Garmin cadence sensor

chatlow

Got the garmin 800 bundle but have never bothered with the cadence unit. About to fit it to my Scott bike but the crank I've got isn't flat (concaved in the middle) which means a cable tie can't really be used to secure the pedal sensor. The sticky back will prob be ok but any ideas on fully securing it?

Also, can the cadence sensor be used for just speed and leave the pedal sensor off the bike?

chippyk

It can be used without the cadence sensor. I have Rotor cranks with a concave inside face and the cable tie does the job. I'd imagine the adhesive might eventually fail in the wet

chris_bass

~I use something like this

http://www.ebay.co.uk/itm/X-2-HIDDEN-ST ... 233f06bb27

it goes on the pedal axle, should do the job.

but you can use the sensor without cadence and just for speed.

chatlow

thanks. I've managed to stick it on a flat section at the end of the crank arm. Other than a few auto pauses and speed issues at first, it now works ok after calibrating.

Have noticed that when I turn the garmin on it is showing a speed of 0.1- 0.3 mph when the bike is stationary. If the sensor isn't moving how is that possible? Was under the impression that the unit will only use gps for speed when no cadence/speed sensor available.

Also - and it might be a coincidence, but I've just done my first ride with the cadence sensor and the activity is showing me riding at 1am this morning! I did turn it on at 1am this morning but it was turned and left on to get signal outside for a minute before I started my ride this evening, so don't know how this has happened.

chatlow

wheel size has also been automatically set to 2068.. i'm on a 700 x 23 which should be 2096. meh.

wongataa

chatlow wrote:

Have noticed that when I turn the garmin on it is showing a speed of 0.1- 0.3 mph when the bike is stationary. If the sensor isn't moving how is that possible? Was under the impression that the unit will only use gps for speed when no cadence/speed sensor available.

That's GPS position drift for you. When you start moving the speed sensor data will be used instead of GPS data for speeds and so that drift will be gone.

chatlow wrote:

Also - and it might be a coincidence, but I've just done my first ride with the cadence sensor and the activity is showing me riding at 1am this morning! I did turn it on at 1am this morning but it was turned and left on to get signal outside for a minute before I started my ride this evening, so don't know how this has happened.

When turned on the device starts logging. To prevent this you needed to reset it to start a new log at whatever time you started.

chatlow wrote:

wheel size has also been automatically set to 2068.. i'm on a 700 x 23 which should be 2096. meh.

I set my wheelsize manually. I don't trust the device to reliably keep calculating the correct size. Just work out your wheelsize and enter the number. Job done and no need to do it again unless you change something.

chatlow

Thanks for the answers. Have just set the wheel size manually and will make sure it's all reset before any rides.

Cheers

ai_1

I don't see why you'd manually set the circumference.
What makes you think the automatic value was less accurate?
Did you measure the 2096mm value, calculate it or get it from some chart? If you measured it, accurately, then the difference is a little more than I might have expected but still only <1.5%.... and are you sure any or all the error is on the Garmin's side?
Any value from a chart etc will also have error even if it specified the rim width, tyre brand, model and size, tyre pressure and bike+rider weight and weight distribution. I've never seen one that did. The Garmin is simply calculating circumference by duviding dustance travelled by wheel revolutions. Unless you're going around in circles so that the GPS distance figure us highly inaccurate the Garmin value should be more accurate than a manual measurement and far more representative than any value not based on direct measurement.

Bar Shaker

The unit auto calibrates wheel size very accurately. It will look at the distance travelled and number of revolutions over a long distance and is always fine tuning the result.

It also only looks at the wheel speed sensor when it loses GPS accuracy, such as when you ride under heavy tree cover. If you want test this, take off your spoke magnet but leave on you crank magnet. A typical 85 cadence while show up as 14mph whilst you are doing 20mph under the trees. The minute you come out to clear skies, your speed will be corrected. The primary source of info is always the GPS time stamps.

The other thing to consider with accuracy is that the tiny positional errors that show you moving at 0.2mph are the same when you are 20 miles away. If the I accuracy is say 600mm it is also that at your destination. Over 100,000,000mm that I accuracy is almost immeasurably small and does not equate to a 0.2mph average on your whole ride.

ai_1

Bar Shaker wrote:

The unit auto calibrates wheel size very accurately. It will look at the distance travelled and number of revolutions over a long distance and is always fine tuning the result.

It also only looks at the wheel speed sensor when it loses GPS accuracy, such as when you ride under heavy tree cover. If you want test this, take off your spoke magnet but leave on you crank magnet. A typical 85 cadence while show up as 14mph whilst you are doing 20mph under the trees. The minute you come out to clear skies, your speed will be corrected. The primary source of info is always the GPS time stamps.

Agreed on the wheel size calibration. Stick with auto.

However, displayed speed is from the sensor not the GPS. I think logged speed may be GPS based but the displayed value is from the sensor and is much more responsive and accurate. Over short distances GPS based speed values are poor due to the relatively large location error on consumer devices. The total distance and average speed will be accurate, thus the recommendation to use the auto wheel calibration but it's not good for calculating instantaneous speed values.

chatlow

okay - well I am using a standard Vittoria 700c x 23 tyre on a Giant wheel, pumped to 100psi. Looking at the Garmin size table it should have selected a 2096 size, but like I said, it's gone with 2068 (which is 26 x 2.1). This was calculated after 5 mins into my 15 miles journey yesterday.

Quite a few people online saying better to input wheel size manually just in case GPS signal is bad (like today).

ai_1

chatlow wrote:

okay - well I am using a standard Vittoria 700c x 23 tyre on a Giant wheel, pumped to 100psi. Looking at the Garmin size table it should have selected a 2096 size, but like I said, it's gone with 2068 (which is 26 x 2.1). This was calculated after 5 mins into my 15 miles journey yesterday.

Quite a few people online saying better to input wheel size manually just in case GPS signal is bad (like today).

I stand to be corrected but I don't believe the Garmin chart is saying that "it should have selected 2096". I think 2096 is the figure they recommend as a best guess in the absence of a measurement. I presume it says this is the suggested size for a 700c 23mm tyre, and doesn't specify the brand/model/pressure/loading? there are too many variables for a simple table to cater for.
Different 23mm tyres have different shapes, i.e. some are higher, some are more low profile. 23mm refers to the width, not the height which can vary a bit. In fact even the width which is the basis of the sizing can vary significantly between tyre models so why expect a similar circumference for all 700c 23mm tyres?
The difference between 2096 and 2068 is 28mm. Ignoing compressibility, that corresponds to a radius difference of 4.46mm. I reckon that's easily explainable by the Vittoria tyres being relatively low profile compared with, say, Schwalbe Ultremos or Continental GP4000S. So tyre selection alone would explain the difference nevermind the other factors which may compound this.
So to repeat my earlier suggestion - use the automatic setting. There's no reason to expect high accuracy from the chart. The GPS based calculation should be more accurate.

Your Garmin selected a value after 5mins you say?
Assuming you weren't climbing or otherwise slowed you probably covered 2km or more in that time?
GPS location is typically accurate to 5m or better. So talking a conservative guess of 2km covered have a look at the likely accuracy assuming you weren't cycling in circles ("corner cutting" would reduce distance calc accuracy).
Distance error expected: 5/2000 = 0.25%
I assume the distance used for the calculation starts and finishes on a magnet pass and therefore there should be no error as it's a digital count.

So ideally the GPS measurement is accurate to 0.25% and even if you take a few corners and have a slightly higher location error you'll still probably be well under 1%.
This is better than I would expect from a chart unless you get lucky. Direct measurement takes all variables into account and has estimatable error. The chart is a median value. The measured data is superior unless my assumptions about the automatic calc are way out and Garmin have in fact made a balls of it - highly unlikely.

I also believe the previous poster is correct and the Garmin tweaks the value after the initial calc. In which case it may be more accurate again.

Plenty people say plenty things. Ask for a reason and then decide.

Sorry for the long winded reply. Too much effort for such a minor "issue"

chatlow

No at all. Thanks for a very detailed reply. Yes I covered around 2k I would say. I am about to head out again now as just stopped raining, so will change it to automatic wheel size, rescan for the cadence sensor and start cycling.

Thanks

Bar Shaker

Ai_1 wrote:

However, displayed speed is from the sensor not the GPS. I think logged speed may be GPS based but the displayed value is from the sensor and is much more responsive and accurate. Over short distances GPS based speed values are poor due to the relatively large location error on consumer devices. The total distance and average speed will be accurate, thus the recommendation to use the auto wheel calibration but it's not good for calculating instantaneous speed values.

Try my experiment you will see that I am right. The speed sensor picks up your crank magnet and displays the erroneous speed data only when it has insufficient satellite reception.

At all other times, the data displayed and recorded is satellite derived.

ForumNewbie

Bar Shaker wrote:

Ai_1 wrote:

However, displayed speed is from the sensor not the GPS. I think logged speed may be GPS based but the displayed value is from the sensor and is much more responsive and accurate. Over short distances GPS based speed values are poor due to the relatively large location error on consumer devices. The total distance and average speed will be accurate, thus the recommendation to use the auto wheel calibration but it's not good for calculating instantaneous speed values.

Try my experiment you will see that I am right. The speed sensor picks up your crank magnet and displays the erroneous speed data only when it has insufficient satellite reception.

At all other times, the data displayed and recorded is satellite derived.

A few weeks ago my data on the Garmin 500 display went weird over the last 3 miles of my ride. The speed was displaying as only 2 or 3 mph when I was riding on the flat at a speed of at least 15mph. The mileage ridden also went up by about half a mile in the last 3 miles. However when I downloaded the ride to Strava it actually re-calculated via the GPS data and showed on the correct distance that I had ridden and therefore worked out the correct average speed.

It turned out that the battery in the speed/cadence sensor was running out and needed replacing. Accordingly this shows that when on the bike, the Garmin display speed and distance is being calculated and displayed from the speed sensor rather than from the GPS.

Bar Shaker

So what was the satellite coverage like when this happened? Were you in a built up area? Was it very cloudy or even raining? Were there lots of trees around? Were there HV cables around?

if you ride your bike without the magnet on the spoke, what does it show?

Your Garmin calculates something called its DOP. This is based on how many satellites it can see and therefore how much confidence it has in the time stamps it is receiving. It needs a minimum of 4 satellites but ideally sees many more. The DOP is based on a circle within which most of the receiving stamps fall. Things that affect the accuracy are this points above and how close the satellite is to the horizon. When it sees 4 or less, it looks for the wheel speed data to display as it has little confidence in the satellite info. It will still record positional data when it can only see 3 satellites but the DOP will be high. You will see this as your recorded route straying away from the path you took, when you zoom in on map on Strava.

I don't doubt you got a very inaccurate display from your GSC10, if the battery was low, but I would suggest the power drain is so low (the battery will last for years) that it would have been low for the whole of your ride and you only saw the strange readings when you lost some satellite reception.

term1te

Chris Bass wrote:

~I use something like this

http://www.ebay.co.uk/itm/X-2-HIDDEN-ST ... 233f06bb27

it goes on the pedal axle, should do the job.

but you can use the sensor without cadence and just for speed.

Be careful using one of these if you have the latest 6800 Ultegra pedals. The only way to get them off is with an Allen key, which the supper strong magnet prevents you accessing. If it is flush with the crank arm, you'd be pushed to get it out.

chris_bass

Term1te wrote:

Chris Bass wrote:

~I use something like this

http://www.ebay.co.uk/itm/X-2-HIDDEN-ST ... 233f06bb27

it goes on the pedal axle, should do the job.

but you can use the sensor without cadence and just for speed.

Be careful using one of these if you have the latest 6800 Ultegra pedals. The only way to get them off is with an Allen key, which the supper strong magnet prevents you accessing. If it is flush with the crank arm, you'd be pushed to get it out.

The magnets aren't that strong (or at least the ones I have aren't, the wonders of cheap things on ebay!) and there is a slight gap so you can get a fine tipped screw driver down the side if you need to.

chris_bass

Bar Shaker wrote:

So what was the satellite coverage like when this happened? Were you in a built up area? Was it very cloudy or even raining? Were there lots of trees around? Were there HV cables around?

if you ride your bike without the magnet on the spoke, what does it show?

Your Garmin calculates something called its DOP. This is based on how many satellites it can see and therefore how much confidence it has in the time stamps it is receiving. It needs a minimum of 4 satellites but ideally sees many more. The DOP is based on a circle within which most of the receiving stamps fall. Things that affect the accuracy are this points above and how close the satellite is to the horizon. When it sees 4 or less, it looks for the wheel speed data to display as it has little confidence in the satellite info. It will still record positional data when it can only see 3 satellites but the DOP will be high. You will see this as your recorded route straying away from the path you took, when you zoom in on map on Strava.

I don't doubt you got a very inaccurate display from your GSC10, if the battery was low, but I would suggest the power drain is so low (the battery will last for years) that it would have been low for the whole of your ride and you only saw the strange readings when you lost some satellite reception.

I have been out before with the magnet set slightly off, it clipped the sensor and spun round the spoke so it wasn't being picked up, I cycled about a mile with the speed displaying as zero, GPS accuracy was fine (i have it on one of my screens). The computer uses the speed sensor when it is linked no matter on the gps signal.

Next time you are out, before you start, lift the back wheel up and spin the pedals, see what speed it displays. also, when i am on my turbo trainer, I have the gps unit near enough to the window for it to get good signal and I have forgotten to turn the GPS off and it still records the speed and distance.

rolf_f

Chris Bass wrote:

Term1te wrote:

Chris Bass wrote:

~I use something like this

http://www.ebay.co.uk/itm/X-2-HIDDEN-ST ... 233f06bb27

it goes on the pedal axle, should do the job.

but you can use the sensor without cadence and just for speed.

Be careful using one of these if you have the latest 6800 Ultegra pedals. The only way to get them off is with an Allen key, which the supper strong magnet prevents you accessing. If it is flush with the crank arm, you'd be pushed to get it out.

The magnets aren't that strong (or at least the ones I have aren't, the wonders of cheap things on ebay!) and there is a slight gap so you can get a fine tipped screw driver down the side if you need to.

Look pedals are also only removeable with the allen key - and like you, I have no problem removing the neodymium magnet. Infact, I think you can get them off with a fingernail without too much bother. You can slide them fairly easily too reducing the effective strength of the magnet.

Bar Shaker

Chris Bass wrote:

I have been out before with the magnet set slightly off, it clipped the sensor and spun round the spoke so it wasn't being picked up, I cycled about a mile with the speed displaying as zero, GPS accuracy was fine (i have it on one of my screens). The computer uses the speed sensor when it is linked no matter on the gps signal.

Next time you are out, before you start, lift the back wheel up and spin the pedals, see what speed it displays. also, when I am on my turbo trainer, I have the gps unit near enough to the window for it to get good signal and I have forgotten to turn the GPS off and it still records the speed and distance.

When I got my new carbon wheels, I rode for 3 weeks with no magnet. I always had speed displayed. If I rode into an area of reduced satellite reception, it would look to the GSC10 for this and would pick up the magnet on my crank, thinking this was the wheel speed, hence the lower speeds. Come back out from under the tree covered road and speed would return to normal. This was not a one off and was a repeatable process. I don't know what GPS you have but what you describe is not what happens with an 800 or 1000. At no time did my GPS show 0mph with no magnet fitted.

Your GPS would be lucky to see two satellites through your window let alone 4+. The GPS signal would be extremely poor. Similarly, if you spin the pedals and rotate the wheel, the Garmin will either take this input as being over riding as the GPS signal will say you aren't moving, or you will be doing it before you have a full lock onto all satellites.

The unit is a bit more advanced than a wireless Cateye with some GPS maps.

slowbike

Bar Shaker wrote:

Chris Bass wrote:

I have been out before with the magnet set slightly off, it clipped the sensor and spun round the spoke so it wasn't being picked up, I cycled about a mile with the speed displaying as zero, GPS accuracy was fine (i have it on one of my screens). The computer uses the speed sensor when it is linked no matter on the gps signal.

Next time you are out, before you start, lift the back wheel up and spin the pedals, see what speed it displays. also, when I am on my turbo trainer, I have the gps unit near enough to the window for it to get good signal and I have forgotten to turn the GPS off and it still records the speed and distance.

When I got my new carbon wheels, I rode for 3 weeks with no magnet. I always had speed displayed. If I rode into an area of reduced satellite reception, it would look to the GSC10 for this and would pick up the magnet on my crank, thinking this was the wheel speed, hence the lower speeds. Come back out from under the tree covered road and speed would return to normal. This was not a one off and was a repeatable process. I don't know what GPS you have but what you describe is not what happens with an 800 or 1000. At no time did my GPS show 0mph with no magnet fitted.

Your GPS would be lucky to see two satellites through your window let alone 4+. The GPS signal would be extremely poor. Similarly, if you spin the pedals and rotate the wheel, the Garmin will either take this input as being over riding as the GPS signal will say you aren't moving, or you will be doing it before you have a full lock onto all satellites.

The unit is a bit more advanced than a wireless Cateye with some GPS maps.

I don't think you're right here.

My understanding is that the GPS displays and records sensor speed (along with satellite derived position) - if one is sensed. If it isn't sensed (eg just the cadence magnet triggers the cadence part of the sensor) then it will display GPS speed as well as record the position.

If you have triggered the Speed sensor with a magnet, remove the magnet then go for a ride then your speed & distance display will be zero (btw I know this for a fact as I've had it happen - to get around it you have to either turn the GPS off and on or de-pair it from the speed/cadence sensor - or just fix the magnet ...)

So as far as I can tell the GPS (in my case an 800) doesn't selectively chose which source to display based on accuracy calculations.

chris_bass

This was taken from the ride where the magnet was not picked up, you can see the speed goes to zero but the cadence keeps going. I have an edge 500.

Bar Shaker

Slowbike wrote:

If you have triggered the Speed sensor with a magnet, remove the magnet then go for a ride then your speed & distance display will be zero (btw I know this for a fact as I've had it happen - to get around it you have to either turn the GPS off and on or de-pair it from the speed/cadence sensor - or just fix the magnet ...)

So as far as I can tell the GPS (in my case an 800) doesn't selectively chose which source to display based on accuracy calculations.

Mine did exactly as I describe, for 3 weeks... until I got a Noly magnet for the wheel section.

There are many factors that could make all of our experiences happen, one being what OS version we are using etc.

The new wheel speed and cadence units, with no magnets, have no such problems. Much simpler.

John.T

The GSC10 always takes priority over GPS for speed if one is linked. If the signal is lost then it will revert to GPS speed sensing. This takes a short while and zero speed will be show until it does. An intermittent GSC10 signal such as with a low battery or misplaced magnet can cause a zero most of the time. This information has been published on the Garmin forums many times.

apreading

Bar Shaker you are wrong. My GSC10 has developed a problem where the reed is sticky and occasionally it doesnt register a rotation or to - when this happens it auto-pauses and then resumes, during which time the speed drops or says zero. The mileage also does not match. This is over training routes where I never have a problem with satellite reception. Disabled the GSC10 and everything reads normal for the whole ride.

Everyone in the world, here and on other forums is quite clear - with the GSC10, once it has picked up that you have wheel sensor, it ignores GPS for speed or distance.

Bar Shaker

apreading wrote:

Bar Shaker you are wrong.

Not wrong, just had a different and repeatable experience to yours.

apreading

Bar Shaker wrote:

apreading wrote:

Bar Shaker you are wrong.

Not wrong, just had a different and repeatable experience to yours.

No, an experience which you interpreted incorrectly and then insisted on believing to the point you wont listen to all the others who know how it really works.

When I got my new carbon wheels, I rode for 3 weeks with no magnet. I always had speed displayed.

Because it didnt detect a speed sensor due to no readings coming it and therefore continued using GPS.

If I rode into an area of reduced satellite reception, it would look to the GSC10 for this and would pick up the magnet on my crank, thinking this was the wheel speed, hence the lower speeds. Come back out from under the tree covered road and speed would return to normal.

No, this wasnt picking up your crank magnet, it just wasnt accurately tracking your speed/movement due to the poor satellite reception. It must have been REALLY poor for this to happen by the way because I only get it on very rare occasions, even cycling off road through dense woodland.

At no time did my GPS show 0mph with no magnet fitted.

Because it wasnt using the speed sensor - using satellites, it will always think there is some drift, even if it has no satellites. It was obviously getting some kind of weak signal and sensing some movement but not accurately measuring it.

Sometimes the whole world isnt wrong, with you being the only one who is right. Sometimes you have to re-analyse what you interpreted and admit we are right.

Bar Shaker

You have made your mind up and are now telling me what happened on my bike, to fit how you see this. As a result, your whole post is based on incorrect assumptions because that's what they need to be to fit your assumptions.

I was surprised by the change in displayed readings and that prompted me to take note of the changes and when they occurred. I witnessed these changes over 4 rides. I know what I was looking at and knew when the display would change. As much as your post above would appear to say otherwise, you were not there.

With regards to what is recorded, speed is not recorded form the GSC10. Here is part of a log from my 800, when speed and cadence were working correctly...


<gpxtpx:TrackPointExtension>
<gpxtpx:atemp>8</gpxtpx:atemp>
<gpxtpx:cad>83</gpxtpx:cad>
</gpxtpx:TrackPointExtension>
</extensions>
</trkpt>
<trkpt lon="0.399215" lat="51.635902">
<ele>59.200000</ele>
<time>2014-03-02T10:26:45Z</time>
<extensions>
<gpxtpx:TrackPointExtension>
<gpxtpx:atemp>8</gpxtpx:atemp>
<gpxtpx:cad>83</gpxtpx:cad>
</gpxtpx:TrackPointExtension>
</extensions>
</trkpt>
<trkpt lon="0.399120" lat="51.635948">
<ele>59.600000</ele>
<time>2014-03-02T10:26:46Z</time>


This is the raw recorded data file. You will see that temp and cadence are recorded but not speed. Speed is calculated by the distance between the two points, over the time which in this case is 1 second. It is a very simple task for the GPS to work out this distance. You will note that the points are in two dimensions and riding up a hill is not correctly calculated, but that is another subject for you to think about.

On the rides with the GSC10 sensor error, my logs have no cadence recorded. The displayed speed was not recorded and the GPS point log was recorded albeit with the occasional high DOP. If you open the log in something like Adze, you can see the individual points and it is from the distance between these that the speed is derived. Speed is not recorded nor is it used to somehow determine how far away the next point may be on the planet, let alone which direction it is in.

The unit calculates its DOP by with a formula based on how many satellites it sees and where they are in the sky. When mine fell below a certain threshold, it displayed the speed which the GSC10 was erroneously picking up from my crank sensor. That is the end of it.

Believe what you will, convince yourself of what you like, but these are the facts.

apreading

Speed is recorded in the FIT file by my 800 when the GSC10 speed sensor is operational - I know because I checked it when mine was playing up. The file did not indicate whether the speed was from the sensor or the GPS though, which wasnt very helpful.

I (and we) are not making assumptions to fit how we see this, I am making assumptions that are the only logical explanation of what you saw based on the facts of how the Garmin Edge computers work with the GSC10.

I will give up now though - and suggest everyone else does the same because quite clearly the world is wrong and you are the only one who is right and wont hear otherwise... I have no interest in continuing a debate with no hope of the other party opening their mind to what they are wrong about.

slowbike

Bar Shaker wrote:

The unit calculates its DOP by with a formula based on how many satellites it sees and where they are in the sky. When mine fell below a certain threshold, it displayed the speed which the GSC10 was erroneously picking up from my crank sensor. That is the end of it.

Believe what you will, convince yourself of what you like, but these are the facts.

No - these are not the facts - these are the assumptions you have made - and our experiences and knowledge are that you're incorrect in your assumptions.

It would be possible to program the 800 to behave in the manner that you describe, but it doesn't happen that way.

I can show you the slowing down of the speed reading when satellite reception is compromised - even to the point of pausing - a repeatable demonstration - and that's on one bike with no speed or cadence sensor present (or paired).

The garmin unit will use speed sensor data (if present) to record the speed (in the fit file) and will use that to display your speed and calculate averages. If it receives duff data it still continues to use that data.
It records your position from the satellite - if speed sensor isn't present then it will use that data to show you a speed, when it gets poor reception it continues to show you a speed, but it is usually significantly lower than your actual speed - it doesn't suddenly decide to use a speed sensor ...