Re: [Golden-Cheetah-Users] power & speed spike when using "train" mode

[Jon Escombe]

I've not seen spikes like that, but have had some odd results when
paired with both my PowerTap wheel and the Garmin Speed/Cadence sensor
(Linux, USB2 stick). I think GC is receiving speed and cadence data from
both sensors which may confuse it a little.

In the end I paired with just the PowerTap, and live with the glitchy
cadence data at low power outputs. Would perhaps be preferable if GC
could use the same approach as the Garmin head units, which I believe
take speed & cadence from the dedicated sensor (if available) in
preference to the PowerTap data.

Regards,
Jon

On 08/03/13 14:51, gh...@sbcglobal.net wrote:
> I used TRAIN mode for the first time the other night. I am using a
> Power Tap wheel, ANT+ HR strap, Garmin Cadence / speed sensor with USB1
> stick on a Windows 7 system. I was able to Pair all of the devices. I
> picked the 70-75erg workout as a test and it seemed to be working fine
> but I was getting somewhat regular Power spikes (sometimes all the way
> to 5000+ watts) and Speed spikes (to 400+ mph).

[Mike Patton]

I think I had something similar with Train mode, except I'm using a Garmin GSC-10 spd/cad sensor and a Quarq powermeter. I don't have a magnet on my cranks for a dedicated cadence sensor, so I'm pretty sure the computer was displaying alternating zero values and the cadence reported by the Quarq. When I disabled the speed/cad sensor in my device options, the cadence worked perfectly and matched my bike computer's readings.

I would prefer if the user were able to specify the preferred source for redundant signals, rather than have GC make the decision for me. The Wahoo Fitness App does this nicely.

Mike

[Greg Hoch]

Thanks for the reply. I was able to do a quick test after turning off
the Speed / Cadence sensor and that seems to have fixed the power /
speed spikes. As mentioned the cadence is now incorrect (especially
at low power) but I can live with that more than the spikes. I will
ride a complete workout tonight to make sure that was the issue but I
believe it was. I agree that it would be nice to have the ability to
specify the device to use if multiple devices are sending a signal or
to just take the cadence from the sensor if it is available.

> > to 5000+ watts) and Speed spikes (to 400+ mph).- Hide quoted text -
>
> - Show quoted text -

[Jon Escombe]

I've raised a feature request (#526) for this on GitHub

Regards,
Jon

[Liam]

I'm seeing the same problem with a Power Tap wheel and Garmin HR strap. Power spikes of 25,000 Watts! It means my averages are way off too!

Is there a workaround for this?

[Liam]

I'm seeing the same problem with a Power Tap wheel and Garmin HR strap. Power spikes of 25,000 Watts! It means my averages are way off too!

Is there a workaround for this?

On Sunday, 24 March 2013 16:16:06 UTC, Jon Escombe wrote:

[Jon Escombe]

Hi, you should be able to fix the ride file afterwards. Select the ride
in question, and then tools -> fix power spikes..

Regards,
Jon

[Liam]

Thanks Jon. I am aware of that but it doesn't help when I'm in the middle of a turbo session. The spikes are so big the Realtime Chart becomes useless (e.g. if I have a power spike of 25,000W 5 minutes into a workout then the blue power target line becomes almost invisible as it's so close to the x-axis). I hope that makes sense.

I've been using TrainerRoad for a few year and love it and I was hoping GC could replace TrainerRoad for me ... and it's so close to doing that. I've changed trainer-layout.xml to completely customize the layout and dials to suit me - in fact, I now much prefer the layout of my GC in comparison to TrainerRoad. I've figured out what each dataSeries integer represents and also the heightFactor and widthFactor. So now I have a screen that looks exactly as I want it to.

[Jon Escombe]

If you could send me the antlog.bin file from a session like this, I'll have a look at the data and see what's actually come from the power meter..

Thanks,
Jon

[Liam Frawley]

Thanks for the reply Jon - I really appreciate it.

I'm in work now so send on the file later on.

[Mark Liversedge]

Its a bug. Will look to fix it.

[Chris C.]

Hey Liam

Would you mind sharing your layout? I tried playing a bit with it but I still had issue with the heightFactor and widthFactor....

Thanks
Chris

[Liam]

Thanks Mark and Jon.

I've done one hour on the trainer and I've uploaded 3 files here:

https://drive.google.com/file/d/0B3GeobV1X2udYjZQVUIzYS1WbG8/edit?usp=sharing

antlog.bin

2014_02_12_18_49_12.csv - create by GC

2014-02-12-18-48-51.fit - Garmin Edge 500 capture of the same workout

There were some serious power spikes in this workout.

Is there an ETA for a fix?

[Liam]

Hi Chris,

I've included a screen shot below of what my GC looks like in Train Mode.

widthFactor and heightFactor are simple enough concepts. If widthFactor=8, then it will be one eighth the width of the screen and you will fit 8 dials across the screen. If widthFactor=2 you will fit 2 dial across the screen ... etc. Same applies for heightFactor.

The XML can be found here

[Mark Liversedge]

You can drag and resize too! You don't need to manipulate the settings, just drag the corners of the charts!!

[Liam Frawley]

Thanks Mark - I hadn't noticed that!

[Jon Escombe]

Thanks Liam,

I don't see anything wrong with the received ANT data that should cause
those spikes, all the values in the standard power and standard wheel
torque data pages look good.

There are the massive spikes in speed too, so I'd say it has to be
coming from the wheel torque message processing, but am a bit stumped atm.

Cheers,
Jon

[Liam Frawley]

Thanks for looking into this Jon - I appreciate it.

I was kinda hoping you'd find the "smoking gun" when you looked through the log files.

Is the "wheel torque message processing" in the GC code base or is part of the ANT+ driver?

[Jon Escombe]

Me too, but nothing smoking anywhere that I can see ;)

It's in the GC code, there are two types of power message from the
PowerTap, but the logged ANT data looks perfectly fine for both.

The wheel torque messages are where GC calculates the speed from too -
which makes me think the problem lies there - but I just can't see it..

Cheers,
Jon

On 12/02/14 23:15, Liam Frawley wrote:
> Thanks for looking into this Jon - I appreciate it.
> I was kinda hoping you'd find the "smoking gun" when you looked through
> the log files.
> Is the "wheel torque message processing" in the GC code base or is part
> of the ANT+ driver?
>
>
> On 12 February 2014 23:05, Jon Escombe <jesc...@googlemail.com

> <mailto:jesc...@googlemail.com>> wrote:
>
> Thanks Liam,
>
> I don't see anything wrong with the received ANT data that should
> cause those spikes, all the values in the standard power and
> standard wheel torque data pages look good.
>
> There are the massive spikes in speed too, so I'd say it has to be
> coming from the wheel torque message processing, but am a bit
> stumped atm.
>
> Cheers,
> Jon
>
>
>
> On 12/02/14 20:40, Liam wrote:
>
> Thanks Mark and Jon.
>
> I've done one hour on the trainer and I've uploaded 3 files here:

> https://drive.google.com/file/__d/__0B3GeobV1X2udYjZQVUIzYS1WbG8/__edit?usp=sharing

[Peter CH]

Has this issue been solved in the latest build?

Yesterday I used the train mode for the first time with my Powertap G3 and Garmin ANT+ usb stick. It works great most of the time but there were occasional peaks in power...  My Edge 810 registered no peaks at all. So I guess the problem should be in the signal-strength (laptop in front of bike) or .. . ?

Could it be related to using the powertap data for power, speed and cadance?

I'm using GC 3.3, Windows 10, Garmin ANT+ USB2, Garmin heartbeat  and Powertap G3.

Thanks for your help!

Op donderdag 13 februari 2014 00:24:06 UTC+1 schreef Jon Escombe:

[Jon Escombe]

Hi Peter,

I don't believe it's solved, as there was another recent report of the
same spikes with a G3 hub.

I've looked at the ANT data for a couple of previous reports, but
nothing jumped out. If you're able to recreate the issue and can email
me the .json and associated antlog.bin file for the workout I'd happily
look again..

Regards,
Jon

[James Manners]

Hi,

I can confirm I have been seeing regular data spikes. The frustration is how it throws out the train graph when it occurs. As per one of the comments above, I use a Garmin 510 at the same time and there are now spikes registered in that data.

Only one example is below:

{ "SECS":1541, "KM":64.129, "WATTS":92573, "CAD":75, "KPH":9861.66, "HR":144, "TEMP":0 },

The activity .json file is here https://www.dropbox.com/s/3cfl2bvfd1a12f5/20160114%20GC%20Dud%20Data.json?dl=0

I have seen these on a variety of bike setups:

• Garmin GSC10, PowerTap Pro Hub, Garmin Heart Rate Monitor & Suunto Movestick Mini
  
• The same without the GSC10.
  

I'm running OS X 10.10.5, GoldenCheetah 3.3 

Unfortunately, I can't find the location of the antlog.bin to attach as well. I'm on a mac, 10.10. Where do I find this log?

Cheers,

James

[Jon Escombe]

Thanks James,

That's pretty much the exact same bike setup as me..

Up to and including 3.3, the ant log file is stored in the working
directory of the GC process. Not sure what this would be on a Mac
though? I generally run GC from a terminal window while testing, so the
log file ends up in my current directory.

btw, the latest code puts the ant log into the athlete directory to make
it easier to find (and also captures a bit more detail). Next dev build
should make this a bit simpler..

Regards,
Jon

On 29/01/16 10:18, James Manners wrote:
> I have seen these on a variety of bike setups:
>

> * Garmin GSC10, PowerTap Pro Hub, Garmin Heart Rate Monitor &
> Suunto Movestick Mini
> * The same without the GSC10.

[James Manners]

Hi Jon,

Sorry for the delay, but I here is an activity completed in train view with the data spikes https://www.dropbox.com/sh/m6brn2xzjni3185/AAD6AkSweodofeblJ07hDj49a?dl=0

Both the antlog.raw and activity file are there. 

Let me know if I can do anything else to help.

Cheers,

James

[Jon Escombe]

Thanks James, will take a look..

Cheers,
Jon

[Jon Escombe]

Thanks James, I see what it is now!

Will get a bugfix in this week..

Cheers,
Jon

On 08/02/16 19:42, James Manners wrote:

[James Manners]

Hi Jon,

Again great news. I was wondering if there any work arounds that could help until the fix and next dev build is ready?

Cheers,

James

[Jon Escombe]

Hi,

Turned out to be a trivial fix in the end (battery status messages were
being handled incorrectly & corrupting the speed/power calcs), but
unfortunately there is no workaround that I can think of - short of
swapping your hub out for an older model! :(

Btw, I don't actually have a G3 to verify the change against, so would
be useful to get verification of the fix from someone who does. Have
submitted a pull request this morning..

Regards,
Jon

[Mark Liversedge]

Lets get a development build out this weekend.

Mark

[Liam Frawley]

This is excellent news. This was the single issue the was stopping me from using GT in Train mode.

Is there a development build with this fix? I can test immediately once a dev build is ready. I could repro this issue easily with my Power Tap

[James Manners]

Morning Jon,

YIPPEE!!! I tried out the new development build this morning and it worked perfectly! 

Tahnk you to Mark, yourself and all the GC team!

James

[Jon Escombe]

That's good news, thanks for the confirmation!

Regards,
Jon

[Patrick]

I'll second James. Just tried the latest development build this evening in train mode and there were no power or speed spikes during my workout.

Thanks Jon

[jeroen mermans]

I still have powerpeaks, they are  65535Watts => the max value of int16...
i get this when callibrating my tacx satori smart, when i have to stop pedaling

is there something i can do to fix this?

[Liam Frawley]

I've used the latest development build of GC over the weekend with my PowerTap ... thank you, thank you, thank you, thank you, thank you, thank you!!!!

It's perfect. No power spikes at all. Well done for fixing it!

[Jon Escombe]

Hi Jeroen,

Can you describe the steps you take to calibrate (as native support is
not yet in GC master)?

If you can send me the .json and antlog.raw files I'll take a look at
the data.

Thanks,
Jon

> --
> _______________________________________________
> 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.