Smaller WIFI range with Annex32 1.5x

[PeterN]

Might be coincidental indeed ...
I will try now to reconnect a wifi client with increased wifi.power when a button gets pressed . But there are no good references for the correct power values.
I found only this: [Local Link Removed for Guests]
Do you remember?


PS: In the 80's we had some perfect sunny weeks near Brighton and Pool with a B&B that served breakfast, which always lasted us the whole day.

[bedesign]

In case you have not checked this already, as it was not mentioned -

1. These symptoms could, with a good likelihood indicate there is
(sporadic)radio channel congestion going on.

Why Wi-Fi stinks and how to fix it -
I use this open-source Android tool:
https://vremsoftwaredevelopment.github. ... usage-tips

https://spectrum.ieee.org/why-wifi-stin ... -to-fix-it

2. As was said, hip-hop switching to different fixed IP addresses can
trip up the Address Resolution Tables.
This requires a clear, alt. power cycle.

There is an early historic example of days of countrywide ISP
outage caused by the same MAC-address presented to
different switch ports

[PeterN]

Thanks for your thoughts ... but for my specific problem: my wifi scanner says no congestion, but intermittent low transmission levels as described ... and I'm not using fixed IP addresses and the DHCP server keeps returning the same correct IP for the MAC, if it's involved at all, because the module undesirably goes to AP mode.

[bedesign]

I agree with Peter's test conclusion about
the possible firmware-related power diff.

For other sporadic problems described by users -
And using the channel analyzer,
I have experienced that some stations are not
visible for minutes, so to be more sure one has to
monitor for an annoyingly long time.

A sporadic problem could be caused by a marginal PSU/cable.
An oscilloscope gives the full truth while monitoring close
to the chip while transmitting at maximum power.

[Electroguard]

Something I didn't mention previously to avoid possible confusion:
My main wifi router (default gateway) is remote, inside a room in a metal 'hanger', with the Sentry and Standby Sentry nearby to it.
Very little wifi signal leaks out from that metal Faraday cage, so most of my other annex devices which are outside about 50m away, connect to a local slave wifi router acting as a local access port, routing all subnet traffic to the default gateway main router via underground UTP cable (which was checked ok a couple of days ago).

So another strange symptom is that some of my local devices can connect to the remote default gateway and are available on the network, while others within a few feet can only connect to the network intermittently.
Therefore I added a local wifi repeater about a week ago to boost local signal and increase coverage, which sometimes helps, but sometimes not.
I suspect the intermittency may actually be because the "route" to the default gateway is not always via the fastest or strongest hop.
It is possible that some devices are sometimes trying to connect to the weak leakage signal escaping the Faraday cage.
So if there is more than one path to the default gateway, sending devices via the wrong weaker route might appear as signal loss or bad connection.

Which again leads me to suspect the possibility that mDNS may be interfering with the routing tables of my local slave AP router and wifi repeater.
Perhaps newer routers are designed to cater to mDNS, whereas old routers may not deal with it properly.

[BeanieBots]

Yes PeterN, I do recall (most) of that thead.
Position and orientation are very critical which is why I did my best not to change position during the firmware changes.
The ESPs should default to full power. If they are not, then that could be an issue with recent firmware.
Have you seen any improvement by doing nothing other than setting to full power?

Electroguard, my understanding (which may well be wrong) is that your system is Linux.
If that is true, then we can at least rule out Windows for playing any part.

[PeterN]

I have tried to adjust the power of an ESP32 running in AP mode with no success.
I was not able to increase (or decrease) the power with WIFI.POWER.
And there seems to be no way to read out the current power setting.
But my Windows Wifi Scanner (inSSIDer) shows that the signal strength of the Annex32 1.5x AP is significantly weaker than that of the 1.49 AP.
I have to admit that I have only done all the tests with M5 ATOM_lite devices at the moment.

And still absolutely confusing: A connection to the web installers console seems to increase the power in WIFI client mode.

I think now I need a PAUSE (1000*60*60*15)

[BeanieBots]

And still absolutely confusing: A connection to the web installers console seems to increase the power in WIFI client mode.

That might simply be ground plain effect.
However, as I mentioned earlier, there seems to be some other (maybe related) quirks dependant on console connection.
Try to test with two cables of the same length, one with data and the other power only. That will prove/eliminate if anything relates to actually making a data connection or if it's just ground-plaining from the cable/PC.

[PeterN]

Sorry; Forgot to mention:
#1 M5 ATOM + Powerbank with 5V-supply only and no USB-data lines : No connection to my WIFI
#2 M5 ATOM + USB-Datalink + Notebook with Amnex32-toolkit-console : No connection to my WIFI
#3 M5 ATOM + USB-Datalink + Notebook withAnnex Web Installer console: Connect as client to my WIFI

Edit: #2 and #3 with absolutely unchanged hardware software setting, except the other serial monitor software

[BeanieBots]

So, I think we can conclude that it's when there is active serial comms that things are OK.
Not sure why you could not change power with WIFI.POWER. Syntax error or just no change in RSSI?
Maybe a little bit of autorun code to switch between levels and toggle an LED to indicate which level it's at.
I think I might be seeing something similar with an ESP32-S3. Too much else going on at the moment to check for sure. But I will!

I have a 2.4Ghz spectrum analyser built into my RC Tx gear, so will dig that out to try and see if it shows anything useful.
I also mainly use inSSIDer for RSSI values. WiFiInfoView and Wireless Network Watcher are two other useful tools.

As for Full English Breakfast in a Brighton B&B, it will certainly keep you going all day, but may also give you a heart-attack the following day