LoPy 4 in a factory

[robert-hh]

@william There was no file attached. But besides that, the picture looks bad, since also a part of the printout message header, the trailing "===========" is either overwritten or not printed at all. And also that data is way too long for a LoRa message. So it looks like an internal bug. Does this happend spontaneous or as result of a message being sent (and received).

[William]

Hi Robert:

Sorry, I can't upload txt file, but it just like what you see.
I found when it received, I guess it can't printed all of message because it's crash.
It's possible happen from signal?

Otherwise, I sometime can see error message "buffer too small"
I already use 1024 to send and received message.
Or It's a stupid idea.

Thanks,

[robert-hh]

@william Lora messages can only be up to 251 byte long. I do not think it is a buffer size question in your code.

[jcaron]

I think you'll have to share the relevant parts of your code, it's quite difficult to guess from your logs what you are doing and what could happen.

Note that it is perfectly possible you could sometimes pick up LoRa messages from other sources, so you need to make sure your code is robust enough to recognise your own messages only, and behave correctly if it receives a message which is not one of yours or is incorrectly formatted...

[William]

@jcaron

Node:
1_1542268619192_main.py
0_1542268619191_boot.py

Gateway:
1_1542268637223_main.py
0_1542268637223_boot.py

I upload the files.
Thanks,
William

[William]

I don't know this is a problem or not.
Those lopy place are nearby "High frequency" machines.
Does is make a messy message?

Thanks,

[robert-hh]

@william What do you call a "hight frequency" machine? Besides that, I would place the Lopy in a well shielded metal box, and add filters to the power supply lines. You may also consider to put the Lopy power supply inside the shielded box.

[William]

@robert-hh
I mean "High-frequency induction hardening", Sorry my bad English....

[William]

@robert-hh
If I don't have any shielded boxes.
This will cause message messy?

[crumble]

@william said in LoPy 4 in a foctory:

This will cause message messy?

Do you mean the LoRa messages? Be aware that Lora is no point2point connecteion. You revceive all messages from your selected channel. So you will get messages from other LoRa users as well. It works like a CB radio or PMR, but with more rules.

You can choose between two solutions:

• Design a protocol, so that you can detect your messages and ignore all other.
• Join a network like TTN. You will get the protocol for free and may benefit from the long range (of the internet).

[robert-hh]

@william The dump below does not look like a noisy LoRa message. It looks more like the LoPy crashed because of other interference, spikes on the supply, ....
A closeby high frequency, high energy induction machine is definitely not a friendly environment for unshielded electronics.

[William]

@crumble Design a protocol? Do not use LoRa?
Sorry I don't that mean.
Could you tall me more?

Thanks,

[William]

@robert-hh
Okay, I need to make a shielded box to protect power supply, right?
I thought that was happen because the antenna, but it wrong.
Thank you.

[robert-hh]

@william Not only for the power supply, but all other interferences into the circuit. The antenna must be assembled into outside the box' wall, and with that box, it will also work better.

[William]

@robert-hh, Thank you. But I still have few questions.

1. In lopy, there have "Power Management IC" and "AC to DC".
   These can't filter interference?

2. I have set watchdog in lopy script.
   It's not work. why?

Thanks,

[robert-hh]

@william About topic 1: I do not know that the LoPy has any of that by itself. There is a 3.3V power regulator, but thta cannot handle string interference, which may com in at each and every signal line.
Topic 2: The watchdog is software, can handle some run-away software situation. You might need a separate hardware watchdog which pulls reset.