A few days ago I noticed some odd stats on my home VDSL2 modem; steadily rising number of near end FEC errors. Perhaps 1000-2000 per hour and handled by the hardware without retransmits so nothing to lose sleep over, but strange because the SNR and attenuation figures are stellar.
It's probably been going on forever but as a matter of general principle I don't allow it access to the LAN (transparent bridging, normally only connected to the WAN port of my router which provides PPPOE) so I'm not in the habit of looking at the web interface. These things have indefinite uptime if you just use them as a protocol converter and don't ask them to do any actual networking.
To make a long story short the errors were coming in bursts associated with the furnace going on and off. Turns out that EMI from the PWM-controlled induction fan on Trane furnaces is a known problem. I'm using CAT5e for the phone line but I'm pretty sure that twisted pair provides little benefit in this application; anyway it runs right by the furnace on the way from the NID to the modem. Rerouting that wire 10 feet away diminished the problem to perhaps one-third as many errors on furnace start and none on furnace stop.
Still, I'm off to buy ferrite clamp beads in the morning to slap on every wire that looks like it has anything to do with that induction fan. If the furnace were still under warranty I'd make Trane do it.
It's probably been going on forever but as a matter of general principle I don't allow it access to the LAN (transparent bridging, normally only connected to the WAN port of my router which provides PPPOE) so I'm not in the habit of looking at the web interface. These things have indefinite uptime if you just use them as a protocol converter and don't ask them to do any actual networking.
To make a long story short the errors were coming in bursts associated with the furnace going on and off. Turns out that EMI from the PWM-controlled induction fan on Trane furnaces is a known problem. I'm using CAT5e for the phone line but I'm pretty sure that twisted pair provides little benefit in this application; anyway it runs right by the furnace on the way from the NID to the modem. Rerouting that wire 10 feet away diminished the problem to perhaps one-third as many errors on furnace start and none on furnace stop.
Still, I'm off to buy ferrite clamp beads in the morning to slap on every wire that looks like it has anything to do with that induction fan. If the furnace were still under warranty I'd make Trane do it.