Steve,

the RSSI is read for the connection in use, -127.0 is the value for the unconnected state. The value needs to be smoothed as it's quite jumpy, so may need 2-3 seconds after connect to get to the actual value.

IP association normally takes long enough from connect to get the RSSI into a usable state, and thus the WifiStaCheckSQ() call in WifiStaGotIP() is meant to switch to the network as soon as possible.

For the case of fast DHCP association, you could try speeding the RSSI adoption up by checking for m_sta_rssi == -1270 and skipping or weakening the smoothing in that case, so the first RSSI read after a connect will be used immediately / stronger. But I remember that first value being far too optimistic in my test cases, which were having a single AP barely reachable from the car. I suggest testing that case specifically, as it seems your current test setup is close to a best case scenario.

OTOH, removing the WifiStaCheckSQ() call in WifiStaGotIP() would introduce at most 1 second delay for the netmanager to switch to the Wifi network, as the next ticker.1 run would trigger the switch. So that's not critical I think.

Another option: WifiStaCheckSQ(NULL) triggers the events unconditionally of the previous state, just by the current signal quality. That was meant to process a changed threshold configuration, but I don't remember why I did that in WifiStaGotIP(). Maybe that's why the "bad" event got triggered twice? You could try passing ms_m_net_wifi_sq there instead of NULL.

Regards,
Michael

Am 17.12.20 um 03:58 schrieb Stephen Casner:

Following up #2:  Perhaps we can just remove the call to
WifiStaCheckSQ() in WifiStaGotIP()?  If the IP address has just been
assigned then the radio must be working well enough to exchange DHCP
packets.  That would avoid the problem I observed that the signal
quality measurement was not yet accurate at that point.

With that call removed and with a script "wifi reconnect" attached to
the network.wifi.sta.bad event I can walk to the corners of my house
carrying the OVMS module and see it switch cleanly among my three APs.
See the attached log.

                                                        -- Steve

On Wed, 16 Dec 2020, Stephen Casner wrote:

Following up: I see that OvmsNetManager::WifiStaGotIP() calls
WifiStaCheckSQ() at the start of the connection, but it looks like the
radio hasn't fully adapted yet or something.  The signal quality can
be -95.5 -99.2 dBm and sometimes even -127.0 dBm, which looks like an
uninitialized value, even though wifi scan says the RSSI is -60.  (Is
RSSI calibrated to dBm for this system?)

If I stop the looping by "wifi mode off" and "wifi mode client" then I
see a good signal quality -79.3 reported after getting IP, and then if
I do "wifi status" the reported signal quality is -56.0 dBm.  This is
all with my test OVMS v3.0 sitting on my desk.

So would it make sense to introduce some delay here if the signal
quality is not good enough yet?

                                                        -- Steve

On Mon, 14 Dec 2020, Stephen Casner wrote:

On Mon, 7 Dec 2020, Michael Balzer wrote:

If you don't mind frequent disconnects in poor Wifi situations, a simple
solution now is to attach a "wifi reconnect" command call to the
"network.wifi.sta.bad" event. I added that command on the Wifi rework some
months ago.

I tested this today with the three APs in my home but ran into
trouble.  It looks like the event was emitted again at the end of the
reconnection process so another reconnect was performed and it kept
looping.  See the attached log.  I have not investigated this
carefully.

For a better, general solution, we could e.g. count the "network.wifi.sta.bad"
events on the same network (or better: check the event frequency) and
disconnect when reaching some (tunable) threshold. The next automatic scan
then connects to the best network available. Getting stuck in the "bad signal"
state for too long would also need to trigger a disconnect.

I'll look at this further to see whether a smooth transition from one
AP to another on the same network is possible.

                                                        -- Steve

_______________________________________________
OvmsDev mailing list
OvmsDev@lists.openvehicles.com
http://lists.openvehicles.com/mailman/listinfo/ovmsdev

_______________________________________________
OvmsDev mailing list
OvmsDev@lists.openvehicles.com
http://lists.openvehicles.com/mailman/listinfo/ovmsdev

-- 
Michael Balzer * Helkenberger Weg 9 * D-58256 Ennepetal
Fon 02333 / 833 5735 * Handy 0176 / 206 989 26