For some reason I seem to have dropout problems specifically on races on Richmond. Four out of six really bad dropout sessions have happened racing on Richmond, large events but also small ladder races, including two different sets of equipment, and over BT, wifi, and dircon connenctions.
Two of those times, I haven’t been able to reconnect anything. Once it happened using BT on the Zwift ride, and today on the Kickr Bike Pro, on direct connect. In both instances the only thing that worked is restarting zwift, then no issues in the activities after or before. It has never happened like that in other routes or events. These two have been on large ECRO events, but I also have had no total connection death issues in other ECRO or other mass races (have had other temporary dropouts outside Richmond, just far less frequently).
Could be just random, but it just seems so odd that it would happen that much more frequently on this world specifically.
I’m on Windows 11, a decent laptop, fast Internet connection (100Mbps down, 30 up, latency 5-7ms). Lots of BT devices and WiFi around (live in a condo). Overall I’m just puzzled by i) Richmond being a curse, and ii) why would dircon still give me problems - so far, just there, but very early testing.
I would start by processing logs through trainerdx.com, from activities where you used both connection types. Aside from the dropouts, I would also be looking for clues that other system performance measures showed something unusual around the time of the dropout. For me, dropouts are like a once a year phenomenon on my Windows PC. It sounds like the common factor here (aside from the venue) is the laptop itself.
When you say DirCon, do you mean a wired Ethernet connection, or WiFi pairing?
I can’t say I’ve heard similar reports about Richmond specifically causing problems, though different worlds and areas of worlds have different CPU demands. I’m also not aware of Richmond being particularly bad in terms of system load.
just having similar issues - I had the Kickr Move before not one connection issue at all - rode everday pretty much since around sep 2024.. had the kickr bike pro for nearly 2 months and in the last 2-3 weeks 5-6 lost connections mainly in group rides - ive also found the resistance to be way too easy, hardly pedalling sometimes… Macbook UniFi Router - work with tech so no problem there (I think) Ethernet Direct Connect..
I used zwiftanalyzer and didn’t see anything. Will try that next time it happens.
My system is actually pretty good, I could run it on the best settings (I don’t because I don’t want to unnecessarily overheat the laptop’s GPU), but the compute is def there.
I tried both wifi and ethernet, I was referring to ethernet being puzzling with dropouts.
UPDATE:
Running it with ethernet directly to laptop and bypassing the modem has been stable. Haven’t had a dropout in this new setup. I think the issue was the modem maybe doing silly connection optimization things or whatever else, despite my Internet connection being fast and all the devices working really well. No Richmond races yet, but I’m thinking it was just a very unlikely fluke to see dropouts more common there across all set ups despite there not being a correspondingly higher number of “tries” within the world vs the rest. The total dropouts are probably well explained by some BT interference in my previous setups due to having a lot of BT devices around me, and then all about (hopefully) the modem messing things up as intermediary either on Wifi or Ethernet.
For me everything has been working well on the bike, except the issues in the races I had mentioned. Never happened on group rides but could be random. Bypassing the modem and going directly into my laptop seems to have worked for me like I mentioned in the other reply.
There is also this discussion of direct connect problems. Honestly I do not understand what is going on there. There have been some changes made to improve it, we still see reports of problems, but I don’t know if those are bugs or just the usual trickle of people having problems that look similar to an acknowledged bug.