SethAZ 

I'm trying to figure out why rides uploaded from my Garmin Edge 500, paired with my new Stages L/R power meter, are showing higher power output figures than my Fenix 5 paired with the exact same power meter, for the same rides.

I've done three rides with the power meter so far, and I paired it with both the Fenix 5 and Edge 500 and uploaded both recordings of each ride so I can see how they compared. On the first ride the Edge 500 was giving me 200W avg. power over the hour-long ride, while the Fenix 5 reported 185W avg. for the same ride. I looked into the settings on both, and I had auto-pause configured on the Edge 500 but not the Fenix 5. I set up auto-pause after that on the Fenix 5, and ensured that the two units were configured as closely the same as each other in every way I can find. The settings aren't going to be perfectly identical, but very close.

The second and third rides showed a smaller power difference between the two of around 5-8 watts. Anyone have any idea how that might be? My ego wants to believe the Edge 500's slightly higher power readings are the most accurate, but I have no idea whether that's true, or the Fenix 5 recording of the ride is closer, or whether the truer figure is something in between.

I suppose it doesn't really matter. I should just pick one to go by, and if it's at least consistent with itself over time, any changes in power output should be reflecting the results of my training accurately, whether the actual power figure is truly spot on or not.

Anyhow, if anyone has experience with different gps units recording different power outputs from the same power meters over the same rides, and figured out why that might be, I'd be interested in hearing about it.

unterhausen

yes, do the thing about your ego. It doesn't matter if they match, just use one. Have to admit it's weird, they probably are doing smoothing differently. My guess is the 500 does less.

SethAZ 

It's funny that this should matter at all. If I only had one device I'd feel good about the data, thinking hey, I'm measuring power, the manufacturer claims it's accurate to within 2%, so this is cool, I actually know what performance I'm hitting. And then for the same power output I'm getting different results that are much greater than the 2% accuracy claimed by the manufacturer. In other words, the power meter may be 98+% accurate to what my body is actually doing, but I'm not getting that level accuracy reflected in my data due to how it's recorded. Doesn't matter really, but it is the kind of thing that will bug me. I'd almost have been better off not knowing, but for some reason I recorded the rides on both devices just to see what the data would look like.

Uploading rides through the Fenix 5 is much more convenient, since it's paired with my phone and auto-uploads activities as soon as I stop and save them. They show up in Garmin Connect and then Strava within seconds. With the Edge 500 I have to unmount it and plug it into my computer over USB. I do like looking down at the Edge 500 during rides, though, as opposed to having to look down at my watch on my wrist. I may find myself using both, and only uploading the rides through the Fenix 5, but looking at the data on the Edge 500 during the ride, then discarding it after the ride is over. That means all my historical data that's uploaded will be from the Fenix 5 from now on, which is the one showing the lower power output. It's still fine, but the ego is slightly annoyed by this.

One thing I haven't yet tried is disabling auto-pause on both units. I turned on auto-pause on the Edge 500 a long time ago because I was sick of having my average cadence data being skewed by periods of time when I was stopped at intersections and the like. The data still gets skewed a little by the time spent not pedaling as I coast into intersections, but at least it's skewed less by eliminating the time I'm actually stopped.

What I really want is a setting that says, essentially, that I want both cadence and power averages to be calculated only during periods of time where I'm actually pedaling. I want to know what I'm doing when I'm actually doing it, and am annoyed that my averages don't reflect what I did when I was busy doing it due to times when I wasn't doing it at all being included in the average.

unterhausen

Average power doesn't mean too much. There are platforms that will calculate more reasonable metrics, like TSS. TSS doesn't count when you aren't moving, but you do need to tell it an FTP.

SethAZ 

I hear ya. It's going to be a learning process to see what data I have, what are the most relevant measured or calculated values, etc. I've got a Training Peaks account that's still active but not being actively used at the moment (not premium), and I've gotten multiple emails about a new FTP being recorded from my rides since I installed the power meter. Those are just the highest power values yet recorded, not reflecting what my FTP would actually be if I pushed it as hard as I could sustain for some extended period. I haven't done that yet. I haven't ridden the last two days so I should be fully recovered from the 7 out of 8 days I rode previously. I still need to get out on a ride today, and I'm thinking I'll make it an approximately one-hour ride and push hard to the max of what I can sustain for that full hour. Then I'll have a pretty good starting value for my FTP that I can use moving forward.

SethAZ 

Ok, so this afternoon I did a 20.7 mile ride in just a hair over an hour (19.3mph avg speed), and the two Garmins measured within 1 watt of each other on power. This was a ride where I tried to push hard for the entire hour, in order to get a good FTP measurement. The Edge 500 reported 218W average, the Fenix 5 reported 217W. The best 20-minute power level was 234W, so multiplying that by .95 gives 223 W, which is what Training Peaks emailed me to inform me was a new FTP value for my account. The 223 W value is what I'm going to use, because the average of 217/218 W included several minutes up front before I'd come up to speed, and that last 1/4 mile or so at the end where I slowed down approaching my house through my neighborhood. So 223 W is what I've plugged in to the apps.

I don't know why why today's ride had very close to the same power readings between the two devices. I suspect that it's because today's ride was a more steady-state affair than my other rides, with fewer times I had to pull back at an intersection. I'm guessing the differences I saw the other days since I installed the power meter were on rides where approaches to an intersection or a little coasting around sharp corners or whatever were more of a factor. My prediction is that on long rides with few interuptions to the pace and cadence the two devices will show very close to the same results, but on rides with more periods of low or inconsistent effort the slight differences in how and when they record will show up in the stats.

The 223 W FTP was slightly disappointing to me, as I'd assumed based on estimates from various programs over the years that it would be more like 240 W, but those estimates were clearly too generous. Anyhow, now I have a baseline measurement, and it will allow me to track improvements over the coming weeks or months as I ramp up my mileage and improve my endurance and efficiency.

Cyclist0108

Power = work/time I am going to guess both Garmins treat the numerator the same and that the difference arises in the denominator, i.e., different moving times. That should be something you can check -- did the two record different moving times when they reported different power outputs? The last ride serves as a positive control -- I am assuming from your description that you didn't stop midway, or did something different in that last ride with respect to moving times.

unterhausen

Head units are just getting instantaneous power readings. Nothing more fundamental than that. They are filtering it somehow. Garmin never seems to re-use any software, so they are almost surely doing the filtering differently.

August West

Not the same situation but I've used an Edge 500 & Edge 830 to record a P2M NGeco & Favero Assioma concurrently. They were within 3 watts of each other. On the next ride I swapped the head units to the opposite power meter and got the same results. The Assioma's were slightly less than the P2M no matter which head unit I used. The point is even the way old Edge 500 seemed to work just fine and record accurately even though it didn't record the pedaling dynamics of the Favero's which was expected. There were slight differences in the times between the 2 and as was mentioned I'd bet that is the issue. But if it were me I would need to find out what was going on since that big of difference would drive the OCD part of me nuts.

I also have Fenix 3 but have never paired a power meter to it since I alway use a bike computer. Maybe I'll give it a try on my next ride just for yucks.

You could upload the .fit file from each and use DCRainmakers analysis tool to compare the raw data. It costs $5 for a day pass to upload up to 3 sets of data or $29/year and you could compare the raw power output between the two.

DCRainmaker Analyzer

ChrisAlbertson

I've worked as a software developer for a long time. It is surprising just how much the customers are unable to tell the engineers. So what I imagine is the engineers found the data from the power meters was "noisey" and varied so quickly that it was impossible to make the display stable. Each group came up with a different idea of how to fix this problem.

I see the same thing with using GPS to measure distance. The GPS receiver finds your location on Earth once every second. So it can plot a series of dots on a map. But is the distance a series of straight lines that connect the dots or a curve of some kind? If you are driving a car on the freeway it makes no difference but if you are hiking on a trail with switchbacks it matters a lot.

Little details in how they process the data matter at the 2% level. It is to bad that they can't tell you the details of how they process the data.

SethAZ 

I think you probably nailed it. I'm betting it was different moving times. On ride 1 the Edge 500 was set to auto-pause, the Fenix 5 wasn't, hence the 15W difference due to different moving times. The course I was riding on crosses multiple streets, and whether I stop for a few seconds or just stop-and-go (or, gasp, just kind of slow-roll through) is going to be a little different from ride to ride. On rides 2 and 3 I had set the Fenix 5 to auto-pause, but it still pauses and restarts a little differently than the Edge 500, so although there will still be differences in moving times, they would be smaller, hence why rides 2 and 3 showed smaller differences of like 7 or 8 watts between the two Garmins' output. Ride 4 (yesterday's ride) I managed to avoid anything more than instantaneous stops, and as I did my laps through this one neighborhood's ring road I managed to avoid having to fully stop at any of the intersections, so whatever moving time differences there were would have been even smaller, hence the 1W difference.

To confirm this I'll turn off auto-pause on both units. They should be within a second of each other, because I start and stop both units within about a second of each other. If that was the reason for the differing results, they should show either the exact same power results, or at most within 1W of each other. My ride today will be deliberately slower than yesterday's ride, because yesterday's ride was right up against my threshold heart rate for nearly the entire ride, on purpose, for FTP measuring purposes. I'll take it easy today so I recover enough that I can ride again tomorrow and Saturday as well, then take Sunday off. With auto-pause off on both units I should know by this afternoon if time differences were the explanation for what I've seen.

Btw, I'm thinking seriously about ordering a Garmin 530. While the Edge 500 has served me extremely well since 2014, I'm really, really loving the features my Fenix 5 watch has, including constant connectivity with my phone, instant auto uploads of completed activities, etc. The Edge 530 has all those same features, plus a larger and easier to read display. The reason I've been comparing the Fenix 5 with the Edge 500 for my bike rides is mainly due to the ease of use benefits of the Fenix 5. If I pull the trigger on the Edge 530 I'll have the best of both worlds.2

August West

FYI, I just got back from a short ride using my Edge 830, Fenix 3 & Favero Assioma pedals. Time was 1:15:01 with 9 second difference between the two. Both are set to auto-pause. Average power reported was 1 watt difference between the two with the 830 being higher while normalized power was identical. Elevation gain was within 7 feet on 1726' of climbing. Distance and ave. speed were screwed up because I forgot to change to the correct wheel circumference on the watch but I'm confident they would match if the correct circumference had beed set.

Seattle Forrest

Try turning auto pause off and compare. Are you calibrating the better with both devices?

SethAZ 

Ok, well I consider the mystery solved. I turned auto-pause off on both Garmins, and have "every second" recording turned on both as well, and after today's ride they were 1W apart on their average power reading. It has to have been the difference in time recorded because of differences in how the two units responded with auto-pause turned on and the fact that on my rides I typically have to stop several times at intersections, however briefly.
@seattle_forrest, the Edge 500 asks me if I want to calibrate each time I turn it on, and I do it, but the Fenix 5 only asked me to calibrate the first time I paired it. The number I'm seeing on the Edge 500 when I calibrate it hasn't really changed each time I did it, so I doubt that was a factor.

This also gives me a real lesson in why avg. power recorded isn't all that useful, given it can change so much just based on differences between recorded time and actual time and moving time and so forth. Now I need to familiarize myself with the other power-based stats like TSS and whatnot. Since I measured FTP yesterday and recorded it in my apps, I believe that TSS will reflect what percentage my ride's power output was compared to my FTP. That should give me a "zone"-like result similar to how I've always used HR zones in the past.

I did pull the trigger on a Garmin 530 today. I got a good deal at 15% off the normal price, so it was just over $250. I look forward to having the same connectivity and features that I've enjoyed having on my Fenix 5, while also having a slightly larger and easier to read and configure screen compared to my Edge 500.

Seattle Forrest

For what it's worth, you can calibrate your PM from your F5 even without being prompted. Long-press the menu button and I think "calibrate power" is the first option. Sounds like you don't need to, though.

If you're looking at average power, you have to include the time you were stopped at lights and whatnot; that's time you were resting, that rest allowed bigger numbers to follow, it's part of the picture. For what it's worth, because auto-pause needs some time to realize you're moving again, you also miss out on starting, a lot of people throw down more watts starting from a stop than cruising. So leave auto-pause off, get the full picture.

Normalized power (NP) is the main thing to look at if you've been using average power. It's what your average would have been if you rode at a steady state. Training stress score (TSS) is what a ride took out of you. You can get big or small TSS numbers with a high or low overall power, because it includes time. Probably the most useful thing though is the power curve from a ride, since you're using Garmin, it's near the bottom, below all the charts, in with "stats" "laps" "time in zone" etc.

caloso

Beyond the auto-pause, Garmin has a setting that will ignore zeros, so even if you are moving but coasting, it will ignore those readings.

SethAZ 

Holy Moly, I was about to post that when the Garmin Edge 530 that I ordered yesterday arrives, I'd have to go through the menus and see what my options are with regards to zeros and recording, but I went to track the package and found it was already sitting on my front doorstep. It's plugged in right now to top off the battery.

I'm not made of money, and buying two new cycling toys within a week is out of character for me in some ways. I think it's a subconscious coping mechanism for taking my mind off the end of the world. Anyhow, to celebrate the fact that I'm still alive and, as yet, uninfected, I'll go out this afternoon, pair the Edge 530 with the power meter, and take it for its maiden spin.

I do appreciate all the comments in this thread. The different power readings were unexpected to me. The explanation for that was a real mystery to me at first, though now it seems perfectly understandable. It just took some figuring out, and I appreciate the comments here helping me do that.

Seattle Forrest

You do want zeros included in your average power. Simplest reason: they happened. There are only a handful of things AP is better for than NP (average, normalized) and they'll all lead you down there wrong path if you omit zeros.

SethAZ 

Hmm. I'd be interested in hearing more about why you think the zeros should be in there. Yes, they happened, and I'm not trying to hide that fact from myself. The reason I didn't want the zeros included is because I wanted to use average power to reflect what power output I was generating during the times I was pedaling. It's possible that other stats are more useful than this, and I've started learning about these. On your advice I just turned zeros back on in the averaging settings on the 530.

I'm actually having a little issue I'm trying to deal with right now, where L/R torque balance and torque smoothing showed up in my ride data from today on the 530, but it shows 66% L and 0% R, so I've unpaired everything and started over, but haven't done a test ride yet to see if that fixed things.

On my short ride today I recorded it on the Fenix 5 and the 530 in order to compare them, and with zeros NOT in the averaging the 530 was nearly 20W higher. While that doesn't reflect the true average over the whole ride, it does at least tell me that my output while actually pedaling was nearly 20 W higher than just looking at the avg. from the Fenix 5 was showing it as. I do recognize that I should probably just ignore this and look at NP instead.

Seattle Forrest

A lot of people think normalized power is average not counting zeros, you have the data to prove that's not true. That's a random aside, but you said you wanted to hear more about this.

The zeros are rest. You wouldn't be able to produce as much power without them. It's like how Strava time doesn't stop when you pause your Garmin.

Your Garmin tells you your best 20 minute average power from each ride. Imagine riding as hard as you can for 10 minutes, resting an hour with auto pause, and then riding hard for 10 minutes again. An extreme and contrived example to illustrate the point: AP wouldn't tell you anything useful.

Power is tired together with time. Most people generally assume your FTP is 95% of your 20 minute power. I'm a natural sprinter, my 10 second power is impressive but you'd laugh at my 2 hour averages. I train to my weakness, it's easier to do that and evaluate progress with the zeros included because it's a more accurate picture of what I'm capable of.

Obviously, it's your choice, I'm just trying to share things I've learned, that felt counter intuitive at first.

If you don't stop pedaling often, you can use the lap button to get the average for the periods you're pedaling.

* Getting back to the idea of splitting a 20 minute ride up. There's a thing called "the beer and burrito rule," if you stop long enough to enjoy those, it's two rides. That's because of the way normalized power is calculated, and down the line you'll probably want to use TSS for "scheduling" and TSS includes NP.

SethAZ 

Awesome. I'll look up TSS and do some reading about it. On my rides so far the zero times are restful by definition, though they are only occurring because I've come up to an intersection, and I don't believe my output would have dropped had the intersection not been there. I do appreciate your tips and advice here, and I'm looking these things up, then looking at my data as it comes in to try to see what they're about and how to use them. I recognize that I'm approaching this from a position of ignorant and assumptions made prior to having power, and prior to having read up on the stats collected or calculated using power data, so I'm flexible in my opinions and approach.

I discovered that the left and right halves of my Stages LR power meter were not in fact linked together from the factory as they should have been. This means the power data from my first four or five rides with it were using left-only power, which was then doubled. I linked the two halves using the Stages app on my phone, re-paired the power meter with my Edge 530 and Fenix 5 and now not only am I getting proper dual-sided data, but on my new Garmin 530 am all of a sudden getting all this interesting L/R split data. It revealed immediately that I've been pushing harder with my left leg to varying degrees (that changes with increasing cadence), so now I'm watching for that and trying to pedal more evenly with both legs.

I was able to cut the zip-ties and remove the Garmin cadence transmitter from my chainstay and pull the rare earth magnet from the pedal stud on my left crank, since the Stages power meter sends internally generated cadence data. That's one less accessory strapped to my bike.

I've seen what I needed to see by recording rides both on the Fenix 5 watch and on the bike computers, and now that I'm confident that everything's set up and working properly on the Edge 530, I'll stop also recording the rides on the watch. As for how I'm feeling about my virus-anxiety-driven comfort purchases, I'll say I'm delighted with both the Stages gen 3 L/R power meter and the Garmin Edge 530. I've got way more info about my rides to geek out over, help me out in planning my rides, and the 530 has new connectivity and features that I've really loved on the Fenix 5 watch, and is a real step up from my trusty and faithful Edge 500.

sean.hwy

I have Fenix 5 also. I have been using it for hiking and biking. I have it paired with garmen cadence sensor. It has been working great.

I have on back order assioma duo power meter pedals.
Should I get a garmin 830/1030 computer or is the watch good enough?

Seattle Forrest

The watch will work with your PM. Neither is inherently better. If you're going to do power based intervals the larger screen on an Edge is useful. But on a technical level they're equivalent.

SethAZ 

My eyes aren't all they once were, and I find it very annoying trying to peer down at a watch like the Fenix 5 while I ride. The larger screen, centrally mounted, of the Edge makes glancing at it many times throughout a ride much more convenient. Was it worth the extra $260 or so given I already had the Fenix 5? To me it was, but others may disagree. Yeah, they seem to be from the same era as far as technology goes. For me upgrading from the Edge 500 to the Edge 530 was all about getting the technology level of the Fenix 5 with a much larger display capable of showing more data that I find useful during a ride. No regrets.

If you use cadence or heart rate data (or power) a lot during a ride it can be very convenient having a Garmin with a larger screen firmly mounted on your handlebar, rather than trying to glance down at your wrist during a ride. Whether that's worth the money to you personally is up to you, but it was worth it to me. I look down at my Garmin quite often during a ride, and it would be much more annoying and difficult trying to do that with the Fenix 5. I love the Fenix 5, btw, and when I'm running, walking, or swimming it's Fenix 5 all the way, but for cycling I want a Garmin Edge on my bike.