Training View Y-axis "Re-size" Event

[Eric Abbott]

Do I document apparent bug (#1350) here or in Github? A search of this forum reveals that one user, Greg Hilton, reported the same issue on 12/15/14. I'll attach the .csv data file here since I couldn't find a way to do that in Github. Anyone know if I can adjust something besides the wattage spike feature?

Let me know please. This is driving me nuts,

Eric

[Eric Abbott]

Edit: I also examined the data file/.csv for the time span around where the spike displayed during my training session; I didn't find any wattage anomalies. 

[Jon Escombe]

Hi Eric,

This has been reported on and off for a couple of years. I had a look at
it about a year ago (Feb 14), but never got to a root cause (and have
never been able to replicate it here).

The ANT logs I looked at then had no obvious spikes/anomalies in the
incoming data, so I think it is a GC thing.. Can you confirm, is your
speed spiking at the same time (looked like it from the charts)?

Do you build from source yourself? If so, it might be worth dumping
debug data out of the ANT_WHEELTORQUE_POWER message processing in
ANTChannel.cpp. I think an incorrect period or event count here could
produce these symptoms...

Perhaps if you can reproduce in a short workout, you could attach the
antlog.bin file for that session also.

Regards,
Jon

On 30/03/15 00:42, Eric Abbott wrote:
> Edit: I also examined the data file/.csv for the time span around where
> the spike displayed during my training session; I didn't find any
> wattage anomalies.
>
> On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:
>
> Do I document apparent bug (#1350) here or in Github

> <https://github.com/GoldenCheetah/GoldenCheetah/issues/1350>? A

[Eric Abbott]

Hey Jon,

I didn't build from source, just installed the (then) official release. I agree that the "Wspike" pic also shows a speed spike, although when I re-examined the .csv Joule GPS file attached; shows no such km/hr spike. 

This sizable wattage spike/re-size event happens occasionally, which is to say maybe 2 to 4 times a training session. I searched my HD for the file you mentioned...didn't find it. Assuming it's volatile, where does it reside during a session?

Thanks for answering. I was hoping someone would read my message.

Eric

On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:

[Jon Escombe]

Hi,

Thanks, it is consistent that the Joule file won't show the spike in
either speed or power. But if both are spiking in GC, then I think it
has to be somewhere in the processing of the WHEELTORQUE messages.

The antlog.bin file should be written to the working directory when in
train mode, but not sure where that is going to end up under Vista (it
is transient - overwritten with each new train session).

If you could email me that file (GC logs all the ANT data messages into
it), plus your json/csv file for the workout, I'll have another look at
the data.

Regards,
Jon.

On 30/03/15 14:07, Eric Abbott wrote:
> Hey Jon,
> I didn't build from source, just installed the (then) official release.
> I agree that the "Wspike" pic also shows a speed spike, although when I
> re-examined the .csv Joule GPS file attached; shows no such km/hr spike.
>
> This sizable wattage spike/re-size event happens occasionally, which is
> to say maybe 2 to 4 times a training session. I searched my HD for the
> file you mentioned...didn't find it. Assuming it's volatile, where does
> it reside during a session?
>
> Thanks for answering. I was hoping someone would read my message.
> Eric
>
> On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:
>
> Do I document apparent bug (#1350) here or in Github

> <https://github.com/GoldenCheetah/GoldenCheetah/issues/1350>? A

> search of this forum reveals that one user, Greg Hilton, reported
> the same issue on 12/15/14. I'll attach the .csv data file here
> since I couldn't find a way to do that in Github. Anyone know if I
> can adjust something besides the wattage spike feature?
>
> Let me know please. This is driving me nuts,
> Eric
>

> --
> _______________________________________________
> Golden-Cheetah-Users mailing list
> golden-che...@googlegroups.com
> http://groups.google.com/group/golden-cheetah-users?hl=en
> ---
> You received this message because you are subscribed to the Google
> Groups "golden-cheetah-users" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to golden-cheetah-u...@googlegroups.com
> <mailto:golden-cheetah-u...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

[Eric Abbott]

Thanks Jon.

My next workout is Wednesday...once I capture the spike, I'll send the files. In the meantime, I didn't submit the ANT+ devices in use for GC. Basically PT hub, HR strap, and speed sensor:

 

I was wondering how GC prioritizes between two devices sending speed data, as is my case here.

On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:

[Eric Abbott]

Jon,

Trapped a good spike this time. Here are some pics:

Before session start. Planned intervals are visible (got some ideas on this, but that's for later.)

This spike doesn't exceed the y-axis limit

Neither did this one

This spike did exceed 1,000. Note how my planned intervals got squashed upon the re-size event. Trying to follow the session graphically at this point can be difficult.

This spike occurred also

I've attached the Joule GPS .csv file and the antlog.bin file. I haven't looked at the .csv yet... gotta make dinner :-)

Am curious to learn what you find out.

Thanks man,

Eric

[Jon Escombe]

Thanks Eric, I'll aim to take a look this weekend.

Could you also email me either the GC .json file for this workout or a
.csv export from GC - I'll need to correlate the reported spikes with
the ANT data.

Regards,
Jon.

On 01/04/15 22:34, Eric Abbott wrote:
> Jon,
> Trapped a good spike this time. Here are some pics:

> ...

[Eric Abbott]

See attached...

On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:

[Jon Escombe]

Thanks Eric,

Just to make sure I understand then, the spikes don't even appear in the
recorded GC data? Just in the train view window?

I was thinking they were present in the recorded GC data, just not the
Joule export, but don't see them in that attached GC export either... So
just want to clarify they've not been edited out already?

For what it's worth, the received ANT data all looks perfectly fine. I
have some python scripts to analyse the antlog file, and it's all
exactly as it should be..

Regards,
Jon

On 02/04/15 16:12, Eric Abbott wrote:
> See attached...
>
> On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:
>
> Do I document apparent bug (#1350) here or in Github

> <https://github.com/GoldenCheetah/GoldenCheetah/issues/1350>? A

> search of this forum reveals that one user, Greg Hilton, reported
> the same issue on 12/15/14. I'll attach the .csv data file here
> since I couldn't find a way to do that in Github. Anyone know if I
> can adjust something besides the wattage spike feature?
>
> Let me know please. This is driving me nuts,
> Eric
>

[Eric Abbott]

That's correct Jon. The spike phenomena doesn't appear in any of the .csv files associated with this training session. The only place the spikes manifest themselves (afaik) is in the training (and maybe realtime view) graph/object. My hypothesis is that when the wattage plot value exceeds the Y-axis upper limit, a graph display re-size event is triggered. I have no way of testing this however other than trapping spikes for [others] to analyze.

On Sunday, March 29, 2015 at 4:09:14 PM UTC-7, Eric Abbott wrote:

[Jon Escombe]

Thanks for confirming.. but I'm truly baffled ;)

You are right that the graph is automatically resized when the power
value exceeds the upper limit, which is why it zooms so far out. But
I've as yet no clue where those rogue values are coming from!

Regards,
Jon.


On 02/04/15 18:27, Eric Abbott wrote:
> That's correct Jon. The spike phenomena doesn't appear in /any/ of the

[Jon Escombe]

Hi,

Afraid I'm still no wiser on this one..

To recap, I believe the received ANT data is clean, with no power spikes
or otherwise unexpected counts or values.

I also can't see any issues with the code that processes these ANT
messages and updates the telemetry. That said, if I were able to
recreate the issue I would add some debug logging to the message
handling in ANTChannel::broadcastEvent() just to be sure.

Dunno where to go from here if I'm honest. I don't have a Windows build
environment, else perhaps putting a test build together with that
additional console output included would be the way forward..

Regards,
Jon