BSX Insight: Lactate Testing and Playing With Power Meters

Just a warning: this is a geeky post…

I received my new BSX Insight which is a wearable lactate threshold sensor. I supported its development through Kickstarter and the final product is a very tidy piece of kit. Your lactate threshold is a key metric in determining one’s training zones. Testing is done by taking regular blood tests as one ramps up the work. What makes the BSX Insight so unique, is that it can monitor lactate without taking blood: the BSXinsight monitors muscle oxygenation via a light array shined into the calf.   It then runs an algorithm to give users lactate threshold (LT) power and heart rate numbers, plus training zones based on percentages of those figures. So the process is to slip the sensor into a leg sleeve, hop on a treadmill or a bike trainer, and then do a test. Sounds easy? If only …

Bike Testing

The first thing you need is an ANT+ or Bluetooth power meter. Fortunately I have both, so I mounted my Powertap wheel on my BMC Transcontinental Race bike which has a Stages power meter. I then put the bike onto my Computrainer which also measures power. Thought this was a good opportunity to compare the three meters. I did a calibration on each and then started the test.

The test is actually quite simple. I downloaded the App to my Sony Xperia Z1 cell phone and it then walks you through a few questions on your training history, volume, etc. it then takes you through a ‘ramp’ test where you aim to ride at a certain power level for 3 minutes, and the power then increases by 20 watts for the next three minutes. I used my Stages as the primary power meter as that is the one that I will be using for this year’s Transcontinental race.

The first challenge I had was getting the App to work. It would start and then crash my phone. I didn’t know why as this appeared to happen randomly. The App would also sometimes connect to my BSX sensor, but other times would not. I tried it on my older Xperia Active phone, but it must have a different version of Bluetooth as the App was not compatible.

Eventually—or should I say randomly—I got everything set up and did the test as per the recommendations. I went all the way through and got to the end, pressed stop and … nothing was saved. Talk about frustrating as these tests are seriously hard.

Put it aside for a month as I was travelling on business, and then tried it again. There was an upgrade to the App and it didn’t crash. The sensors all connected which was great. Then did the test. When I got to about 260 watts it stopped recording the heart rate and power data. I pushed on and finished the test, but once the test was over it would not upload any data. I had hoped that it would at least upload the BSX data without the heart rate or power data, but not. So I’ve dropped them a note suggesting that they need to put a bit more work into the system.

It’s a really nice concept, well engineered … just wish it would work!

Comparing Power Meters

While I was not successful in getting the lactate test done, the data were still quite interesting to play with.  I imported the data to WKO+ and then recorded the 3 minute average power and heart rate. The data are below.


Stages vs Powertap: The Stages power meter is on the left crank. I have a weak right leg—the result of four knee operations—and the Computrainer Spinscan suggests that my power balance is typically 55%/45% between the left and right legs. So it is not surprising that the Stages readings are all higher than the Powertap readings, as the latter is at the rear wheel (and so also does not include the 1-2% losses in the drive train with regard to energy efficiency.

The data from 180 – 300 watts were more consistent so I decided to play with them to see what the implications are of leg imbalance. Using Excel’s ‘Goal Seek’, if the left leg did 4.2% more work than the right leg (i.e. 54.2%/45.8%) the average difference between the Stages and Powertap readings are 0%.  Given that these meters are advertised as + 2% (Stages) and + 1.5% (Powertap) there was very good agreement once I corrected for leg imbalance.

What about at low power levels? For one thing it was very difficult to maintain the target power level, as evidenced by the noise below. Also, I expect that my left leg does much more work at these levels than at higher power levels.





This bias at lower speeds is reflected in the relationship between Stages and Powertap where the intercept is 14.4 watts.  If you set the intercept to zero the relationship is: Powertap = 0.9513 Stages, or Stages = 1.0512 Powertap: again suggesting about a 5% leg imbalance in favour of my left let. So to keep things simple in the future, I’ll just assume that the Stages is 5% higher than the Powertap.


Computrainer: The Computrainer read higher than both the Powertap and the Stages, again with strong linear relationship. What is interesting is that the intercept for the Powertap was 0.8 Watts, vs 16.3 Watts for the Stages. There is definitely a large bias entering in with the Stages given that we have such a large intercept when compared to the Computrainer and the Powertap.



If you constrain the intercept to zero, the regressions are: Stages = 0.9709 Computrainer and Powertap = 0.9237 Computrainer. I would have expected the Computrainer to be much closer to the Powertap. It was interesting that from 80 to 180 watts there was close agreement between the Stages and Computrainer, but above that they diverged. Wheel slippage? Must look at this more again.

A final analysis of no value was to look at heart rate as function of power. Here’s the graph below.


So the main conclusions from looking at the power results were:

  • Strong linear relationship between the different power meters
  • My leg imbalance is introducing a bias which is about 5%
  • The Stages reads 5.12% higher than the Powertap; 2.91% lower than the Computrainer
  • The Computrainer reads 3% higher than the Stages; 8.26% higher than the Powertap

Training Zones

Once you have your MMP, you can use this to estimate the power that you should be at for different training zones. The table below from shows the recommended ranges.


Here is an example of the resulting training zones by Powermeter.


It is more common to use the ‘Functional Threshold Power’ (FTP) for the training zones which is based on a 2 x 20’ or 1 x 30’ test. I did one a few months ago and the resulting training zones using Coggan’s method and the calculator at: are below.


What is interesting is that Coggan suggests that the MMP is 2-20% higher than the FTP. If you take 20% off of 309 watts you 258 watts, which is quite close to the 254 watts from my November test.

One thing you can estimate from the maximum power is your VO2 max. This calculator uses the Hawley-Nokes equation which, using Powertap, suggests my VO2 max is 56.6 mL/kg/min. A lot lower than the 70.6 mL/kg/min I measured on a treadmill in 2005!

Now, if I could only get my BSX Insight to work I would have more data to play with!


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s