Wrong power value

Hi, my buddy noticed that Intervals, shows way much higher power values than Garmin Connect and Strava for the same activity.

Here is an example where it’s noticeable - Intervals.icu

20m best:
Intervals - 320w avg. Strava - 296w avg, Garmin Connect - 299w avg

57:06 - 1:02:10
Intervals - 347w avg, 547w max. Strava - 304w avg, 561w max

Whole activity
Intervals - 278w avg, 301w NP, Strava - 284w avg, Garmin Connect - 284w, 292w NP

I can’t show comparison with all Garmin Connect values, but by his opinion it’s much closer to Strava values than Intervals.

Can you explain, why it’s shows that way?

He has Assioma Uno power meter on left pedal.

I don’t understand this one … I downloaded the streams CSV from the activity data page. The best 20m power is from index 873 to 2073 (1200 seconds) which comes to approx 320w. Thats a very simple sum of the watts and divide by 1200s. You can check it yourself easily.

How do Strava and Garmin have such different values?

I downloaded CSV file from Intervals and everything looks fine there, just as you said.
I converted the original FIT file (which I downloaded from Intervals) for this activity to CSV with two different tools: Convert FIT Files to CSV, and GoldenCheetah. Then calculated some averages.

Rows 873 - 2073
Intervals: 320w avg
Gotoes: 296w avg
GoldenCheetah: 296w

Rows 2000 - 2500
Intervals: 263w avg
Gotoes: 295w avg
GoldenCheetah: 295w

Rows 2500 - 3000
Intervals: 265w avg
Gotoes: 293w avg
GoldenCheetah: 293w

As you can see there is some difference. Each row with same index in each file represents same point, checked it, these values must represent same period of time.

So, it seems on Intervals power values parsed differently from FIT file. I checked HR values, they seems OK.

I can’t upload here CSV files where I did my calculations, forum won’t allow me. I uploaded them to dropbox, it’s available here Dropbox - wrong-power-value - Simplify your life

2 Likes

Tx for investigating this. That file has an extra power field “C Power” and a subtle bug was causing Intervals.icu to use that instead of “power”.

The bug would only trigger if the first record had “null” (65535) for “C Power”. So it should be very rare. If you spot any more such activities you can do Actions → Reprocess File to fix them.

1 Like

curious what you’re using paired to the Garmin that has “C Power”?

You mean power meter? Favero Assioma Uno, signle side power meter on the left pedal

No… the one which David mentioned that there is a “C Power” data in the stream. (It’s not in the CSV cos like david mentioned, intervals.icu didn’t pull from the normal power stream, it pulled from “C Power”)

So, the (org) FIT file has 2 power numbers/columns… “C Power” and “POwer”

I’m curious what is “C Power”.
YOur Assioma should be paired and using the “Power” column

I found another topic which related to “C Power” Reload from garmin failed - #4 by rus89
Seems “C Power” is Calculated Power.

Thank you very much, for quick fix.

I think you’re using a 3rd Party Garmin CIQ Data Field which generates these data.
These are “developer fields” in garmin speak.

Yes, you are right. He’s using some Connect IQ apps, dunno which. Up until some time everything was fine.
And now it’s got back to normal :slight_smile:
Thanks for the help

1 Like

I’ve just noticed strange behaviour for Time in Zones. Garmin Connect and Golden Cheetah are calculating different amounts than intervals.

GC example:
image

Intervals:
image

check the boundaries?
rough look, the overall total time looks different as well.

Intervals.icu uses a 5s moving average of power for time in zones calculation. Maybe it is that?