Re: Raspberry Pi 4's IP Address 169...

On Thu, 13 Jan 2022 20:26:05 -0000, "NY" snipped-for-privacy@privacy.invalid declaimed the following:

But when you've got something like 6 neighbors all using "optimal" channels 1/6/11 -- you might be better off on channel 3/4 where you might not be getting channel usage congestion/collisions...

I always use channel 13 and I'm always the only one on it. With everyone elso on "auto" even channel 11 is rarely used and if so it's usually very weak. It seems they all begin on 1 and only move up when needed.

No, that's not how its done. At least not for original wifi. OFDM is only used in later implementations..802.11a and g DSSS is used on the rests.

Most routers support 802.11b,g and n. for 2,4Ghz And 802.11a, an and ac on 5GHz

You should have read it. ADSL 'broad band' uses 'bins' like that. Wifi spread spectrum used DSSS - it multiplies it (convolutes) by a very high frequency pseudorandom code. What comes out is pseudorandom noise across the whole two or three channels. You convolute that again with the right code, and there is the wanted digital signal. Whereas any non random single frequency noise in the channel becomes convoluted to noise.

Any other AP on nearby frequency will use a different code, so it will just become 'noise' once convoluted and deconvoluted by the wrong code.

I was supposed to know this stuff, and I got as far as understanding the principle, but the bloody maths was beyond me.

Frequency hopping is much slower than DSSS - Its an extra layer possibly on top of DSSS used to shift from the centre of one channel to another.

It appears from your interesting PDF that bluetooth uses frequency hopping on much narrower channels - 1MHz instead of 22MHz. Without using DSSS

But the interesting thing about 2.4GHz is that the ONLY restrictions legally placed on it are spectrum power and width. What you use within it can be almost anything

That might be the case if you've got an old Raspberry Pi with a knackered SD card acting as a file server, but a Pi 4B with a USB3 spinning rust drive does 100MB/s which is 8x faster than 100BaseT. Another of my Pi 4Bs with a SATA3 SSD easily saturates the gigabit Ethernet.

---druck

The whole 2.4Ghz things is a messy delight of fun. I fly model planes on

Another funny story is that one particular very well known 'brand' of transmitter went out with an individual code in its flash memory. Essentially a MAC code. In use you press a button on the receiver and it then 'pairs' with that MAC code...so different transmitters can coexist....

...until someone discovered that switching the transmitter on to check the battery charge state and then immediately switching it off crashed the NVRAM and erased the code...to zero. So all the transmitters ended up with that code.

As far as wifi goes, even access points on the same channel will only really be an issue if the receiver close to one trying to receive the other. In DSSS the convolution code will sync with one or the other but not both.

And I *think* that there is in general radio silence except when data is being transferred. Or SSIDs being transmitted

Certainly when I ran my wifi scanner in a hospital last, there were about 20 channels of varying strength on the three main channels, and it all worked

But I hate wifi. Bloody unreliable. The spark igniter on my central heating oil boiler reliably disconnects any wifi device in the house within 20 feet of it. My Pi-zero W maybe has 5ft reliable range through a wall, and the worst wifi chip in the world

Oh dear. Its in the dining room and its managed to connect itself to the living room, 6 meters away...7Mbps instead of the kitchen

Probably that one second power cut the other day..

wlan0 IEEE 802.11 ESSID:"LivingRoom" Mode:Managed Frequency:2.432 GHz Access Point:

let's see what rebooting gets me...

iwconfig wlan0 wlan0 IEEE 802.11 ESSID:"Kitchen" Mode:Managed Frequency:2.457 GHz Access Point:

Ah. That's better. 12 dB better. Still rubbish.

I get -70dB and 40/70 quality in the laptop about 6 meters away from the same point and that manages 150Mbps....or thats what iwconfig says anyway., Realtek twin channel chip.

No. That's the worst option.

Wifi networks on the same channel detect their presence and schedule their traffic so that only one transmitter is active at a time, reducing garbled packets and retransmissons.

Wifi networks operating on adjacent, overlapping channels cause unpredictable random RFI, reducing performance for all users due to retransmissions. The same is true for a mix of 20Mhz/40Mhz networks.

For best performance, switch to 5GHz on a frequency not in use by neighbors with enough spacing to enable 80+MHz channels.

- the range is less because 5 GHz is attenuated more than 2.4 GHz for the same location of router and computer (ie the same distance and the same obstructions)

- some older devices (my old laptop, our Foscam security cameras) don't support 5 GHz, so you may need 2.4 left on for compatibility

Nowhere in our house does 5Ghz offer any advantage, it's only if I put a client virtually on top of the router or AP (I have two which offer

Where I am using my laptop at the moment 2.4GHz gives me 300Mbs and

We are lucky in that we have virtually no neighbours close enough to have any effect on our WiFi, I can only see one, faint, signal from our nearest neighbour who is across the road from us.

The Pi-zero W's WiFi is fine for a single antenna of that size - as long as it knows what to connect to.

If you are using multiple consumer access points you are always going to get issues like this. Proper managed WiFi can be set up to ensure a device connects to the most appropriate access point.

---druck

The problem is its set up to try the kitchen first, then the living room

But, under power cut conditions the kitchen - a repurposed Netgear ADSL router - takes longer to boot than the Pi..

Ergo it found the nearest and best access point it could.

The problem is that i dont know how to make it switch to a strionger one if there is one, later.

My nearest neighbour is 200+ yards away.

All I can see are my three access points, one of which is almost useless, and came with the router, which is in a room with foil backed plasterboard walls.

It has been quite a struggle to find good places for the access points, based on where people like to use WiFi.

Only if "a device" cooperates. Modern devices *should* have the latest WiFi standards updates in them so that they understand and usew the hints given by mesh systems. However it is always down to the client to choose which AP to use and older (and other badly configured) clients may well not choose the right AP or move when sensible.

An actively managed system can send a drop to a device on an undesirable access point, which forces it to re-establish a connection to an access point with a stronger system.

---druck

Yes, but dropping the connection will break whatever is going on. lost phone conversation, broken streaming, whatever. It's an absolute last resort to drop the connection.

It's usually done at the point the device first connects to the unsuitable access point, so nothing is going on then, and it just appears to take slight longer to get a WiFi signal.

---druck

That assumes you have a client that does a "first connect" after not being used for a while. I'm much more familiar with the way my laptop does things and that simply never disconnects unless I explicitly tell it to. Thus there isn't a time when "the device first connects", it has been connected all night and I've just walked through from the lounge where I used it last night to the breakfast room where I'm using it now.