njkayaker

The 305 is ancient.

The preference for the speed sensor existed in the still-old 800. I'm not sure what the situation was with the 605/705.

Who knows why Garmin didn't have the 305 prefer using the sensor. I wouldn't call the change "a big about face".

JohnJ80

Do you want me to more loudly agree with you?

J.

njkayaker

You aren't being clear.

You called it a "big about face". I disagreed. But we agree? Huh?


GPS speed works ok (usually). The speed sensor works better.

Maybe, Garmin initially intended that the speed sensor was needed only for a bike on a trainer.

(It's amazing that people are "complaining" about what the ancient 305 does.)


No, GPS (even "more advanced" ones) are weak for speed for low speeds.

The issue is that the error in GPS position is about the same as a couple of wheel rotations.

Counting wheel rotations is very accurate.

With a properly calibrated circumference, it's much more accurate over short distances to count wheel rotations.

The limitation is inherent in GPS units that are cheap (and small and that are moving at slow speeds and in places where signals are blocked).

It's not going to change (any time soon).

So, it's not "surprising" that they prefer the wheel sensor at all (even for current units).

Garmin figured it out a long time ago (with the 605?).

rebayona

I own a 530 and I've noticed the speed reading is moving around numbers while climbing (mountains/trees around) despite I have a speed sensor. It is like 20... 21... 11...14... 21... and I'm at a steady pace!

Am I missing some setting somewhere? It is like it is not using the speed sensor and the speed is being calculated through the GPS only.

JohnJ80

bike computers fall back to the gps if the sensor isn’t reporting accurately. I’ve had this happen a number of times when the battery gets weak - it can even seem like it’s connected but the readings are flaky.

Replace the battery is the first step. Then I’d unpair it and pair it again.

Tourist in MSN

Old thread resurrected after five years.

Seattle Forrest

Somebody had a detailed question about this topic. And bumped it to ask. That's what threads are for. It wasn't a drive by spamming.

rebayona

I replaced it few weeks ago. I'll try re-pairing then, thanks for the suggestion!

thank you! So I found a thread having the same issue I'm having, how would it matter how old is it? How can it bother some people? 🤷

Iride01 

Because your issue is not anything about the issue describe by the starter of this thread.

So someone reading this dug up zombie thread might not realize you ask a question and then begins answering the OP's questions. Or perhaps you might re-kindle and old argument that is better left in the grave.

It will always be better IMO to start a new thread. Although I'll give you credit for not doing what many do and just say "I have the exact same issue, help me". Because usually when you get to questioning them it's not really the same issue at all.

If you find a thread that is pertinent to your issue, then you can simply reference that in your thread by just linking to it.

Seattle Forrest

This is a forum for talking about bikes. Some people get really upset if they think you're talking wrong. I mean it's about bikes, but not in the right kind of way. You can't please everyone. 🙂

GhostRider62

If you are climbing at 21 km/h.....you must be a really fast racer. If you are climbing closer to 6-8 km/h and the displayed speed is jumping around and it is a device that measures speed with wheel rotation, you are seeing aliasing error (nyquist frequency).

The other possibility is GPS. GPS speed is inaccurate over the short run, especially on curvy, tree covered climbs. Have you checked that your speed sensor is enabled in "Sensors"

rebayona

The thread is titled: "GPS vs Sensor for speed" and guess what: I'm wondering if my issue is caused by the GPS overriding the speed sensor.
If I start a new thread there would be probably some other person asking me to "search before posting duplicate questions". I guess "You can't please everyone", quoting Seattle Forrest

It is a 3% grade climb, so no big deal for the 20km/h. I just checked and they're visible there:

The sensor is paired, the behavior is consistent with a GPS miscalculation (there might be tall trees during that part). From what I've read in other forums the speed sensor will be the main source for the Garmin regarding speed by default, saving power as the GPS will be used for altitude and route only. I don't see any setting to tell the Garmin to use the speed sensor as its main source, but from what I see in the field and also in the graphic it doesn't seem to be doing so.

Any ideas or theories here?

Iride01 

Almost never will you see someone say that. I'm sure it's happened though, but not often and not that I can recall.

I'm not trying to be rude to you. I'm just pointing out that it's not the normal thing here, nor is it the normal thing on many other forums going back to the days of BBS's. I was new to BBS's in the early 80's (as were BBS's in general) and I was similarly doing what you did and ask my own question in another's thread. So you aren't the first one to be admonished for this.

Certainly you should be able to see that if everyone thinking their issue was the same as another and put them in the same thread then we'd have a very long thread with many replies and many times some replies not obvious about what member or issue they are intended for. And as I said, frequently the issues are not the same issues, nor do they have the same solutions.

rm -rf

(That was a bit confusing when I landed on the first page of this old thread.)

You can test your wheel sensor: while off the bike, spin the wheel with the sensor. if your GPS (Garmin?) shows random very slow speeds, it's using the GPS signal, which randomly drifts around your location if you are stopped. The wheel sensor will show the fast wheel rotation speeds.

I like having a wheel sensor on my Garmin. It's more accurate on slow, steep climbs, and handles starting and stopping a bit better. It's not critical, though.

I'm not sure why your climbing recording showed those odd dips. Try uploading the recording to ridewithgps.com, and see how it looks there. They have different smoothing methods.

rm -rf

Ride recordings need to smooth out the second-by-second raw data!
Each second of the ride is another data point. Slight inaccuracies in location and elevation will produce big swings in the data from one second to the next, up and down.

I save my rides in the free software "My Tourbook" on my PC, along with Strava.

My Tourbook has settings for the smoothing method and amount. I kept the defaults. Too little smoothing, and there's fake peaks in speed and elevation gain. Too much, and top speeds, gradients, etc, are all rounded off. It's a tricky balance.

Here's a climb from a recent ride.
The green graph is elevation from my Garmin 1030. The red line there is satellite elevation data (SRTM data) which can be inaccurate on climbs, since it's data points aren't always on the road surface. You can see where the red line dips when the road crosses a small creek on a bridge. (The strava option to substitute map elevations for the device readings can have the same problems.)

The blue graph is bike speed. I have a magnet wheel sensor on the back wheel. The smoothing looks pretty good here, no sharp spikes.

The yellow graph is grade percentage. It's probably more variable than the actual road, which has changes in grade, but no "bumps" or sharp grade changes or switchbacks.




Here's the smoothing settings. (No, I don't know the theory of this.)



~~~~~~~~~~~~~~~~~

And here's the same graphs with smoothing turned off!
Totally unusable.
Elevation is reported by the Garmin, and it's smoothed while recording. (I can tell, because there's a 6 second grade% delay before a sudden, steep climb, compared to the power and cadence data.)
Note that the wheel sensor is reasonably accurate, no wild swings in bike speed. It's interesting that even these relatively small speed changes produce such a wide swing in the grade percentage. I'd think it would use the already smoothed elevation to calculate grade, not needing the speed data?



~~~~~~~~~~~~~

There's also a simple averaging method, here with a 6-second rolling average. Better than raw data, but still way off.



~~~~~~~~~~~

The raw data zoomed in with no smoothing, to see each one-second data point. Each corner of the blue graph is the next data point. It does seem to have some alternating fast-slow wheel sensor speeds, due to slight changes in the exact timing of where the magnet actually passed by the sensor?




~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

...

GPS location accuracy
GPS can be quite inaccurate with tree cover, with rock hillsides nearby, with reflections of the signals off taller buildings, etc.

But here, out in the open, it's precise.
Each dot is another one-second recording position. Color coded by speed.
See the scale at bottom right, 0.05 miles. That's 264 feet / 80 meters.
I did a rolling stop crossing the highway, no traffic, slowed to see if the other riders got across, then headed downhill to the right side of of the map.

njkayaker

fooferdoggie 

I need one because I go underground and the GPS keeps pumping the same speed so when I end a ride the gps auto starts again. its really annoying.

rebayona

I followed your suggestions: I deleted the sensor, re-paired it and started a ride indoors, spun the wheel and the reading was quite accurate. I can confirm the sensor is doing its job. Not sure if I should rely on the automatic calculation or if I should manually enter the wheel's size though.

That was very useful, thank you. I loved the app. I wonder if it is possible to trim tours in order to generate Garmin courses based on this data?

I tried it and it shows very interesting data indeed:

Still doesn't explain why the reading real-time on the bike reading it from the Garmin's screen at a constant pace sometimes is erratic, but I guess the re-pairing may fix it. I'll know the next ride tomorrow morning if the weather helps.

Thank you!

njkayaker

The Edges (like your 530) use track files for navigation. You won't get "course points" but you can get "turn guidance" (the big white arrows) using any track file (either a ride you recorded or a synthetic track created using route planner software).

There is various software that will let you edit a track (I don't know if My Tour Book lets you do that). It's also not that hard to click along the roads to create a track (so, there often isn't that much need to use an actual track).

===============================

I believe the speed from the wheel sensor and that calculated by GPS are both saved to the FIT file.

The wheel sensor speed will be better than the GPS speed. The wheel sensor is actually a bit quicker to react too. GPS calculates the speed based on change in location over time (which means it's less accurate at low speeds).

Golden Cheetah is a program that might let you see that.

It won't. The point of pairing is to tell the head unit (the 530) what id number your sensor is using (nothing else).

unterhausen

ride with gps will take your ride and make it into a route which can then be used on your garmin. I have done it a couple of times. You should be able to change the route after it's converted, not sure about before that.

JohnJ80

a couple of things - usually and in general your bike gps updates its position around once a second. The wheel sensor is a lot faster than that so it tends to both a lot more accurate and a lot smoother.

all of this bike wireless network stuff ultimately has to rely on the black magic of RF and that gets into parts of the device that are difficult to test.

presuming you’ve updated the batteries, it’s a lot more likely that your having a problem with either the sensor or the computers radio (more likely the sensor).

I'm inferring here since you say it’s a back wheel sensor so I’m guessing it’s the old style with a magnet pickup on the wheels and probably on the crank for cadence. These were known for being difficult for two reasons: first, the air gap from the magnet to the sensor is critical. If that isn’t perfect you may have an intermittent pickup from the magnet. Secondly, the Bluetooth sensor is fairly far from the computer and if the sensor has a weak radio it may be not a great connection to the computer. This is doubly true if your Garmin computer has a weak receiver (although they tend to be pretty good).

You can sort of test this by spinning the bike wheel on a bike stand and taking the computer off the bike and walking a distance away keeping an eye on indicators that show the wheel spinning or the sensor connected. With some weak sensors, you’re lucky to get even 4-6’ away. For a good sensor you can get 15’ or more.

if I’ve guessed your sensor correctly, I’d advise dumping it. Those were tricky and inconsistent and they haven’t been built for a while. Newer sensors rely on a switch in the sensor that trips every revolution and no magnet is required which eliminates the other problem. The newer sensors since they’re not magnet driven work very well on the front hub where they’re easier to get at and they are directly in line with the computer with nothing in the way to disturb it’s tiny radio.

If I’ve guessed the last sensor wrong, then it could be you just have a lousy sensor. You can test this by putting another sensor in its place and seeing if that works fine.

I went through a similar deal with trying to get my Stages gen2 PM to work with my Hammerhead K1. I spent hours and hours with both Hammerhead and finally stages. Turns out the gen2 Stages PMs have a very wimpy radio that is fixed when they (for free) upgraded my PMs all to gen3. At gen2, the PM could barely reach and then intermittently the handlebars. With Gen3, I could connect 20’ away. Given I’m an electrical engineer, this drive me completely nuts and I wound up with my being on a first name basis with both Stages and HH tech support over an entire summer.

in my case, It was not practical to swap out power meters for another one to verify if it was the PM. Fortunately for you, it is easy and practical to swap out wheel and cadence sensors. Give that try and if it doesn’t work, then borrow a bike computer and see if the problem is inherent in that computer. highly likely that’s where the issue lies and not deep in the weeds of smoothing and sampling theory (which is actually part of my engineering specialty). Almost always with all my tinkering on this stuff, it’s first off a battery issue. If it’s not that then it’s a sensor/Rf issue. If it’s not that, then it’s a bad receiver in the computer.

good luck. I’ve been down the data integrity rathole too. I feel for you.

here’s a good article in gps accuracy on bikes.
https://www.bike-mag.com/how-your-gps-lies-to-you

rebayona

So I followed your suggestion and tried Golden Cheetah (love the game by the way)
I can see this:

So the speed sensor is there!
And also this:

I can't tell what is the speed source here but within one second: same cadence, constant slope, same HR... how come the speed can change that much within seconds?

Any ideas on how to find out what is going on here?

It is located at the front wheel, Garmin speed sensor so no magnet on wheel. It is the 010-12104-00 to be precise. No BT but ANT+ for these two as well.

However, your concerns about the sensor make complete sense. I'll keep an eye on it during future rides on that particular route. I'm still looking for a way to know what was the source for those numbers, sensor or GPS?

Thanks for your reply, it was very informative.

njkayaker

The speed is from the sensor (if there's a sensor). I haven't looked at Golden Cheetah in a while (and I'm not running a sensor).

The distance is from the sensor (calculated from the wheel rotations). The distance also shows a drop.

With digital sensors, the failure mode is losing rotations/counts, which means the speed and distance will be lower than actual when things aren't working properly.

One could calculate the speed/distance from the GPS coordinates to see how they compare.