Here’s a weird one …

Attend a fault today, punter says dire speeds via wifi , but all OK via a wired connection . All quite usual so far.

All tests of the pair are good, VDSL test is good 72 sync, a handful of FECs in five mins ….

(Sky) router has been replaced already.

Throughput speed tests are 69/70 via ethernet, and a ping of 9ms same test to both his mobile and mine get between 3 and 7 and around 300+ms ping

It still seems quite usual….

I power cycle the router, unplugging the ethernet connections to the router. Once back up, wifi speed tests mimic the previous ethernet ones repeatedly.

So we methodically try reconnecting the various ethernet cables in turn , till we find connecting a 6 port switch instantly drops the wifi speeds, we narrow down the culprit to a new (series 4 I think he said) Raspberry Pi box.

This box is powered up, connected via ethernet cable to the switch, but nothing connects after at present. Tried the REIN gear on it, nothing doing. The guy tried various bits, removing the fan and the like, and it seems to have cured it for now.

So here’s the thing ….how could this bit of kit , connected to the same switch as the wired laptop which is never reporting slow speeds, be causing only the wifi from the router to be going tits up ??

I have an RPi 4 wire connected through an 8-port switch to my router (FritzBox 7530) with no issues, wifi or otherwise. But the RPi 4 is vanilla - no fans or other gizmos attached. It's not being worked hard at the mo - just running a Pi-hole server - CPU temp doesn't get much above 45C so no need.

Sounds like it must be interference at a frequency that only affects the router/wifi circuitry, but ...?

It really is weird …. how can it affect the wifi but not other stuff on the same network connected, as it is, via ethernet.

If its radiating stray RF garbage...pretty easy as WiFi is a nice open bit of ether.....

On the other hand, a decent bit of twisted pair CatX, should be fairly immune to external interference. Ditto the electronics on the switch ports which have transformer / baluns across each of the pairs for de-coupling and noise rejection.

Copper still has its place

Edited by Pheasant (Tue 17-May-22 22:12:34)

Out of curiosity, chargeable visit or not ?, or was it some sort of SFIO visit paid for by Sky…what seems fairly clear is that , whatever the issue was/is , it’s not really anything that Openreach are responsible for

Can't those Raspberry Pi boxes be used as an AP? if so maybe it was on the same channel as the router wifi.

@dect: Sorry, forgive my ignorance, AP?

The RPi 4 has both wifi and bluetooth. Mine aren't turned on, but if the customer was using one or both, that could be a possible interference source.

[ My RPi 4 is actually intended for another project, which I likely won't get around to starting till the autumn now. RPis have been hard to come by recently due to the chip shortage, and when I saw this one I grabbed it in case I couldn't get one later on. Till I get around to using it properly it's just passing the time being a DNS server. ]

Zarjax

Either the Pi had Wifi turned on interfering or the Fan motor is causing interference due to a freak frequency conflict at it's rotating speed although I would have expected this to show as REIN unless it is a very narrow frequency match.

Thank you all for your suggestions so far.

In reply to a post by Thaumaturge:

@dect: Sorry, forgive my ignorance, AP?

Sorry for the confusion, I meant Wi-Fi Access Point.