PIR started reporting motion every 20 minutes [SOLVED]

[Strixx]

Hello,

I have been testing my latest node for a couple weeks. It's a node with a double relay, a DHT22, one reed switch and one motion sensor (HC-SR501), with repeater function. It is intended for maneuvering my garage door. And should be situated in my garage (of course) and the reason for repeater function is that I also have a temp node at my fridge in the garage that some times is not reaching my gateway. This new node is placed halfway between the gateway and fridge node.

It has been standing in the garage at the same distance as final mounting place and I have been logging the function. And all have been working perfectly.
But yesterday I mounted it and connected the garage door to the relay and extended the reed wires (4 meters), and now the motion sensor (not changed at all) is reporting motion every 20 minutes.

So last night I found this thread:
New nodes, new problems Motion sensor [Solved]
wich reports the exact same behavior.

But @xydix solved the problem with replacing the PIR from another model to the one I am using. Others suggest it is a power problem.
So I thought it was due to the reed switch. That was not tested at "closed" for a longer period under testing. Drawing current when the switch was closed.
But I disconnected it again, and no change. Still reporting motion every 20 minutes.

So:
Any one having any ideas what I should do next?
The only remaining differences from testing period is location. I was standing on a shelf before, and now it is placed in the sealing, with the same distance to the gateway (about 7 meters)

And if it is power issue due to radio radio transmissions (as suggested in the earlier mentioned thread) why is it only on discover request broadcast, and not when the node is sending data from DHT22 or reed, or when I send data to the node for maneuvering the relays?

[ahmedadelhosni]

@strixx actually a have a similar problem but my OpenHab sends random notifications maybe every 2 days that motion was detected but i didnt debug yet. Whatever .. do you power your PIR with 5v ? With 3.3v for sure this will happen.

[electrik]

You could try to add a big capacitor on the power line (1mF). Maybe it is overdone but then you at least know if it is caused by transmitting of messages.
Did you disconnect the long wires completely? Did you decouple the wires with 10n capacitors?

[Strixx]

@ahmedadelhosni Yes. This is a 5V sensor powered from a phone charger.

[Strixx]

@electrik I disconnected the extender wires for reed. And it has been running all night like this. But no difference. So it can not be due to the extender wires.

[Yveaux]

@strixx Power is everything with the HC-SR501. Small fluctuations can cause false triggers. Fluctuations can e.g. be cause by radio traffic, bad/unsifficient power supply, decoupling etc. And even with a good power supply (3xAA batteries) I've found these sensors to sometimes produce bursts of false triggers.
I'm afraid you get what you pay for in this case.

[Strixx]

@yveaux Yes. I have understood that by reading up. But what I can't understand is that it not happens every time I use the radio. As it seems now it is only when the gateway is sending the discovery broadcast. Not when the node is sending data from the other sensors on the node, or when I send data to the node for operating the relays. Strange...

[electrik]

@strixx maybe the amount of packages/messages makes the difference?

[Strixx]

This is strange...

Last 24 hours I have had the node standing almost next to the gateway, and no false motion has been detected. So:

• Close to the gateway there is no false indication.
• Testing area in garage is no false indication.
• Final place in garage there is false indication.

PIR sensor is driven from 5V USB charger. All sensors are connected to the 5V charger as well as the Pro Mini. The radio is NRF24L01+PA+LNA connected to a AMS1117 5V-3.3V converter.
USB charger is capable to deliver 700mA. And if I calculate correct the node has a maximum of about 270 mA consumption.
So there should be more than enough power.

And why is it different due to location of the node? Does power consumption differs from radio depending on the quality of the radio connection to the gateway?

[Yveaux]

@strixx said in PIR started reporting motion every 20 minutes:

Does power consumption differs from radio depending on the quality of the radio connection to the gateway?

Yes, bad connection causes retransmits, eating more power.

[Strixx]

Well. It was the radio that caused this false triggers. The final placement of the node must be in a radio shaddow.

I really "funny" thing is that I solved this problem by changing the radio. I had a NRF24L01+PA+LNA for better range. This was for future project (node in my greenhouse) which would need this node as a repeater for shore.
But by replacing the radio to a "normal" NRF24L01+ made the problem go away. I did not change the location!