Tips‎ > ‎Indoor Trainers‎ > ‎

Computrainer

Before the advent of the Wahoo Kickr, I was a big advocate of the Computrainer, and in particular, Erg-mode based training.

This page is to share a few lesser known tips, thoughts and ideas that Computrainer users may not know about. It's not a lesson in how to use a computrainer.

Golden Cheetah - Alternative software instead of Computrainer's 3D or CS16:

http://goldencheetah.org/ (release versions)
http://goldencheetah.stand2surf.net (development versions, currently unstable for me)
One great feature of GC3 is being able to upload a map/route created in, say mapmyride, and then GC uses Google-Street-View to create the images you would see on that ride. Finally you can ride through Times Square at 28mph.

For me, GC3 doesn't work when connected directly to the serial port, but it does work when I have CT-ANT running in between the two.
I'm still evaluating this compared with CS16, I'll write more in the summer.

Ditch your old Racermate HRM:
I use an ANT+ HRM on my CT. Data is passed to a $20 garmin dongle on the PC, then ct-ant passes it to the Computrainer Software (3D or CompCS).
You can also use this method to view your CT and Quarq/SRM/PowerTap power at the same time.
You can also use your ANT+ strap+dongle with Golden Cheetah. That doesn't require CT-ANT+.

Keep your calibration number high (and other calibration thoughts):
The default calibration number is 2.00, but that doesn't mean 2.00 is ideal. By increasing the resistance on the tire you can raise this number, which will give better CT accuracy. This is documented in the new CT manuals, but not the older ones. A calibration number greater than 3 is suggested. Of course, you should calibrate after warm up on most rides. I often re-calibrate at the very end of a ride, and then adjust my reported numbers accordingly, based on a calibration number of 0.02 equaling 1W. thus, if I calibrate at 2.40, then ride an hour at 320W, then recalibrate and get 2.50, I'll use the later number, and add 5W to my reported average (0.1/0.02).

Calibrate in the same body position in which you ride:
I used to sit back on the bike while the wheel slowed down during calibration. However this increase weight at the back of the bike increases the calibration number, and would thus add about 3W to the reported performance.

Keep your bike speed low:
The following graph shows an interval workout done in Erg mode, where the "requested power" (set at the head unit) was the same (9 x 4 mins @220W w/ 2 mins recovery) on each interval. The cadence was varied on each interval, which ought not to affect the power in erg mode. This ought to be an overly simplistic way to determine HR response at different cadences.
The upper purple line shows that the actual power recorded by the CT differed on each interval, depending upon the cadence/speed.

2011 11 08 - Intervals at same power and differing cadence.jpg

As the following thread explains, the computrainer doesn't accurately account for the change in aerodynamic drag of the wheel at different speeds.
The solution, and the tip here, is KEEP YOUR SPEED LOW. Keeping the speed between 12 and 20 mph will reduce the above effect, and thus, in erg mode, allow speed power to remain roughly constant regardless of cadence.
This is done simply by selecting a lower gear, so the wheels go slower and more of your power is absorbed in the CT resistance unit rather than in moving the air in your basement.

Exporting your data
If you use the Coaching Software (CS16) or Golden Cheetah then within the options you can choose to have CS16 export your data to a file.
By default CS16 will export to a .cdf file. However many tools (eg Sporttracks don't recognise this format). There is a CS16 option  File -> Export -> Export Options -> "Auto Export". Ensure that's enabled, you may also uncheck some of the useless parameters. The .txt file that results can be imported into many analysis tools. However it's very big, so you'll want to zip it if emailing it. There are other options worth messing with in CompCS, such as ways to reduce the amount of displayed data, and scale the graphs appropriately.

Spinscan
Don't worry if your numbers suck. Don't gloat if they rock. Spinscan doesn't really matter. Here are some thoughts of one of the authorities such subjects, Andrew Coggan: