5Ghz Wi-Fi completely jammed for 100ms every ~450ms

Hi folks. I'm fairly confident that the problem isn't OpenWRT, but I don't know where to ask my radio questions :frowning: and with any chance, someone here is enthusiastic about this sort of thing.

Since late December, I started having latency issues over my 5GHz network. I ruled out the router itself or any of my particular devices as the cause by trying to connect my devices to each other directly and seeing the same effect. This evening, after a frustrating match of League of Legends (due to the state of Wi-Fi and not my team, for once!), I decided to do some more data collection. Pinging my router every millisecond and plotting the time shows that nothing goes through for 95ms at a time, roughly every 450ms (95ms stalled and 355ms working):


The pongs for all the pings sent during the 95ms jam arrive at the same time, which creates the sawtooth pattern.

Somewhat frequently, the jam lasts for a lot longer–hundreds of milliseconds to entire seconds.

This could be a channel thing, but I already moved channels (from 36 to 40, FWIW) and I'm still seeing the same pattern. I haven't tried lighting up the 2.4GHz radio, but I can't imagine that things are going to be better.

I was wondering if this pattern reminds anyone of anything that I could investigate more. OpenWRT realtime graphs say the signal strength is -64dBm and noise is -91dBm, with small variations.

How congested is your 5GHz?
LuCI/Status/Channel Analysis.

It could still yield diagnostic data; try it.

Does this happen on both wired and wi-fi?

I don't know what this should look like, but I assume this isn't too bad:

I removed the network labels, but I'm the gray-ish network centered on channel 40.

What diagnostic data should I try to get with the 2.4GHz radio?

Well, I think we should try wired at this point.
I was looking for someone overpowering you with some crap reason but that does not poll quickly enough to catch an intermittent rouge transmitter and you dominate the channel.

On second look: who is surrounding you with a 160 width channel?

Yeah, I'm looking to go wired but I need my landlord to let me run a cable through the wall.

You have nothing to plug in just for diagnostics?

That's okay; let's look for the lag on 2.4.

Regarding who's surrounding me: I believe no one is, OpenWRT attributes the whole gray-ish segment to me (both the 20MHz-wide and 160MHz-wide gray blocks stopping around -10dBm).

I forgot to answer your other question. This only happens with the wireless connection. Wired is steady.

I'll check if it spikes on the 2.4GHz range too. This is going to take a bit because I have some shuffling to do.

EDIT: well, nevermind; looks like it's not doing it anymore right now. First time in weeks. I'll check out 2.4GHz if it comes back (which I imagine it will).

When it starts up again look at your system log for something that happens ~500ms; should be plenty of the same error; then switch to 2.4GHz and leave it there for 30 minutes because it is now intermittent.

Now we know it is, at the least, a 5GHz issue.

Ping at that interval isn’t really a meaningful way if measuring radio interference.

The only thing you measure is data congestion in the data line or the targets possibility to answer.

To be able to measure radio interference you need a radio scanner and a spectrum analyzer.

And depending on your whole setup these spikes are either the ping target getting angry at you or your other clients also wanting to send data.

So how many clients do you have on the 5GHz to begin with?

If the trouble is sport livestream and not streaming or data flow in general you maybe need to look at data control like sqm and multicasting.

2 Likes

OK, tried it and it also happens over 2.4GHz. For good measure, I also tried with an ad-hoc network between my computers, over 2.4GHz and 5GHz channels, and it happens there too, with my router fully out of the equation. They're both Macs, but my router pinging my Switch (Nintendo) has the same spikes.

Yeah, I don't mean to claim it's any good. I don't know what I'm doing when it comes to radio interference. Interference looks like a convenient explanation for what I'm seeing, but mainly what I'm asking is where to go from here.

I have 5 devices connected over Wi-Fi: two computers, two phones and a HomePod. The HomePod and the phones are usually sitting doing nothing at all, but even if they were, that's not that many things to have on Wi-Fi. It worked great up to December 23rd and then this started happening.

Okay, that is crazy. That is a huge block of spectrum. That is the largest radio spectrum I've ever seen. Consider calling the FCC for advice.

Turn everything off except one laptop and check again. if same->
turn it off, try another laptop. No change->

I hope you have an Android device so you can download wi-fi analyzer. It is free.
turn off all your devices except the device using wi-fi analyzer; airplane mode is not enough..

Use the Time Graph and look for a spike since it is ~1/2 second but not exactly, it should pick up a spike..

Started Christmas Eve's Eve....
If you have cable boxes in your router's room and you game room their are adapters that will convert cable to ethernet and back.

I don't have an Android device to use WiFi Analyzer on, unfortunately. If there's a similar tool for Linux, I could probably manage to run it on the router.

I filed an "Internet Services Issue -> Interference -> Jamming/Blocking (including Wi-Fi)" complaint with the FCC. I'll update this thread if I hear back before it ages out.

(I also asked my landlord if I can pass Ethernet cables through the wall, and I suspect that this will become the permanent solution.)

While this won't help your issue, but with the typical 80 MHz channel bandwidth for 802.11ac, ch36-ch48 are one single channel - so changing the channel "from 36 to 40" won't actually change anything at all (in either case it will use all of 5170-5250 MHz).

There's no direct frequency correlation between the WiFi 2.4 and 5.8 spectrum that would provide a single source of interference that wouldn't also severely impact multiple other services (such as cellular, and trust me, a troubled cell tower would get a lot of attention very quickly). Therefore, you either have someone operating a very illegal and potentially unintentional wide-band radio transmitter very close by, or you have something else going on that is making it look like something is wrong with wifi.

You say you tested this on 2.4 and it was doing the same? Did you perform the ad-hoc test with the wifi router still powered on? It's possible the wifi router itself and/or one of the clients could be the issue. The next time you perform a test via ad-hoc network, try doing so with the wifi router completely unplugged and all other wi-fi clients either unplugged or set into airplane mode. This will eliminate (or point to) one of those as the culprit, and at the very least, will help rule out a few more possibilities. Also, try directly between multiple sets of clients but not at the same time.

If this doesn't return anything usable on its own, it's time to think outside the box. Start with "known good" and "known bad" date points, to narrow down exactly when this started. Once you have that, what changes occurred around you between those two dates and times? Don't rule anything out yet. You might be surprised as to what can go wrong and cause unexpected signals to pop up.

Kali would have something that would work; I think it will run off a USB drive.

1 Like

FWIW, I imagine it's not one thing blowing the whole spectrum between 2.4GHz and 5GHz, it might just be one neighbor with a misbehaving AP blowing both the 2.4GHz and 5GHz ranges, or it could be two neighbors with each a misbehaving AP.

I find the date it started to be the most interesting:
December 23rd.

But in the absence of spinning radar, I'll accept Hanlon's razor for a while.

It could be a neighbor's AP or even one's own AP or one of the client devices that's decided to start acting up. It could even be an unintentional radiator that was brought into the house or into the immediate surroundings. Without RDF gear and the knowledge and experience to track down a problem signal, we can't so easily rule out a neighbor's devices unless all neighbors are physically distant enough to put them well into the "far field". We can, however, rule out everything in your control and also use local knowledge of the surroundings to advantage in trying to narrow down the culprit based on known changes around that time period.

1 Like

Do you have an ISP modem with builtin Wifi?
If so then verify if the issue occurs as well when connected to the ISP wifi