July 31, 2012

CycleOps Powercal versus Quarq Review - Is Powercal Any Good?

Short answer - YES, it's surprisingly decent. Long answer - see below

Well if you're like me you have been wondering how you can measure power in a cheap way, but you don't want a piece of crap. And you want something that can take a hike, a run, (probably not a swim), hockey, etc., etc., and convert it to a cycling equivalent that I can put in my WKO+ to find out how tired I am and how fit I am getting. Theoretically, this thing does all that - if it works. Multisport power measurement nirvana.

Reading the reviews on line I have found them to be lacking significant details - like plotted comparisons. And all the armchair experts whining about how heart rate and power aren't fully coupled and therefore it can't possibly work made me think that there had to be more to this thing than a simple heart rate to power correlation. Any high schooler could do that and everyone who has used a powermeter would know it wouldn't work very well. Obviously, change in heart rate had to be part of the calculation, but nobody seemed to mention it. Where are the data?

Anyways, I felt the potential benefits outweighed the potential for a piece of crap and I bought one for a very reasonable price. Really, price is the other huge advantage for this thing compared to other power meters (can you say "an order of magnitude"?) - if it works of course. Seemed to be vaporware for a long, long time, but finally I was able to buy one - arrived yesterday.

So I took my Quarq - recently calibrated and rebuilt to boot - hitched to the Garmin 500 and my brand new Powercal hitched to my Garmin 705 on a 3+ hour training ride today. Talk about data - two displays to watch while listening to tunes on my Blackberry. I was wired to the gills. Radiation nirvana.

The value of a long ride is that I could put the thing through a variety of activities - short sharp efforts between stop signs and lights, small hill attacks, longer hill attacks, some steady riding and some slacking off of course. About the only thing I didn't bother with was high intensity sprints because I doubted it would work for those anyway.

Now for the stats:

Surprise! The Powercal isn't simply using heart rate as predicted by power. Actually not a surprise as that would be pretty useless. When riding, it is obvious that the meter is using the rate of change of heart rate as well. When your heart rate is increasing, power is high for a given heart rate, when your heart rate is dropping, power drops, even to 0 at high heart rates. Which is entirely realistic if you stop pedaling. Looking at the data, the algorithm is complicated - e.g., for a rising heart rate followed by a steady heart rate - power rises and then drops - for both the Quarq and the Powercal. That's comforting.

Looking at the scatter, it is similar to the Quarq, but notice the banding of the Powercal relationship. I could speculate that it is using discrete categories of rate of change of heart rate instead of a continuous relationship. But I prefer not to speculate. The important part is that this scatter looks realistic. Since points are obviously plotting over top of each other due to the discrete nature of the relationship, we can't tell if it matches the Quarq really well or not. Who cares. The next graph is what you should care about.

Well I'll be a monkey's uncle. That's not too bad is it? This is without calibrating anything. Apparently, calibration didn't tend to make anything better for most of us, so CycleOps removed all the videos on how to calibrate that used to be out there. That's a timesaver. And you don't need another powermeter to use this thing. This graph shows that the Powercal does a decent job of tracking all kinds of efforts.

I wanted to zoom in on the first part of the ride where my power was more erratic due to lights, pedestrians, stop signs, etc. Here's what I saw:

Again, amazing. There seems to be a noticeable delay in peak efforts which is understandable as there is a time lag for heart rate to respond. But only a few seconds and the peaks are mostly all there.  Anything over about 20 seconds is nicely tracked so you can use this for most types of intervals. Nice and not unlike comparisons of powermeters that measure power directly I have seen.

Unfortunately, for me it does tend to underestimate power at higher power levels and slightly overestimate power at very low levels. I expect results will vary depending on your unique heart rate versus power relationship. I think the relationship is a bit better than the trend line shows - the very low values are pushing the line down a bit at higher wattages. The scatter is expected due to the time lag relative to the Quarq inherent in heart rate.

How important is this? Well for comparing to other cyclists, this suggests I'm not working as hard as I really am. Might matter for bragging rights for example - (My power was X for 5 minutes, what was yours? Yeah but who got up the hill faster? Just export your files as GPX for upload to Strava and no one will ever know.). For planning and evaluating workouts, this lower power doesn't matter at all. Relative to itself, it is consistent and therefore doing a pretty good job. There are occasional random spikes, but not a big deal.

There you have it. I can self calibrate - on this ride the Quarq TSS was about one third higher than the Powercal (actually 38%, average power was 13.5% higher, normalized power was 17.6% higher, VI=1.25).  Maybe do a couple of more rides to check. I'm not too worried that the power values themselves don't match but I want a good measure of training stress. If you didn't have another powermeter, you wouldn't have to care about calibrating one to the other. You don't need to match anything.

Other comments - I really like the heart rate strap on this thing - very comfy but it may be a bit short for big guys. Works fine for me though and I'm over 180 cm (5 11 - used to be 6 feet, but I shrunk). I did have to loosen it almost all the way to fit my gigantic chest.  Maybe it stretches, I don't know. No trouble syncing to my Garmin. Heart rate average was 1 bpm lower than the Garmin strap. Go figure. Makes it that much harder to hit threshold I guess.

Multisport nirvana? - check. Low cost? - check. Good buy? - checkmate. Unless you are a bit anal, this thing is all you really need for training with power. Buy one, determine your threshold power using it as described in many training books and blogs, and then train, dammit. You just saved yourself a couple of grand or so and guess what? You will get faster just as quickly with this as you would with the expensive powermeters.

Additional Calibration Rides
Did various rides and portioning of rides to get Variability Indices of 1.01 to 1.43 for 30 to 200 minute chunks. And lots of variability of effort from cruising to time trial mode, to intervals. No apparent relationship to variability index. Stopping does tend to give you a power reading on the Powercal when you move around so that of course is wrong - it reduces the difference between the Quarq and the Powercal. Normally not significant (e.g., traffic lights) unless you do it for a good chunk of your ride - pause that Ant+ device when you stop for lunch. Other than that I am getting an average of 31% higher TSS, 11% higher Ave Power, 14% higher Normalized Power for the Quarq on this hodgepodge of data.

August 24, 2012 Update - Well I have had some quality time with the Powercal and I still like the way it works... most of the time. I have encountered one problem that also occurs with my Garmin heart rate straps, but not as badly. They only work well when you are sweating at least a bit. When you are a bit chilly, like today unfortunately, it takes some time for the monitor to give you a reasonable reading. On the bright side, I have never hit over 3000 watts until recently. But I'm a little concerned that might be incorrect. I am finding the Powercal heart rate strap is a little more temperamental in this regard than the Garmin strap. Once you are fully warmed up and sweating a bit, it works well. That can take 15 or more minutes though depending how cold and windy it is. And yes, I put spit on it.

October 14, 2012 - more update and some comment cleanup - I am finding the meter works well when I'm sweaty and works really poorly when I'm a little chilly as happens all too often for the foreseeable future - until spring. I don't have to be hypothermic chilly, just that "warm enough to keep riding and not put on my jacket" chilly. Basically, the heart rate readings get sporadic and the power readings get really wacky. I did try an assortment of cremes and some helped a bit. But none solved the problem. The heart rate monitor definitely does not pick up your heart rate as well as the Garmin strap and it seems to be getting more sporadic over time for some reason. Here is a very good thorough discussion of the issue and solutions. This can be quite irritating if you are a detail person. Plus your TSS can be somewhat out of whack compared to the TSS when you manually remove the spikes.

The strap doesn't receive data from the Garmin, it only sends. I made my weight 270 kg and my power readings were unchanged. I'm pretty sure I was the fastest climbing 270 kg person in the country but I didn't check. To do list for Strava? Course then I'd have to pay.....