RAK7240 / RAK7258 LTE boot with WiFi AP/LP PC Bug

LPWAN Industrial Gateways / Concentrators
1

Hi Everyone,

I have potentially found a bit a strange bug relating to both RAK7240 and RAK7258 have similar issues.
RAK7240 running 1.0.2 and older (not tested on newest version however)

Another bug also listed below relating to RSSI reporting.

a) WiFi set as Access Point mode on the RAK7240 - e.g. broadcasting a WiFi allowing people onsite to take advantage of the SIM card as a hot-spot essentially while still allowing the Gateway to connect back to the internet.

b) LTE setup is all correct and proven to work, connected and working well with WiFi giving access to the laptop computer currently connected to broadcast WiFi from said RAK device.

Bug:
When rebooting the Gateway, OR powering down/reconnecting at site I have found that the Gateway will refuse to use the LTE as primary backhaul even when logging into the online console and viewing its cellular status as OK.

This is issue goes away if I first remember to disconnect a Windows 10 laptop from the RAK gateways broadcast WiFi WHILE the gateway boots fully, if I wait a few minutes assuming the RAK gateway has fully booted, then turn on the computer WiFi again the LTE has connected and allowing the gateway to connect to the internet perfectly.

It seams there is a bug with the WiFI being used as backhaul instead of LTE - this would make sense in the case of the gateway connecting to existing router - but appears to use the Laptop in hopes of allowing internet access through its adapter - which is wrong for Access Point mode.

Be interesting to hear feedback of similar issues and have a resolution to it.

2nd Bug, visually the Cellular value readouts make no sense for the overview page:
Cellular (though disabled) reports as “RSSI: 25”, what exactly is this 25? is it -25dBm?, if it is negative it should show this for diagnostics as its not clear.

Regards

Toby Partridge
ICT International

2

I’ve seen a very similar issue on a different openwrt-based gateway, and what happened was that there ended up being a stale global route pointing to another interface (wifi or ethernet forget which), and didn’t get superseded by a default route pointing to the LTE interface.

Can you get in to a shell in both states and capture the output of route -n for comparison?

The system I was working on had no gui, so I think the solution was to kill any pre-existing default routes in the script that runs when the LTE comes up - probably something like /etc/ppp/ip-up

There could also perhaps be issues with the nameserver - can you figure out if it’s routing or DNS which is not working?

Would also be interesting if you can confirm the issue with a Linux WiFi client or a phone or something other than Windows, just to rule any oddity there out.

3

Hey Thanks for the tips.
Unfortunitally I don’t have specific access to this gateway again for a while.
But we are making more LTE enabled RAK7240 shortly, so I can try and experiment to see if I can get it to lock up trying to use the laptop instead of the cell network.

Regards

Toby