You Can Test Drive GoldenCheetah/Aerolab on Windows Right Now!

aerolab_first_patch For those of you who were curious about Aerolab, there’s now a patched Win32 build for it in a development version of Golden Cheetah 1.3.   Thanks to Mark Liversedge for building my patch into GC1.3.   Keep in mind that this is a very early version of the Aerolab module and that it is not currently part of Golden Cheetah.  Aerolab is just a prototype right now, but it will eventually be a complete suite of aerodynamic tools for virtual elevation work.

For those of you not familiar with it, Aerolab is a virtual windtunnel that helps you analyze your ride files.  Under certain conditions, Aerolab will help you estimate the rolling resistance (Crr) and aero drag coefficient (CdA) of you and your bike.  It does this by helping you match a computed virtual elevation curve with the real curve.

[Edit:  Please keep in mind that this is a pre-release patched special build of GC1.4, brought to us by Gareth Coco.  It's meant to give you a taste of what's coming. ]

Download it here. Give it a try and let me know what you think!

17 comments to You Can Test Drive GoldenCheetah/Aerolab on Windows Right Now!

  • Robert

    Very cool. I like the visualization quite a bit, and the sliders are very very smooth.

    Here are some random comments.

    1. The calculations appear to be off.
    2. You should allow for rho < 1.0
    3. What are the units for elevation?
    4. Distance on the x-axis is way off.

    IOW, niggling details are off but functionality is very slick.

  • Robert

    Also, I’d restrict the range of the sliders a bit and increase the resolution.

    F’rinstance: rho should be range between maybe 0.9x and 1.25ish. Crr between maybe .001 and .01. CdA between maybe .1 and .5. You’ll need a drivetrain efficiency constant.

  • andy

    Ok, I’ll go home tonight and make sure the sliders are right and recheck my distance calculation. This stuff is begging for 4th-order Runge-Kutta, which I’ll have to put into future version.

    Gotta leave *some* work for later versions. :-)

    Thanks for the feedback, Robert.

  • Ben

    Hi, I doubt it’s related to your part of the software, but I can’t get a file to import (.csv or .wko (2.2or 3.0)) Sounds like a great idea, but… Any thoughts. I may be pushing things by trying to run GC under parallels on a Mac, but…
    Thanks Ben

  • andy

    Ben, it may well be because of Aerolab. I’m looking into it right now. Stay tuned and check back soon for updates to the binary.

    Thanks for your comment!

  • I am trying to upload training files from a WKO+ data file. This is not working and it is suggesting I open the file and save as… What should I do here? The set up looks really good and provides charts that are not offered in other programs. Great work!

    • andy

      I am currently working on a new, clean patch for Windows, Mac, and Linux. Stay tuned for the announcement and we’ll get you a proper version.

      Thanks for trying it out — we appreciate you using Golden Cheetah!

      Andy F

  • James

    GC had no problems importing one of my field testing WKOs from last spring. This ride consisted of a set of Lim opposing-direction runs, followed by a Chung session on a running track. Aerolab shows me a flat plot when I plug in the Crr calculated from my Lim runs, along with the CdA I got from the Lim runs and my original Chung analysis spreadsheet. So, looks like your math is at least close :-)

    1. X-axis is labeled Distance but it’s pretty clearly time.
    2. Inputting values in the Crr/Cda/Rho/Total weight text boxes causes the values to go haywire.
    3. Would be helpful to be able to zoom in on an interval

    • andy

      A new clean patch is imminent, and it will fix the reported problems. Thanks so much for trying Aerolab, and stay tuned for plenty more features in the coming versions.

  • Chris

    It’s looking good

    The sliders are good for rough estimates but it would be good to be able to type in specific values so you can vary CdA by o.oo1 and Crr by 0.00001

    It would also be useful to be able to plot more than 1 run so different conditions could be compared.


  • Ian

    Hi, is it possible to make the Eoffset range greater than 300m? I ride with my polar reading actual(ish) elevation, and the starting point will always be higher than 300m above sea level.

    On a related thought, does air density come into play in the calculation too? (I would assume it must) If so, a range able to include Colorado passes would allow this variable to be set too.

  • andy


    First, thank you for trying Aerolab and Golden Cheetah. I really enjoy it when people use it!

    I can get a patch in to Gareth Coco and will advise you when it is ready. How much altitude do you need?

    Yes, air density (rho) is in play. You will need to compute this number separately, though, for the moment. Just get this spreadsheet ( and go to the “Air Density” tab and fill things in. It’s easy!



  • Ted

    Just wondering what the status of a Linux version of GC with aerolab is?



  • Tom

    Tried this out today – I think it generally worked pretty well. One question, though. If I’m doing multiple laps around a circuit and the barometric altitude drifts up with each lap (about 1′ per 1.25 mile loop), should I match the curves as closely as possible or would it be more accurate to try to get the VE curve to be flat? I’m thinking the latter.

    If so, it might be useful to have correction variable for the “actual” elevation curve to flatten it out (pre autosolve when it comes) given that so many of us rely on barometric measurements for this.

    All in all, this is a great tool and I can’t wait to start doing comparison runs.

  • andy


    Thanks for using Aerolab.

    You definitely should try to flatten out the VE curves. A secondary problem is that as barometric pressure drifts, air density does, too. Aerolab currently only supports a constant-in-time air density model.

    When I get the aero sensor finished, I will be turning my attentions to making Aerolab use it, and to making Aerolab more usable. My “todo” list includes:
    1) interval selection
    2) loop and out-and-back superimposed graphs
    3) yaw variation in CdA

    If you have any other ideas, please send them.



  • Tom

    Andy –
    I did a very flat 40km ITT today. The barometric conditions were very steady, which can be summarized as follows: 10km head wind; 10km side/head wind; 10km side/tail wind; 10km tail wind.

    In Aerolab, the actual elevation profile looks as it should – very flat, but the VE profile looks like a mountain (if I set the beginning and ending points to the same elevation. I think this is correct – the headwind appears as a climb, but thought I’d check.

    Also, will a CdA derived from this ride have any relation to reality?

    • andy

      Hi Tom,

      Yes, you’re right — headwinds look like a hill for VE. And if you turn around and go back with the same wind as a tailwind, you don’t quite get back to the same elevation as you start, because it’s a quadratic term, not a linear term.

      No, the CdA will be over-estimated. If you know how much the headwind added to the virtual elevation profile, you might account for it. If you had a constant headwind, for example, your VE should look like it was rising, rather than level. So a levelled VE curve would over-estimate your CdA. Although I’ve only looked at it a bit, probably Adam Haile’s virtual work method might be helpful if you can guesstimate how much headwind you had.

Leave a Reply to andy




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>