Wifi error log for Bookworm?

All I can say is that my PiZero running bookworm became stable with wifi after a more powerful PSU was fitted to it.

It turns out that moving the Pi5 within three feet the access point restores normal WiFi operation. Signal quality reported by wavemon went from 90% to 100% and the WiFi connected automatically in parallel to a wired connection. However, neighbor's APs using the 2.4GHz band appear with 67-76% signal quality. If I were to move the Pi5 back to its original location, its signal strength at the AP would be much lower. It's difficult to improve the Pi5's antenna, but my AP has an SMA connector. Does anybody have experience with directional antennas? At the moment, the Pi5 is showing 5.05 volts at the GPIO header. I think that's slightly lower than I saw previously when the Pi5 was in the troublesome location. The PSU doesn't seem the prime suspect.

Thanks for reading, and any suggestions.

bob prohaska

Can you not use ethernet cable?

Not conveniently. The AP is at the back of the house, the Pi5 is used near the front. A neatly-routed cable will be close to 50 feet long, with lots of zigs and zags through three rooms. Physically, it's possible but not desirable.

If a directional antenna is practical that would be much better.

Thanks for writing,

bob prohaska

Back in the early days when broadband Internet was still exotic and expensive and wi-fi was new, people experimented with using it for longer range point-to-point connections. I remember reports that empty Pringles potato-chip cans were a remarkably cost-effective way of achieving the necessary directionality.

Memories, memories. Back in the day all I could get was dialup Internet service. My neighbor however had one of them new-fangled satellite connections (quite s***ty by today's standards). He rigged up a cantenna so that I could "borrow" his WiFi. I was allowed to use it in the middle of the night. Otherwise I'd override his allocated bandwidth.

I do now wish to relive that time in Internet history.

As I understand it, the Pringles "cantenna" was intended to be maximally directional. If I understand my problem correctly the need is to "shadow" my access point from the neighbors more than to direct transmit power to my clients (signal level at the client location is ~90% per wavemon,

There seem to be a few complaints about wifi behavior on the Forums, since my problems appeared out of the blue after an upgrade it's possible that software is involved also.

Something very similar happened after a bookworm upgrade a few months ago. After some hair-pulling and a few more upgrades it simply went away.

As an aside, now wavemon seems to crash on its own, reporting:

│Waiting for scan data ...free(): double free detected in tcache 2 │ │ Aborted

That seems clearly to be a software issue, though whether in wavemon or somewhere else isn't clear to me.

Thanks for writing,

bob prohaska

Moving transmitted frequency can help a lot

My AP is using 2417 MHz ch 2, none of the nearby Aps are using it. The second strongest is using 2437 MHz ch 6. That's with the Pi5 within three feet of my AP. At the moment Wifi won't connect at all, though both my cellphone and a second computer farther away connect successfully.

Another puzzle is that wavemon reports 5825 MHz ch 165 after the connect process fails. The targeted access point uses only 2417, so it's as if an unprompted search has started. Meanwhile my cellphone has no trouble connecting to the AP, near or far.

There are 15 APs within range, most reporting ESS including mine. Anybody know what ESS signifies?

Thanks for writing,

bob prohaska

snipped-for-privacy@www.zefox.net wrote: [snipped] Just to add to the confusion, a few minutes ago the wifi started to work. The client is still about three feet from the access point.

The signal level is -33 dBm, the strongest interferer is at -60 dBm, roughly the same situation as earlier when the wifi wouldn't connect.

A quick test of the voltage at the GPIO pins reads 5.05 volts, a few millivolts lower than seen when wifi didn't work. No software updates have become available. Basically no controllable changes between not working and working.

Thanks for reading!

bob prohaska

This sounds very familiar. I fought the same pitched battle here with Wifi and Pi5 under Bookworm. Finally discovered that HDMI output interferes with the 2.4 GHz band. Try running headless over ssh and see if that's any better? Made an improvement here for 2.4 GHz, but did not help 5 GHz connectivity.

I finally ended up replacing an 8 year old Wifi router and that fixed the 5 GHz band. Something about the Pi5 Wifi implementation really does not like older routers.

I have read this elsewhere. All my Pis run headless

One other oddity that probably isn't germane is that my pi PICOS would not connect to an extreme 'edge of band' frequency which is UK legal until I removed the code that said 'use only UK legal frequencies'

What router are you using? Mine is indiputably old, D-Link DI-524, but it works with everything else now and worked with the Pi5 except for two major episodes, a couple months ago and again now. In the meantime it was fine..

The idea that hdmi is the culprit can't be ruled out, but I'm skeptical it's the source of the trouble. This particular Pi5 has _always_ been used with hdmi operating. Wifi has been a problem only intermittently, usuall a day after an OS update. In the first instance, the problem simply went away after a subsequent update. In the present instance, the problem appeared a day after an update and persisted through yesterday, when I moved the Pi5 next to the router to check signal strength. Initially wifi worked, then it stopped working, failing silently and not able to re-connect.

Last night wifi connected and remains up as I write this. After turning wifi off, it turned back on without trouble while hdmi was active, something that didn't happen yesterday. Wavemon presently reports -31dBm on the active connection and -58dBm for the strongest interferer. I'm using 2417MHz ch 2, the strongest interferer is 2437MHz ch 6, which has been consistent over time.

Thanks for writing!

bob prohaska

You really need to get a router that was already considered obsolete by people like Marconi and move into the modern era.

On 2.4 GHz only use channels 1, 6 and 11, as the others overlap those and cause worse interference than sharing with a stronger signal on 1,6 or 11.

---druck

Aye, but it were good enough for my father, and his father before 'im.

:)

---druck

Perhps you've hit the nail on the thumb 8-)

Channel 1 is used at 2412 MHz, channel 2 is used by me, channel 6 is used at 2437 MHz (5 signals) and channel 11 is used at 2462 (3 signals).

The channel 11 signal appears to be about 10 dBm down and the weakest of all the interferers at my AP. The situation will deteriorate if I move the client back to its working location but still might be an improvement over the status quo.

I'll try channel 11 and see what happens.

Thank you very much!

bob prohaska

Attempting to set channel 11 b/g appeared to work but resulted in password rejection, repeatedly asking for the password which is known to work on other hosts and on this host previously.

For some reason it's difficult to get rid of old connection presets, making it a little unclear if changes I'm trying to save actually work as intended.

In the meantime a large batch of updates became available. I'll install those and reboot.

Thanks for reading!

bob prohaska

ROTFL