Saturday, October 27, 2007

Body -> Comparing Power Meters: Polar CS600 to PowerTap

This is one of those reviews that I really didn't want to write.

Mainly, because I knew it would be a bunch of work. However, seeing that I needed to ease back into training after the forest fires, I decided to take a short ride to do a bake-off between my two power meters: The Polar CS600 and the Cycleops Powertap.

Conventional wisdom would tell you that the Powertap would be more accurate than the Polar unit. I pulled my Powertap off of my racing biking and crammed it onto my commuting bike. This highlights the first issue: The Polar was a pain to set up. While I didn't like moving the Powertap, I really didn't want to move the Polar.

Moving the Powertap was a bit easier, but the biggest challenge is that my commuting bike is based around a mountain frame, and the spacing is 5mm too wide on the rear wheel. However, you can clamp hard and a road wheel will stay in a mountain bike dropouts. However, I rode a bit gingerly, not wanting to dislodge the wheel.

Although I did not have an ergometer to test them against, I could compare them against each other and against my perceived exertion. If they both measured the same, then I had confirmation. If they varied, then I would use my body to judge which was closer.

Once I got home, I dumped the data and parsed it into an Excel data file, and plotted the two meters against each other.

The first half of my trip was up a hill, and while I didn't grind it out hard, I did have a moderate pace that was above 150 watts for the most part. I immediately noticed on this section that the meters seemed very close. This was better than I expected.

Here is a chart of this section. Click on the chart to make it bigger.

The polar samples at a lower frequency. You can see that the power curves are simply smoother. However, for the most part, the Polar curve and the Powertap curve track nicely together. The TSS score from this data was close together, as were the calories burned.

Net-net of this section showed that if you are pedaling at a moderate pace to harder pace, the two power meters track nicely together. Now the Powertap samples twice as often, but in my mind you don't vary your instantaneous power that fast. So the higher sample rate is inconsequential. A bit like a Nyquist sampling test.

However, once I crested the hill, I had a long down section where I had no desire to pedal hard, due to the jammed in nature of the rear wheel. So, I was just pedaling at an easy pace. This is where the two power meters split. The Powertap had under 100 watts of power. The Polar, on the other hand, was showing 30% higher readings.

As you can see on the graph, this is reflected in the data. I don't need any rational on what was happening. The Polar unit measures vibration from the chain. As you put power on the chain, oscillations get faster. If you could hear the chain, it would sound higher. However, at easy pedaling, it is difficult to catch the note. So, I clearly believe the Powertap. It felt right.

So what are the take aways?

1. The Polar is surprisingly accurate. I thought it would be okay, but under moderate to high cycling loads, it is very close to the Powertap.

2. Therefore, as long as you are really working out, and keep the chain under tension, the two units will be very close. The only issue is if you are easy pedaling. However, unless you are pedaling easy all day, I don't see this as an issue. While I had a tracking issue, it only happened under a particular condition, which I don't do very often. If you are tourist that pedals at 80 watts, it will over read the power. However, if you are an easy tourist, I bet you don't care about power meters anyway. For me, this was a short bit of high biased data that doesn't ruin the whole data set. It is a "who cares?"

Now, in some sense. I have also seen reports that some people have removed a link or two to keep the tension a little higher, and this has removed bad data. I haven't tried it, so your mileage may vary.

However, I am pretty happy with the CS600. I bought it mainly for other things like the altimeter and the great pulse monitor. However, it does a very good job for power also.

"Body" -> Historical Pulse Levels

Review of the Oregon Scientific Data Logger

Generally, pulse rate monitors are ubiquitous. You can find them very inexpensively, and virtually at any large store. To get an accurate measurement, you normally need a chest strap, which picks up the electrical impulse from your heart. Your pulse rate is then displayed onto a special wristwatch that you wear.

As you get to more sophisticated levels of pulse monitors (higher end Garmin, Polar, and Powertap units), you will also get some ability to store the workout normally onto the wrist watch (or bicycle computer). Monitoring pulse rate during endurance activities is still the simplest measure of activity, although some people are replacing simply pulse monitoring with a more sophisticated power monitoring. However, for the day in and day out athlete, pulse rate is great in helping to get another data point on how hard you are training.

The biggest problem with the normal data logging systems are two fold:

1. They tend to be proprietary
2. They are expensive
3. And due to #2, if you lose or break the pulse monitor, you are stressed (thus leading to weight gain)

Oregon Scientific, gadget maker, has broken through some of this with their data logger. For an inexpensive $30, through Walmart, you can have the ability to record your historic pulse rate for up to 30 hours.

Over the 25 years that I've used pulse monitors, I have a bunch of heart rate straps all over the house. Generally, I've destroyed or lost the watches or cycle computers, before I destroyed the pulse strap. The nice thing about the data logger is that you can use most of you pre-existing straps as long as they are in the 5.3Khz analog range. This means that all your old equipment will probably work.

(Now, the new more sophisticated pulse monitors from Garmin and PowerTap may feature the ANT+ protocol, which is 2.4Ghz, and won't work with this logger, but at this price point, you'll have all the data logging you need with the systems that feature this more sophisticated technology.)

The system is very easy to use:

1. Put on your pulse monitor strap
2. Place the data logger less than 4" from the transmitter
3. Wait for acquisition (solid green LED on the logger)
4. Now you can move is up to 30" away.
5. After the workout, press the button for 2 seconds to turn off
6. Return to PC to upload your work out, which is time stamped

The logger comes with a small app program and a USB cable to download the data. Below is a screen shot of the app.

Photo Sharing and Video Hosting at Photobucket

The logger program is very unsophisticated, but does the job. I do, however, like the ability to export the data into a CSV file, which you can grab with any spreadsheet program.

For example, below I've posted just sitting at my computer typing. I've then exported the data to Excel, and I've graphed out my resting pulse.

I've set the logger to record every 1 second. As you can see, I've had a couple of movements, which sent my pulse up a bit. However, if you put a trend line on the data (see the equation), the last bit of "47" is where my average pulse is resting just typing. In other words, I have a typing pulse rate of 47 beats per minute. However, you could also do more sophisticated analysis on your heart rate data: RMS, Area Under the Curve, or ramp ability of the heart.

Photo Sharing and Video Hosting at Photobucket

Some additional features:

There is a finite amount of memory in the system. On the low end, you can record 15 hours at 1 second intervals, but you can set it to record higher intervals. As you would expect, 2 second sampling would give you 30 hours of data logging.

Geek Of The Week Ideas:

1. Obviously, if you can stand the chest strap, it would be interesting to log 24 hours worth of data. How low does your pulse go at night?

2. If you are into weights, interesting to see your pulse during lifts.

3. You can see how your Tabatas went. After all, if Tabatas don't raise your pulse, nothing will.


All in all, for $30, this is a pretty cool little gadget for endurance athletes. If you don't have a pre-existing chest strap, you may want to get an entire ensemble from Walmart with the strap and the watch for an additional $35. However, I cannot vouch for the rest of the system. For me, it would appear that the chest strap is a bit more tricky to get correct. Regardless, Walmart is pretty good about taking things back, so try and return if it doesn't work.

While it is only available on-line, you can always ask them to ship it to the nearest Walmart for free, thus skipping the shipping fees.

Saturday, October 13, 2007

The Power Of Weight

"How many watts can you put out?"

Hardly seems like a question that you'd hear as a young kid. Certainly, not as common as "how many pull-ups can you do?" However, it is in the same vein. It is a measure of fitness. So, how many Watts can I put out?

As already discussed, you can measure power in Watts, and this can be measured by special instrumentation. However, raw wattage is not all that important in bicycling. Why not?

The answer is the same for athletes as it is for cars. A big engine in a light chassis means fast speed. A big engine in a big chassis means mediocre speed. So what you want is as big as engine as possible, with the least weight. (This is why I am currently dieting, but that is a story for my other blog.)

So, how would we use this information? The answer is as close as a race I did today. This morning, I did an 11 mile time trial. It was not flat, it was a lot of up hill. It took me 33 minutes to do the time trial, and I averaged 269 watts of power.

As mentioned before, Andy Coggan has done a lot of work on the theory of training with a bicycle power meter. In his work, he has defined a term called "functional power." Functional power is basically the power that you can hold for 1 hour. While, I did not go a full hour today, I can take my 33 minute time, drop it by about 7%, and I can get my functional power. This is about 250 watts.

However, I am a big guy carrying a fair amount of weight. Part of this is muscle, but a lot of it is fat. As a matter of fact, I think that I carry around roughly 40 lbs of fat! Every time that I pull that fat up a hill, it does nothing for me. The same athlete, with the same power plant, will be faster if they are lighter.

So, what is really important is the watts that you can put out per kilogram of weight.

So here are my vital stats:

Functional Power = 250 watts
Weight = 90 kg
Watts per kilogram = 2.78

Now, click on the table below. Look at the FT column and 2.77.

You can see that 2.77 is about right in the middle of the Category 5 Racers. If you care serious about bike racing, you'll do some training and the first category you will race in is Cat 5. Now, lets say that by some miracle, I could lose 22 lbs of fat (and I would still have 18 lbs left!), I could move one entire category up.

Another way to look at Andy's table is the above graph. What I've done is take the numbers in the first chart and graph them. Every category has a different color. The different category of bicycle racers are roughly the same length. The graph here shows that you see a steady increase in power as the success of the rider goes up.

Now lets say that I could lose a bunch of weight, but keep the same power out. The following chart shows what my weight would need to be at 250W worth of Functional Power.

If you've ever looked at the upper echelons of bicyclists, they are all legs, and super lean. They will be 6 foot and 140 to 150 lbs of weight. In the following chart, I have graphed my needed weight @ 250W for me to get to higher categories of performance.

If I was built like some professional bicyclists, I would only weight 150 lbs. I have shown this on the chart. However, even at 150 lbs, I am not putting out enough power to get beyond a Cat 3. So, while I'm in okay shape, I would need to do more than lose weight to get above a Cat 3 racer.

Therefore, it should be clear from the data, that just decreasing your weight is not enough. You also need to be powerful. However, as you probably know from experience, a powerful rider with a lot of body fat will also do poorly.

Now, in bicycle time trialling, there are two types of races: flat and hilly. As the bicycle course becomes more and more hilly, the power per kilogram becomes more and more important. When you are on the flat, carrying a few pounds isn't noticeable. However, if you are climbing a lot, the extra weight becomes very important.

However, at the upper ends of performance, you need everything going for you. You need to be lean and powerful. But even us mortal and older bicyclists need to understand that for optimum performance, we can also improve both by power and weight.

Sunday, October 07, 2007

"Mind" -> Fat Brain and Bromocriptine

One of the biggest self deceptions is wide spread belief that some people eat a ton of calories, don't work out, and are skinny.

The fact is that nobody naturally has a high metabolism that allows you to consume all the calories you want. The best way of understanding this to to look at what it takes to burn a lot of calories, and an an example of this is as close as my other blog.

I have a 62 mile loop that I do on my bicycle. On this loop, there is a lot of climbing, and it is very difficult to do this loop in an easy fashion. On my bicycle, I have special instrumentation (see posts on power meter) that measures direct power and this allows me to get an almost direct measurement of my calorie burn.

This route burns between 2100 to 2300 kilocalories (commonly called just calorie). Yesterday I did this loop in 4.5 hours. During this time my pulse was elevated (120-130 beats per minute). I was constantly sweating profusely from the effort (although I drank 72 ounces, I still lost 6 pounds of body weight or roughly 96 additional oz of body water). Considering the liquid loss from m body, on my skin, appeared over a gallon of sweat. On the bicycle, I rode an average of 13.5 MPH. Without this wind, I would have been drenched with sweat, dripping everywhere I went. (This is just like dumping ove1 quart of water of your body every hour.)

All of this effort, just to burn a little over 2000 calories.

You can replace all of those calories with the following:

1. Two Big Macs
2. 8 Powerbars
3. 2 pints of Haagen Dazs
4. Or a variety of other means

Let me be clear: I mean you can have 1 or 2 or 3 or 4. You can't have all of them. It is well within most people's ability to eat 8 Powerbars in a sitting. It is not, however, within most people's ability to bike my 62 mile loop.

It is really, really hard to burn calories. It is really, really easy to take them in.

See, if you had a metabolism that burned 2000 extra calories per day, you would be in a constant state of being "way too hot." Your heart rate would be constantly elevated, and you'd be twitching all the time because muscle have to move to burn calories.

Now, there are two other ways of not taking on weight, although we don't see them in humans: Non-shivering thermogenesis (which humans really don't do without drugs), which also would make you very, very hot. The other way is if your body simply doesn't digest the food that you eat. If you aren't digesting the food, this is going to be very, very obvious next time that you go to the toliet and examine your stool. If you are like everybody else, what comes out is very, very well digested.

In reality, no matter how much we'd like it to be different, everybody's body works pretty much the same. Now, could there be a variant in the population? The answer is obviously yes, but the range of the variation is going to be more along the 20-30% percentage range. Not the 100-200% range. And this variation may often comes because somebody simply fidgets a lot. Yet, it takes a lot of fidgeting to get skinny.

Below is a chart that shows some of this. This was taken from an issue of the American Journal of Clinical Nutrition. Although the chart is in mJ, you can extrapolate this to be kCalories. For roughly the same weight, certain people can burn 500 calories more than somebody else.

So, is there some variation? The answer is yes. Some people can eat 500 calories more per day than somebody else with the same body mass. This translates into one 36 oz Coke per day. However, this is at the extremes. On average, 80% of the people (at the same lean body mass) are within ~240 kCalories of each other. This is like a couple of glass of skim milk.

So, what does happen with that friend that "eats all she wants?" She really does eat all she wants, but if you measured the food that she eats, you'd find out it is pretty small compared to the people that are increasing in weight.

So why do we put on weight?

In a simple way, it is our brain. In our brain, our hypothalamus controls several fundamental drives. Two critical drives are thirst and appetite.

If your hypothalamus isn't working, you could get dangerously dehydrate, or you could starve to death.

The hypothalamus gets set at a preferred "fat setting." This is commonly called "your set point." What is interesting is that the set point tends to protect us against under eating and overeating. If you are suddenly forced to overeat, you will put on weight. If you are then allowed to stop eating, many people will find out that their appetite goes away until they get to a lower weight.

However, the set point is a bit tricky in that in a majority of people, once you have set a higher fat level, the set point goes up and doesn't come down. The set point also tends to increase as you age.

I'll repeat this for clarity, I have never seen any indication that the set point ever comes down.

So, what happens when you get below your set point?

You feel hungry. All the time. Hungry. This is why the vast majority of people will regain any weight that they lost on a diet. Your hypothalamus is telling you to "eat, eat!"

Now, take some comfort in that if you raise up your fat levels, you can push them down a bit without feeling hungry all the time. Therefore, some people have suggested that while set points are an okay terminology, you might better have "settling points." By eating the appropriate diet, you may be able to settle you body fat lower than you have it now.

What types of food help with controlling appetite?

Generally, we will have a tendency to eat less if we eat more protein and fiber.

What types of food may trigger a relapse? Generally, fatty food.

There is a research project for tracking people who lose weight and keep it off, "The National Weight Control Registry." There are some characteristics of this group.

1. They don't eat a lot of fat.
2. They cut back on their calories on a constant basis
3. They get physical activity
4. They eat breakfast
5. They don't watch television

Now, there are several other things that I believe contribute to keeping weight off due to some of the research and thinking that I've done.

1. Getting more sleep is correlated with lower fat levels
2. Having less stress is correlated with lower fat levels

Finally, the safest drug to temporarily trick the hypothalamus into dropping the setpoint is Bromocriptine, which foul mouthed, but highly intelligent Lyle McDonald has written about in his book "Bromocriptine."

If you are interested in this drug, let me give you the short version:

1. It fools around with your brain since it is a dopamine agonist (it activates dopamine receptors)

2. Although its weight loss properties are not studied extensively, it looks as if it does help with getting below the set point.

3. It is a temporary trick. Once you go off, the effect ends.

4. You need to do some work to get it.

5. It appears to be fairly safe, with a few annoying start up problems.

In my mind (pun intended), the best defense is a high protein, high fiber diet, with physical activity.

However, if all else fails, buy Lyle McDonald's book and decide for yourself on bromo.

Saturday, October 06, 2007

"Body" -> Part III: Ways of Using and Issues

Now that you've read the last two installments on tracking training stress, and you are now probably saying, "Hey, that stuff is for me."

Let's recap what we need.

You can track training stress on both the bike and the run. Of the two areas, the more precise is the bicycling instrumentation. It is also the more expensive.

In summary, what you need:

1. Bicycling: PowerMeter ($700-1500) and Cylingpeaks SW ($100)

2. Running: GPS ($100-300) and Topofusion Pro (Demo = Free)

As already mentioned, the GPS for the run is pretty simple. Go buy a Garmin forerunner. There are no other options.

Now there are a variety of different Garmins, and in my mind they are all just find as long as they can import data to your personal computer. The two that you really want to look at is the Garmin 205 (no heart monitor) and the Garmin 305 (with heart monitor). These are the latest Garmins as of the posting today. However, Garmin hasn't revised these models in a while, so perhaps a new one will come out in a short while.

On the bicycle powermeter front there are four major options for a prebuilt system.

1. Ergomo @ $1700
2. SRM @ $2100
3. Powertap @ $1000
4. Polar CS600 @ $700

What did I do? I bought the parts of a Powertap wheel. You can buy the hub and the computer in a slightly heavier "Pro" version for less than $800 on sale. Then I bought spokes and a rim for less than $50, and I made my own wheel. (I've made a lot of them in the past.) So, it cost me around $830 to get my system.

(Now, I also have a Polar CS600. However, for most people, the Powertap is the better choice for its rock solid reading, and ease of setup. The Polar, however, does work better if you have a Currie electric motor on your bike. If you don't have this, the Polar is a bit more fickle.)

Without getting into the details, the Powertap is probably the best system for an entry cyclists. It is accurate, and reasonably cheap. The other system that appeals to me is the Polar CS600, however, it is a bit difficult to install, and is fickle when you use some gears, which means it won't pick up the power correctly.

In anycase, you want the Cyclepeaks software.

So, you'll end up with a Bicycle system and a Running system.

While both systems give you an indication of training stress, they are clearly not completely compatible with each other. This is a frustration to triathletes like myself. What is clear, and not surprising, it is easier to recover from bicycling than to recover from running. Thus, while both systems give training stress, it is more important to make sure that you are adequately allowing for recovery from your GOVSS score.

According to Skiba in forum posting, the GOVSS should indicate the following:

  • <75>
  • 75-125 residual fatigue next day
  • 125-200 residual fatigue about 2 days
  • 200+ residual fatigue 3 or more days

After my run yesterday, I do admit I am a bit sore today.

Finally, the tools that are available today are all a bit sloppy. You cannot find a package that grabs all the performance data and pulls it into one place. You will need to continue to have at least three packages for the data:

1. Topofusion for running
2. CyclingPeaks for Bicycling
3. Something like Sporttrack for logging routes and weather

However, your tools for training smartly have come a long way.

Monday, October 01, 2007

"Body" -> Part II: Modern Tools For Training

In the first installment of this series of posts, we discussed the idea that you need to track your workouts. Not only do you need to do this from a chronological standpoint, but you should also track this from a effort standpoint. However, distance alone (or time) is probably not sufficient to capture the state of your training body.

In cycling circles Andy Coggan (exercise physiologist and national age grouper bicycle time trialist) has derived what he calls a Training Stress Score or TSS. While having good success in using this formula for his own training, Andy released the formula in 2003 on the Wattage mailing lists.

While the Wattage mailing list has a small following, it is a powerful following. Very soon there were many people looking to use this TSS idea, and commercial implementation was not far behind.

The most widely used package using Coggan's formula is the CyclingsPeak WGO+ package. The user must have a powermeter. Now, bicycle powermeters are worth a post of their own, but for simplicity, there are at least three commercially available systems on the market. I happen to own the Powertap model for my racing bike and I also have a Polar model for my commuting bike. (I will do a post on this later.)

A Powermeter for a bicycle allows you to do your workout, and it tracks and logs the effort (power) that you are putting into the back wheel.

After the training is done, the user downloads his data to WGO+ to see his TSS. You can also use an Excel spreadsheet on Powertap data, which is found here. However, the spreadsheet is a bit old, and may not work with all versions of the Powertap.

As already mentioned, the easiest way of getting this information is to buy the WGO+ package. On the left is an example of the ride that I did this weekend with my Powertap computer. After my ride, I took the Powertap computer, which is the size of a normal bicycle computer, to my computer. Hooked it up to a USB port, and downloaded my training session of 63 miles. As you can see, part of the summary data is a TSS score. In this case, my TSS was ~277 at an intensity factor of .811. This means that I went far, but I didn't go all that hard.

~277 is still quite stressful, and I can always feel it after my 63 mile loop. I'm tired! Now, once you have the TSS, you can start tracking this in all of your workouts. What you'll quickly learn is pretty obvious: shorter hard workouts have a bigger impact per training minute than long workouts.

By monitoring your workouts in this fashion, you get a much better external feedback into your training methodologies. This method of training is being implemented by all levels, and you can find the professional bicycle racer is even tracking his (or her) output in races.

The one part that I've skipped over in all of this is the "relative starting point." If you are superfit, a hard workout might be 120 miles of hammering on the bicycle. If you are just starting out, maybe a 10 miler may leave you sore. To get the base level for fitness, you are required to take a time trial (or race pace). While there are many ways of getting this information, the easiest method is doing a 25 mile (40 km) time trial. Once you have this data, you can then base all other workouts against this key workout. The key statistic in determining your base fitness is figuring the power that you can generate for roughly one hour. There is a place in the software to insert this data.

However, this is a blog about triathlons. We know that bicycling isn't the only sport in the triathlon. Has there been any work on other sports?

The question of how to do this for running has been, in some sense, solved by Philip Friere Skiba with his GOVSS (Gravity Ordered Velocity Stress Score). You do need to have a GPS for running and Garmin with their Forerunner series of GPSs have pretty much cornered the market on this application. You must have a GPS that logs the running data! A few of the cheap Garmin units do not log this data.

Once you have your GPS, you load it into TopoFusion, which has a free demo package of their software and decide to implement Skiba's formula. Once you have selected your run, you can do an analysis on your run, and you will see a box like the following.

Now, similar to the TSS of bicycling, the software needs someway of figure out what your base fitness level is. Therefore, you also need a time trial to establish base fitness. Once you have done this, then you can analyze any run in the package to get an "average power."

How do you do this? Average power is calculated for every run, and you'll use this calculated figure to set the base.

Similar to the bicycle methodology, you take a very hard run (or time trial), and it will give you an average power. Now you'll use this as the GOVSS base.

In this case, I generated and average power of 345 watts in a 6.5M time trial in my local area a few months ago. If you look at the box above, I then inputed 345 watts into the "input" box. This now becomes the calculated score.

I'll leave the average power there for a long time. However, if I got a lot more fit, I should take another trial to establish a new base.

With my svelte 200 lb frame on my 8 mile run last night, I had a GOVSS score of around 114. The lactate score is a prediction of the power that I would generate for a 1 hour run. The average showed what I did last night. Now, while my peak power was 345 Watts for my time trial, I "only" generated 286 Watts during my run. However, I went for a long time, thus while I went at a lower intensity, I went longer. Thus the stress score was pretty high.

Also, note that the Kcal expended is post conversion. The software shows calories as work, not calories burned in the body. When you do work, 80% of calories get lost as heat! Now you know why exercising makes you so warm. Working out at 200 watts is like having 800 watt heater on you.

If you keep in mind that the body conversion to work is around 4-5:1, my calories burned was around 1312-1640. Considering I climbed almost 700 ft in elevation during the run, and the 1312 calories doesn't sound too far off.

What is interesting, is that last night's training running is almost exactly the same GOVSS as as my 6.6M time trial, which took 52 minutes. So, it I was doing the same number of GOVSS per week, I could do it as 3 sessions of my 6.6M time trial or as 3 session of 8.1M longer run with more hills.