Train Mode clock reliability

75 views
Skip to first unread message

Kevin Hays

unread,
Mar 27, 2015, 12:53:02 PM3/27/15
to golden-che...@googlegroups.com
Hi Folks,

I have been using GC in Train mode with my Computrainer.  When running a Garmin 810 at the same time, and starting them both together so that they display the same time (within fractions anyway), I have noticed that the GC clock and the Garmin clock become offset over a long period of time.  The difference is about 1 second every 15 minutes, so that at the end of an hour, the GC clock is about 4 seconds ahead of the Garmin.

This is no big deal, I'm just curious if anyone has an idea why.  I'm not sure which one of them is the accurate time.  FWIW, I'm on an old dedicated XP laptop (maybe computer or OS-related?)

Background: I have had much trouble with the HR pickup of my Computrainer over the years.  I've tried everything Racermate suggests, including replacing HR belts and pickup sensors.  So I have taken to recording HR on my Garmen, exporting both files to CSV, and pasting the HR data into the Computrainer CSV, then back into GC.  This is a PITA.  Time to get a life!

--Kevin

Mark Liversedge

unread,
Mar 27, 2015, 12:56:39 PM3/27/15
to golden-che...@googlegroups.com
Its almost certainly GC that is to blame :(
We use a timer signal to trigger the update at 5hz and it will drift if we're hitting the CPU hard (e.g. video).

We could possibly fix it, but its not that easy to fix (unless anyone has a good idea for this ?)
Mark

Mark Liversedge

unread,
Apr 25, 2015, 6:24:50 AM4/25/15
to golden-che...@googlegroups.com
I'm working on fixing the issues related to precision and have found that GC is actually doing the right thing for the GUI -- the time being displayed *IS* the absolute time elapsed according to the PC timer, to the nearest 100th of a second every 200ms.

The data that is written also uses the same timer but only writes 1 sample each second so that will not drift either. 

I am fixing a bug related to accuracy. The data written only reflects the measurement for the last 200ms rather than for the full second. I'm fixing that, but for the most part it isn't a big deal for most users and probably will go largely unnoticed.

I suspect the timer on the Garmin and PC run at slightly different speeds. I cannot comment on which one is correct, but would be surprised if your PC is wrong ................

Mark
Reply all
Reply to author
Forward
0 new messages