I'm using a Garmin Forerunner 35 (pretty basic running watch but it does what I need), and a Garmin Speed Sensor 2 on my gravel bike.
When I run it on the turbo trainer with the GPS off, I get pretty reasonable figures for speed/distance. All good so far.

I left the GPS on for a trainer session (you configure it independently of the rides) and found that the speed sensor was showing 0 or near 0 for large chunks of the ride and assumed it was broken until it uploaded a scatter map to Strava.
It seemed that it was sometimes using the GPS speed (0) and sometimes the sensor speed (correct) but I can't see any particular pattern to it beyond it maybe relating to the points where the GPS thought I was moving.
Turning the GPS back off and the speed sensor is fine.

Does anyone know how the Garmin decides which one to use?

I want to start taking the bike back out on the roads, so should I be OK to go on the road with the GPS and speed sensor or should I take the speed sensor off? I'm aiming to use it outside at weekends and during the week on the trainer so I'm expecting it to change from indoor to outdoor and back most weeks which will be a bit of a pain with the sensor, and though I'm tempted to get another bike just for trainer use I can't really justify the space or money.

Stupid question, but does your watch have a outdoor versus indoor cycling setting? Mine does.

As for your actual question I can't help.. Is there an airplane mode / GPS off option for your watch?

(F5 + 2nd gen speed/cadence)

My experience is that the watch will display speed based on sensor data, whenever available. Noticeable at traffic lights - w/o sensor, you'll see speed going down while you're stopped, and pick up slowly when you resume riding. With a sensor, readings are properly (instantly) updating.

I've never noticed my watch using GPS data to calculate speed if my sensor is linked.

Garmin's do strange things when you use them while stationary. Most want the gps disabled when you do stationary stuff. As for wheel and cadence sensors on a stationary bike, I'd hate to begin to guess. When I participated on the Garmin Community Forums many years ago, use of a GPS and sensors on a stationary bike or trainer seemed to have a lot of people having unexpected results.

You might check with them, the forum is still fairly active......

https://forums.garmin.com/sports-fit...ng-multisport/

https://forums.garmin.com/sports-fitness/cycling/

Annoyingly there's an option for indoor/outdoor runs but only a single option for cycling. I need to then go into a menu to toggle the GPS on or off. So I can use it indoors with the GPS off and outdoors with the GPS on.
There's no control over the speed sensor though; it's either in range and spinning or it's not. So if I want to do a run without the speed sensor I need to take it off the bike.

With the less expensive Garmin's you'll have less features. Garmin Forerunners are first and foremost a running device. The higher up in the Forerunner price range you are willing to pay the more they'll have features for cycling and other things.

I'd think that like other Garmin Devices you can make profiles and associate particular sensors to that profie. So maybe try making a running profile, a stationary bike profile and a "I really am riding a bike" profile. Then if you make certain what profile is active when you are doing what ever, then you'll have better results.

By default, the Garmin will use GPS data to calibrate the speed sensor. The speed sensor is the primary source, but something like a 500m segment the GPS can get a really accurate measurement of the wheel roll-out.
If you're stationary and the GPS is enabled and locked, I could see that doing really strange things. I never had an issue with my 510, but it usually couldn't acquire satellites indoors.

Yeah this is the cheapest one currently available as I wanted to move to a watch for running/cycling and don't need it to be *that* accurate since I just want to track progress over time and I'm not going to be setting records. Being able to toggle the GPS on/off for cycling is a nuisance but I can live with it for the budget

That's pretty interesting; so in theory using both on a run outside should give me a better calibrated speed sensor to use indoors too? The speed sensor was set up using the app and the ETRO number on the side so may be a bit off. I might just leave it on and see what happens then.

Speed on a stationary bike really has no meaning. It's not a good measure for whether you'll be able to keep up with anyone else doing real cycling. Nor is it a good comparison with other stationary cyclists as by itself, you don't know what resistance level they were at. Distance on a stationary bike is the next silly term I hear people comparing.

Power is about the only useful comparison. Next to that just time. IMO of course.

My garmin 1030 + with wheel sensor is not instant :/

I bought the speed sensor thinking it was going to be night and day different. At least for me it was not. I am sure it's more accurate and probably reacts faster but I come to full stop before my 1030+ ever says zero.

I know it's reading the sensor because I can lift on the front wheel and spin it the gps shows I am moving.

I'm not so sure, time is a pretty meaningless comparison because an hour at 5mph is going to do a lot less cardio/fat burn than an hour at 30mph.

As a comparison to anything but itself I agree it's kind of meaningless, but I'm wanting to try and get some meaningful metric from the static biking to try and track progress and push myself. So I've started logging the miles. Time, distance and average speed seem decent enough for that.

I'd disagree about the 5mph vs 30 mph. I don't use a trainer or stationary bike enough to relate to them. On a bike I can do slow speeds at a very high cadence and will have a very high cardio benefit. I can do faster speeds at a lower cadence and be below cardio benefit levels.

On my wife's stationary bike, it gives me a target HR when set up for my weight and other metrics. So if I cycle for a specific time to that target HR, then I think time will still win out. Or you can just set your own target HR. Personally for fat burning, I'd think high zone 2 or low zone 3 ideal. Even on a bike, if out to burn fat, staying below my LTHR seems to do the best. I don't seem as ravenous after the ride. If I allow myself to go hard, then I feel as if I just eat back those Calories I burned off.

If you're avoiding the extremes (spinning in the lowest gear and grinding in the highest gear) then I think speed/mileage should be fairly accurate. My top gear is about 4.5 times faster than my bottom gear so it'd be really hard to use either at an overlapping "road" speed.

Heart rate may be the best to use, though.

Heart rate would be better, and for stationary-bike workouts where you aren't changing resistance very frequently (either through gearing or trainer setting), it's pretty good. If you are changing resistance frequently, it takes your heart rate a while to catch up with each change, so it's a less accurate reflection of your true effort.

Power is the gold standard according to current training philosophy, but you either need a power meter or a smart trainer to get that.