, leading the controller to think the high zone was triggering. The zones are distinguished by differing resistors in the wiring, the low using a 3k74 and the high side uses a 6k98 resistor, all being 1% resistors. When I went to test the zone switch after replacing it it had finally gone open circuit, probably having corroded internally inside the potting compound. Annoying in that it almost always faulted at 5AM.