Intermitent speed recording with magnetless sensor

It could be some slight magnetization of the chain or the crank arm.

If the problem is magnetization, it might be the chain, the cassette, something inside the freehub, the spokes or bearings. Everything else is aluminium (aluminium doesn't magnetizes). Most of those components are not present in the front hub, or are much less present (e.g bearings: 4 in the rear hub, only 2 in the front).

Do someone know for sure if proximity with steel parts can interfere with the sensor? If so, what is the mechanism? Magnetization? Attenuation of the radio signal?

What kind of sensor there is inside those "magnetless sensors"? If we know what sensor it is, it will be easier to find out what can interfere with it... we won't need to keep guessing...

The sensor detects the movement of the sensor through the earths magnetic field so it can be thrown off by something that has a magnet field of sufficient magnitude.

Well, some hundreds of Km ridden since switching to the front wheel hub, I think it's safe to say the problem is solved. The only times the speed sensor failed since then was when I was crossing bridges over electric railway lines. I think that there was a lot of electromagnetic interference there.

So, it's true: the steel parts of/near the rear hub may interfere with the sensor. I'm sure it's not a distance sensor-main unit problem because I've tried to near the Edge 500 from the sensor and the problem persisted.

My evaluation of the sensors: although they are more interference prone than their magnet-based cousins, they do the job. I have a feeling that their readings present a little more jitter, specially the cadence, but nothing that bothers.

Thanks to everybody.

Faraday cage!

Well, some hundreds of Km ridden since switching to the front wheel hub, I think it's safe to say the problem is solved. The only times the speed sensor failed since then was when I was crossing bridges over electric railway lines. I think that there was a lot of electromagnetic interference there.

So, it's true: the steel parts of/near the rear hub may interfere with the sensor. I'm sure it's not a distance sensor-main unit problem because I've tried to near the Edge 500 from the sensor and the problem persisted.

My evaluation of the sensors: although they are more interference prone than their magnet-based cousins, they do the job. I have a feeling that their readings present a little more jitter, specially the cadence, but nothing that bothers.

I want to thank everybody too.
My GSC-10 is about 2 or 3 years now, and I transferred it to a new bike, so I can use it in my house with a super-Crono Hydro (Elite-it.com) trainer. Well, the system and comfort are amazing, but I noticed that my speed was going crazy, 0 to 30, never standing more than 2 or 3 seconds at the same speed, despite my continued pace.
I am ashamed to say that, without any research, I thought (?) that the sensor was faulty and bought the new magnetless sensor speed/cadence (damn 1-click buy!).
After installation, to my surprise, the speed stability was a lot worse, useless. Only then I started to do the right thing, i e, search and got here... :(
When I came across the reference about the interference of magnetic fields, I researched how to make a simple Faraday cage and, guess what? it worked!
I mean, the old sensor (GSC-10) is now working perfectly, completely stable at speed (cadence was never a problem for either of them), and the new sensor, magnetless, improved, and now it looks like the old behavior GSC-10, without going from 0 to 30, floating "only" between 9 and 27 for a pace 20, for example.
I made the Faraday cage (http://www.thesurvivalistblog.net/build-your-own-faraday-cage-heres-how/) covering up the "engine" of the magneto-fluid resistance with plastic that wraps motherboards and other computer components; I put a layer only; may test with two layers or aluminum foil too other day.
Ahn, and to compare, I did two tests:
One test was concurrent with 2 Garmin: 620 paired with a sensor and magnetless 920XT my paired with GSC-10; 620 went crazy and stable 920;
The other test was with two 920XT profiles, each with different sensors; changing profiles, the speed goes from crazy to stable according to the recognized sensor.

If u need more guidance, I can post video and photos, but is very simple.
PS: it is really ugly ...

and so the summation is that there is next to nothing that can work for the magnetless variety? If some bike have magnetized components... we had no idea what magnetized the, or to what level they are magnetized couldnt rare earth magnets or something be used to "de-magnetize" something perhaps. I was planning on getting the new magnetless sensor, mostly for my indoor training. but if the mag-resistance traininer is going to interfere. Sessh. should i even bother. or will the hub be far enough away from the rear/base of tire on the trainer

same problem

Bought the magnet-less sensors for both speed and cadence. Went riding and saw the intermittent problem of speed dropping to zero and my garmin Edge 800 beeping because auto pause kicked in. My speed sensor was mounted on the rear wheel. Will try the suggested fix by moving it to the front wheel. One reason I´m thinking of that could cause an issue and would be good to get Garmin´s feedback about, is that I have the Shimano Di2 electrical shifter. Obviously there is electronics in the rear derailleur. Normally it is only active when you change gears, but perhaps it causes interference. I would suggest for garmin to fix this in SW to ignore the dropouts when contact is lost. This is easily detected and values could be ignored. A real live bike doesn´t stop in a fraction of a second and connection state is monitored by the Edge.

Tested things a bit further:
Read somwhere WiFi can cause interference. I had my iPhone in my backpockets of my cycling shirt. The wifi will become active when you pass homes with accesspoints. I tested today by switching off wifi and Bluetooth. Unclear if it made a difference. What did have an impact is to take off the cadence sensor and only leaving the speed sensor active. It's clear the cadence sensor is interfering with the speed sensor, even though I have the settings on "Speed and cadence" in the bike profile. Each time I stop pedaling, the RPM goes to zero and even switches off, I have often coinciding issues with auto pause kicking in. I have not tried moving the speed sensor to the front wheel. I can imagine it would suffer less from an interfering cadence sensor. I don't understand how the cadence sensor can cause interference. The haertbeat sensor is also an Ant+ transmitter and that seems not to cause issues. Shame on Garmin for producing sub optimal sensors.

I had dropout problems with one of my speed sensors when it was new. I believe it was a battery contact issue. One of the contacts is through the battery cover. Try rotating the cover between the locked to unlocked position a few time to get the contact better seated.

drop out problem solved

Placing the speed sensor on the front wheel hub, solved the issue for me. Still not sure why the cadence snsor seemed to interfere, when the speed sensor was on the rear wheel, but anyway. It's stable now. No drop-outs anymore.

I am in the same boat, GSC-10 broke so got the magnet-less speed sensor from Garmin. Was excited to try and it has been nothing but trouble. Very inconsistent readings and dropping to 0. My primary use for the speed sensor is on the trainer as when road riding I count on GPS. So needless to say mounting to front wheel won't work. What I am gathering here is that its likely the trainer I am using that is magnetized? Also very disappointed in Garmin customer service here... have called them 4 times to troubleshoot the issue, and not one person had any clue what I was talking about... .yet on this thread there are lots of people with same issue. Are they playing dumb here? Garmin customer service has been outstanding with exception of this issue. I may need to buy another computer now... booooooo