Power readings sent to Garmin

[KenS]

When the Newton auto corrects its tilt or after things like a cal ride, tilt correction etc... with the GT key it sends a huge instantaneous power reading to the Garmin which messes up the average power and graphs in Training Center. By huge I mean 30000, 40000 even 60000+ Watts.
Is it possible to suppress this? It's especially a problem with the autotilt correction as it can happen anywhere in a ride.
This is actually something that always happened with my GenIII as well.

[GeirSkog]

Ken

I just bought my first iBike, a Newton+ with the GT-key preinstalled so that I can use it with a Garmin Edge 800.
The iBike hooked up beautifully with the Garmin sensors and the Garmin 800, producing very reliable info after a few cal rides and tilt cals.

However I am experiencing the exact same problem that you describe.
One 50000 watt spike (due to a wind cal) in yesterdays 75 minute ride is enough to seriously distort the power calculations in Garmin Connect.
The 20 min Max avg power in Isaac for the same ride is 310 W, here are the numbers from Garmin connect, as you can see the rest of the numbers are even more unlikely:

Max Power: 47.929 W
Max Avg Power (20 min): 349 W
Normalized Power (NP): 581 W
Intensity Factor (IF): 2,419
Training Stress Score (TSS): 557,8

Hopefully the next firmware upgrade will include a filter function that stops these spikes from being transmitted via the Pout ANT+ connection to Garmin units.
Manually deleting the spikes from the files is no solution for those who prefer to be more on the bike than on the computer....
Is there any other solution to this problem?

[racerfern]

The only time I've seen these crazy high numbers is if I enter the setup screen, for example to view wattage in the coast screen while rolling. I do get lots of dropped data with Garmin units and I blame Garmin not iBike for this.

I don't see where auto tilt could affect power output numbers because then they would also exist in the iBike file. The same goes for wind cal. When and perhaps more importantly why are you doing a wind cal during a ride? Why are you doing a wind cal at all with a Newton? Also, if you do a wind cal, you should be completely stopped and no power numbers should be transmitting. Are you starting off after doing a wind cal before exiting the setup screen?

Also please post the ibr file and the fit file.

[GeirSkog]

The only time I've seen these crazy high numbers is if I enter the setup screen, for example to view wattage in the coast screen while rolling.

You're right, this has do do with setup functions. I have again checked the 3 rides I have done with the Garmin and the power spikes (varying from 2500 to 50000 watts) are produced by the Newton on these occasions:
1. ride: I did 2 cal rides with some more riding before, between and after the cal rides. I stopped the bike before entering setup to start the 2 cal rides, I didn't stop at the end of the cal rides. The Newton stores this as 5 separate ibr-files. The Garmin stores the whole ride in one file. The garmin shows 4 power spikes, each of them corresponding exactly to the start and end of the cal rides.
2. ride: Isaac said bad tilt, did a new tilt cal. Had to do 2 cal rides because Newton said the first was bad. Anyway, the exact same pattern shows up in the garmin file, 4 power spikes corresponding to the start and end of the cal rides.
3. ride: warm up, then stopped to do a wind cal immediately before starting a fit test. The garmin file again shows a power spike at the start and end of the fit test. I think it is the fit test and not the wind cal that produces the spikes.
Conclusion: It seems that the Newton transmits power spikes every time one ibr file is ended and the next starts when doing cal rides and fit tests. This is probably why the spikes don't show up in the ibr files. The spikes are produced regardless of whether the cadence is zero or not.

I do get lots of dropped data with Garmin units and I blame Garmin not iBike for this.

I have checked some parts of the garmin files omitting the spikes, comparing it to the same stretch in Isaac, the average watts in these parts of the garmin files are the same as in Isaac, so no dropped data in my opinion. Anyway, dropped data is not a good excuse for accepting power spikes that make the affected garmin files very inaccurate. The GT-key is a paid for iBike-option designed to deliver powerdata to garmin units and I dont think it is unreasonable to expect proper data.

I don't see where auto tilt could affect power output numbers because then they would also exist in the iBike file. The same goes for wind cal.

Good point, therefore starting the fit test is probably the cause in ride 3.

When and perhaps more importantly why are you doing a wind cal during a ride? Why are you doing a wind cal at all with a Newton? Also, if you do a wind cal, you should be completely stopped and no power numbers should be transmitting.

Being a newbie to iBike I was eager to have a proper fit test, my 3 first cal rides were not good and I didn't get the tilt correct the first 3 times i tried, so reading the forum gave me the impression that a wind cal could improve the data.

Are you starting off after doing a wind cal before exiting the setup screen?

Yes, In 3.ride I was standing still, did a wind cal, started pedaling still in set up and then started the fit test.

Also please post the ibr file and the fit file.

iBike_10_21_2012_FTP_test.ibr

Merged 3 ibr files

(409.8 KiB) Downloaded 343 times

[GeirSkog]

And here is the fit file.

[racerfern]

So the bottom line is that for regular riding, intervals, etc. there are no spurious wattage numbers sent to Garmin. Unless you stop to do a wind cal or tilt cal both of which are not needed once you have completed your cal ride, you won't see numbers that shoot out the top of the graph. Even then, if you come to a complete stop and enter and exit the setup screen without riding off you still won't see any crazy numbers.

It would be nice if these numbers weren't broadcast, however for all that Newton does, if we just train/ride without worrying too much about tilt and wind the numbers will be good.

[GeirSkog]

Good to know that the spikes are mostly related to cal rides and maybe to other set up procedures that I don't need to do very often, now that I hopefully have a good cal ride.

However, if the power spikes are sent by Newton to Garmin every time I (and others?) do a fit test, I think it is a good idea for velocomp to look into the GT-key firmware and see if it is possible to supress these spikes. A fit test is a procedure most of us will repeat at regular intervals. I have checked that the average power calculations, TSS and IF also get skewed in TrainingPeaks when a garmin fit-file (fit test) with spikes is uploaded, whereas Golden Cheetah has a "spike supression algorithm" that seems to work rather well.