Garmin 800 speed spikes

theBlob

Bokeh master
Sep 28, 2011
2,906
1,520
129
...
As much as I would like it to be true, lately I have been hitting top speeds of 100km/hr on my commute!:cool::cool:

http://connect.garmin.com/activity/148180619


What are these speed spikes caused from and how do I fix it? OR should I not and just start racing while my garmin has me in such good form:D
 

AlanW

Maximum Pace
Jan 30, 2007
1,214
438
103
Tokyo
Do you have the speed sensor?

I regularly get maximum speeds of over 500km/h reported. It's from when the GPS loses signal (e.g. in a tunnel) then regains signal the other side. It jumps from one location to the next immediately (or at least, in one sampling cycle) giving HUUUUGE SPEEED!
Naomi-san has the speed sensor for her 705 and doesn't get the speed spikes.
 

joewein

Maximum Pace
Oct 25, 2011
2,574
1,236
133
Setagaya, Tokyo
joewein.net
It's probably a little different from that, because if it simply took the last accurate GPS data point before the tunnel and then the next accurate GPS data point after the tunnel, there's not just a huge distance jump but also a huge time jump, so the extrapolated speed within the tunnel should still work out as a reasonable average.

What seems to happen instead is that it needs signals from multiple satellites for tracking, which come in at certain intervals. The more satellites, the more precise the location. If after the tunnel it only acquires the minimum number of satellites (I think 3) to give a data point it may be off by quite a margin. Then when it gets more satellites, it will jump to the more correct location. It's this jump between a rough position and a precise one that is likely to be the culprit.

The better the GPS, the smarter the algorithms behind the position calculations. This is the real difference between low end and high end GPS equipment. So take GPS based speed calculations with a pinch of salt.

The same goes for power calculations: When I joined the TCC ride on Miura, according to Strava I got way better peak power output than one of the fellow riders whose rides I'm following on Strava, but checking the location I could see we were also going through some tunnels at the time, so basically the data is garbage :p
 

AlanW

Maximum Pace
Jan 30, 2007
1,214
438
103
Tokyo
Well, I haven't been into the GPX file to see exactly what's happening....

But, for Garmin units, the GPS signal lost alarm is only displayed after a few hundred metres of tunnel. I don't know if the clock continues to run from the time entering the tunnel until the signal lost alarm. If it does, but the unit assumes the position at the start of the tunnel, then regains signal, that would give a high speed reading.
I've certainly noticed that for rides with no tunnels, the speed spikes and super fast max speeds don't appear.

But I agree with your conclusion - data around tunnels is useless!
 

joewein

Maximum Pace
Oct 25, 2011
2,574
1,236
133
Setagaya, Tokyo
joewein.net
I don't know if the clock continues to run from the time entering the tunnel until the signal lost alarm.
Each GPS data packet from each satellite basically contains a clock reading. The position is calculated from minute time differences between packets from different satellites. So the answer is, yes the clock keeps running because it's actually up in orbit.

This is what datapoints look like in a GPX file (which is actually an XML file):

Code:
 <trkpt lat="35.44481317476449" lon="139.65036938004468">
  <ele>15.601366711097453</ele> 
  <time>2012-02-05T08:16:23Z</time> 
  </trkpt>
 <trkpt lat="35.44472091302681" lon="139.65054342157114">
  <ele>15.236774509090463</ele> 
  <time>2012-02-05T08:16:26Z</time> 
  </trkpt>
With each longitude and latitude reading there's a time stamp in UTC.
 

theBlob

Bokeh master
Sep 28, 2011
2,906
1,520
129
...
W I thought I was running off the sensor on the wheel. And even so there were no tunnels!!
 

trad

Maximum Pace
Dec 4, 2006
393
30
48
Tokyo
may be a different problem. I have a speed/cadence sensor and i had errant speed spikes recently on my 800. changed batteries on the cadence/speed sensor, reset everything but problem persisted. Initially, sensor and cadence alignment seemed all OK, but it turns out that the magnet on the speed sensor was about a mm out of alignment - so the unit was reading both GPS speed and sensor speed. Reset magnet position and problem went away.

also, before I got the cadence/speed sensor I also got errant speeds - when biking under the Shuto Highway (i.e, with poor sat coverage). Speeds would dramatically drop and (when it lost signal) and then do a "hurry up and catch up" spike when it found sat signal.
 

StuInTokyo

Maximum Pace
Dec 3, 2010
1,662
62
78
may be a different problem. I have a speed/cadence sensor and i had errant speed spikes recently on my 800. changed batteries on the cadence/speed sensor, reset everything but problem persisted. Initially, sensor and cadence alignment seemed all OK, but it turns out that the magnet on the speed sensor was about a mm out of alignment - so the unit was reading both GPS speed and sensor speed. Reset magnet position and problem went away.........

I had the same problem on my Polar CS200 Cad, and I found that my sensor and magnets were out by just a little bit. Sometimes on longer rides I'd get speeds of 120km/h!!

Joe's explination on the GPS problem is great!:D