AeroPod, PowerPod and Newton

I've been doing some tests recently to see if I can determine what causes these two factors to drift in the hopes of getting better accuracy. Now that the iBike self-corrects the slope based on the altitude, that is no longer a serious problem... but unfortunately on most rides there is no way to make sure that the determination of wind speed is correct, since the wind is usually different on different parts of the course and there is no way to check it.

As an example, I typically will see a variance of +- 2 to 4 in the wind offset value during a ride. If you go into the software you can see how much effect this has. On a recent 50 miler, increasing the wind offset by 2 points resulted in the avg power changing from 211.7W to 199.8W... about a 6% difference. So even if all the other measurements were accurate, this is a pretty large error.

I'd hoped that maybe it was consistent with temperature changes, but so far my testing indicates that this is only partially true. I temperature cycled my unit in the house, using a lamp to warm it and the refrigerator to cool it. I left it in each spot for at least 1/2 hr to give it time to stabilize. These are the results:

Temp (F), Offset
69, 0.0
99, -1.7
39, -4.3
65, +6.8
91, +5.0
71, +1.1
41, -2.2
67, +5.4
70, +1.1
74, +2.2
69, +5.0
68, +3.8
69, +4.0
67, +4.6

The next day I zeroed it again
Temp (F), offset
67, 0
112, +.5
37, -7.6
67, +2.0
101, +3.5
67, +3.7

So far there does seem to be much of a temperature effect at room temperature and higher, but there is a consistent drop in the offset value at lower temperatures. Unfortunately there is also a trend of the offset increasing during the day which is independent of temperature.

I also took a notecard and pen along on a ride and stopped several places to record the offset, temperature, and altitude to see if there was a correlation. The temperature during the ride was ~65 (~75 with sun), so close to room temperature.

Altiutude (ft), Temp (F), Offset

7330, 67, 0.0 (home)
6148, 76, 1.0
6880, 78, 1.0
7088, 78, 2.5
7315, 72, 3.0
7321, 66, 4.5 (home)

The only correlation I can see is that the wind offset gradually drifted higher. I thought maybe altitude would have an effect, but that wasn't shown here.

So... I wonder if there is any rhyme or reason to how your wind offset varies? It's pretty easy to do these experiments... you just have to make sure that your iBike *never* zeros the offset on it's own.

I agree that this is an unstable measurement. I can just turn on the iBike at home, auto wind and then go to set up. Most of the time, the readings start out close to if not 0. Come back a few minutes later after the iBike has been idling and you can get + OR - 3 mph and then zero it again. This "zero" seems to have a lot of drift in it - which is also going to throw a wrench in any wind scaling correction and ultimately have an effect on CdA.

Tom


Yes - I think you are correct Fernando. Still it is not stable.

FWIW, the number in the wind offset is not MPH, I think it's pascals. Search the forum there was a previous discussion about this.

Yep... it is supposed to be in Pa (N/m^2) and this can result in a large airspeed error at low speeds. As you go faster the airspeed error goes down, but the power error goes up.

I just finished a 25 mile loop ride, stopped after 39 minutes at DQ for one of their, now on sale for 1.99, banana splits!

Of course at start, wind offset was zero'd. When I stopped at DQ, wind offset was about -0.8.
When I left DQ, wind offset was zero'd. Back at home wind offset was 0.

The temperature profile varied a bit between weather stations but roughly peaked about when I started out at 61F and held for awhile then dropped to 56F at the end of my ride.

Usually I stop at about 5-10 minutes out and rezero, I did not do that today. John, I think, suggested rezeroing after about 5 minutes or so.

Overall for the loop, Analise Wind tool showed -0.3km for the whole thing.
(39/142) * .8 = .3 This shows that the 142minute ride with -0.3 at 39 minutes and 0 from 40 to 142 minutes exactly works out to yield the Analise Wind result. For whatever that is worth.

This was with my iAero which behaves the best over the iPro backup unit.

Russ

Do you normally get a wind offset that is that stable?

Another interesting thing happened recently... right after I did the temperature cycling, the wind scaling appears to have dropped from 3.1 to ~2.1. It has held steady at that for about a week. I have no idea why that would happen.

Here is my results for a 55 mile ride today.

Altitude, Temp, Offset
7330, 72, 0.0 (home)
6749, 79, 0.0
7511, 75, +0.5
8164, 58, +4.0
6693, 63, +5.5
7223, 58, +3.5
7381, 72, -.8 (home)

So... I'm still not seeing any effect from altitude. There looks like a possible effect from temperature, but this is exactly *opposite* the effect I got earlier when testing in the house. Then the cooler temperatures produced a negative offset and here it was a positive one. From what I can tell so far the variance looks random. This isn't good since being off by 5 on the offset makes a big difference in power. I noticed it increase pretty suddenly on a long climb, when the power increased ~30W and my perceived exertion did not. The 30W increase put me over FTP, so I knew it was wrong. On the way back down the mountain I was getting 500W without a struggle.

I would appreciate seeing what other people get if anyone is interested in taking the time.

Miscellaneous ramblings:

@ rruff You mention the wind scaling dropped from 3.1 to ~2.1. Permanently?

I'll pay more attention to wind offset on future rides. However, any time I notice a significant rise or drop in temperature AND I happen to stop at either an intersection, rest stop, whatever, I do a wind offset. I cup my hand over the unit and press the center button. I don't obsess over it but I certainly don't ignore the potential for wind offset error. When I get home I always check the wind offset to confirm I'm fairly close to zero. So far so good.

I placed a few drops of soapy water on the clear window today and blew air through the wind port. I was expecting to see bubbles but there were none. So my unit is pretty much airtight. I can see if there is any airflow around the window then a piece of clear tape would make a huge difference in wind scaling.

Another thing I do is glance at the elevation at times and if because of a large change in barometric pressure the elevation is off, I correct it. Obviously the auto tilt works off the barometric elevation but I don't know if the barometric elevation ever looks at the displayed elevation. I do know that analyzing the power in iB3 definitely adjusts the calculated barometric elevation to the stored elevation number. I took a nearly flat ride and forced the elevation column to increment by 3" every second. So at the end of 19 miles the stored elevation went from 30 ft to just over 1000'. Analyzing the power took an average power of 136 and raised it to 208. Just a test but the power profile is adjusted post ride to the stored elevation. Anytime I know the elevation and there is error, I adjust it. The attached screenshot is a flat ride that I manually changed the elevation. You can see the software adjust to these numbers.

racerfern wrote:@ rruff You mention the wind scaling dropped from 3.1 to ~2.1. Permanently?

So far. I've heard of this happening due to rain, but not warm and/or cold temperatures.

If you can zero your wind scaling often during a ride, then that should help... but this doesn't always happen. Around here I basically never need to stop on a ride for other reasons, and if it's a race then you can't. A few days ago I did a 35 mile loop with a couple other people and we stopped once... the offset read +3 and I zeroed it. I noticed it reading too high later and when I got home it was +7. I seem to have had issues with the offset creeping up during rides lately, but I know this summer it often went in the other direction.

rruf asked:
"Do you normally get a wind offset that is that stable? "

As mentioned earlier, above 70F for the iAero and above about 83F for the iPro is usually very stable for me, excepting possibly a need to ride a few minutes with my black iAero if it has been sitting in direct sunlight, which I usually avoid so I am not totally sure that is a problem in the summer.

I do notice greater swings with the temperature dropping more and at colder temps. Now my iPro may swing as much as you experienced on your 55 mile ride. I wonder how well iBike 3 software corrects the wind speed after the fact and how well, after the correction, the software's offset value correlates to your ride observations? For example does it kind of come up about in the middle, assuming an even amount of time in each temperature range? Or if you could weight each temp range for the greater and lesser times to try to correlate it?

I also wonder how much temperature latency between displayed temperature and actual air temperature (say due to a rapid change in elevation) might make the calculation less accurate?
This might be true if the air pressure sensor (being exposed to the air more directly) sees a different temperature air than the iBike temperature sensor feels due to the several minute delay we know about when starting up a ride and when bringing the unit first outside.

This is why I said it would be a great feature to add to track offset corrections by temperature and build an improved correction table in the ibike for on the road real time use.

Russ

Like many other sensors - more that the ones used in the iBike - I wonder about linearity and zeroing for the wind scaling, let alone the unit-to-unit consistency of any corrective response. So if the factors involved in the WS stability and so on are predictable, do corrections post-ride (offset and zero) made in the software really do the right thing?

So there may be thermal effects, at least at some range of temperatures, and not sure what else to include and also a specificity to each unit that may not be corrected in the current software/firmware.

I have, a couple of times, taken my iBike and started it up at home, zeroed everything on the bike, then driven the bike on a car rack to a park - 2 miles away. - Several observations when I do this - the altitude can go crazy with a few hundred feet shift when there should be virtually none and also the wind zero shifts. - Some of this may be thermal from just the air movement. NOTE that the wheels are not rotating so that the cadence or speed are being recorded as part of a ride.

I do not see this altitude delta when I mount start the iBike at the park. What I do see in the other cases is the same effect as a massive tilt correction because of the altitude change in the first .1 to .25 miles.

You can't use software to correct for wind (scaling or offset) unless you have reason to believe that it canceled out... which won't be the case most of the time.

If you start the unit and then change altitude via driving the bike somewhere, then it will think that you were on a very steep slope. To avoid this, reset the unit before you start riding.

rruf,

A nit:
"which won't be the case most of the time"

This seems to be the prevailing claim or statement about this on the board but wait a minute...

Much feedback exists about start and end points being the same for many rides. 95% or more of mine are that way.

These rides present an opportunity, as a class, for correction and evaluation. Also this is the case that the wind tool in the software exists for.

Another correction that should be added to the software is one for the loop ride case and would correct the altitude, start to finish, on the basis of same start and end points. While this might not be perfect it would be a step closer to perfection. Even if not same start and end, if you could select a two points that constitute a sub loop that is large enough to be meaningful, that could also provide a good correction to the whole ride. Of course this isn't usually needed if the barometric pressure stays the same through the ride (occasionally for me). I don't know but maybe the more accurate alitude profile would then allow the wind correction to be more accurate as air pressure affects aero power and the pressure against the sensor based.

Russ

My rides are essentially complete circuits where my start and finish are the same and all points between are covered out and back. So this is not like riding "around the block" but rather the true retracing of the path in the opposite direction.

Tom

In that case so long as the wind doesn't vary and is the same on both sides of the road, then you should be ok with zeroing the wind. Neither of these is normal where I live.

Did another 35mi ride today. Temperatures were right around room temperature throughout, but a bit cooler towards the end, so I stuck the unit in the frig to see if slightly colder temperatures could be the cause of the increasing offset. At 32F the offset went from zero and 70F to -9.5... so cooler temperatures seem to make it lower, not higher.

0.0 (home)
1.0
3.5
5.5
3.0
0.4 (home)

This last being after it had been in the house for awhile... and the 3.0 value occurring right at the end of the ride but before I came in the house. So the habit of it increasing during the ride seems to hold true, but there is no correlation with temperature or elevation.

Did another 50 miler and this time the wind offset stayed +-1.0, which is nice.

Anybody else noted their variations during a ride? Trying to get an idea of what is "normal".

+/- 1.0 is completely within the norm and is similar to my experience with my unit.

What about a random drift of 5 or more?

Rode four 50+ milers during the T-Day holiday week.

Temps: 50s, 60s, and one day at 70.

Each day, the temps rose during my ride.

Each day, I reset the wind as much as possible. (Not many opportunities on group rides).

Each day, the wind was about 4 to 6 points off.

Each day, the power readings seemed about 20-30 watts low.

If my power meter had a wind offset of, for example, 5 during large portions of the ride, how do I use the software to adjust that? Should I set 'avg ground speed' to zero or should I set 'wind offset' to zero?

You're better off IMO doing a wind offset adjustment than trying to second guess the iBike. It's easy to display wind as the top number with power in the middle. Evaluate your riding conditions and you'll know if you need to do a wind offset.

Doing wind offsets during a group ride depends on your group ride. I'm used to significant temp changes along the Pacific coast, especially when we head inland for a bit. I have to consciously think to reset my wind offset at any rest stop or given enough time at a traffic light. I don't go nuts about it, but I do pay attention when the temperature changes more than 5 degrees, which is a huge amount in the scheme of things.

Simply cupping my hand in front of the unit and going to wind offset works just fine. When entering the wind offset screen I do wait a second or two for the unit to settle in, then hit the center button. It's not more effort than is needed by some other PMs to adjust to current temperatures.

sweet spot wrote:Each day, the wind was about 4 to 6 points off.

Not a wind offset problem... it's a bad wind scaling.

How far off was the wind offset each time you set it?

Thanks. Still getting to know the jargon.

On my ride Sunday, during which temps went from 61 to 66 and back down to 62, I reset the wind four times. Before the ipro scaled down, the number it showed ranged from 4 to 6.

Colder weather == weaker battery == scaling problem?

I'm stilling learning my way around the wind analyze function in the software, so I don't know whether that could correct any ill effects related to the wind scaling problem. What I do know is that the ipro gave realistic power readings after a wind reset --- but not for long.

Frustrated because I use the wko software to monitor my form. If I can't get accurate TSS scores for these 'winter' rides, I can't really trust what the chart says about my progress.

sweet spot wrote:On my ride Sunday, during which temps went from 61 to 66 and back down to 62, I reset the wind four times. Before the ipro scaled down, the number it showed ranged from 4 to 6.

That is neither a large temperature variation nor cold. I doubt I will see a temperature in the 60s for a few months now...

So you are saying that each time you went to the "wind offset" screen, the iBike was showing 4 to 6? And then you reset it and it went to zero... but then the next time it was up to 4-6 again? I noticed something similar on a couple rides.

Temps in the 60's should not be a battery issue. Yes, a jump in temp can, it seems, change the wind offset.

The battery will give you noticeable trouble in the upper 40's to mid-50's and down (F). I have not tried it, but I suspect sub-freezing (0C or 32F) and I would not expect the battery to run very long at all - but you would also not find me out biking


Tom

rruff wrote:

sweet spot wrote:On my ride Sunday, during which temps went from 61 to 66 and back down to 62, I reset the wind four times. Before the ipro scaled down, the number it showed ranged from 4 to 6.

That is neither a large temperature variation nor cold. I doubt I will see a temperature in the 60s for a few months now...

So you are saying that each time you went to the "wind offset" screen, the iBike was showing 4 to 6? And then you reset it and it went to zero... but then the next time it was up to 4-6 again? I noticed something similar on a couple rides.

Yup.

Get some more data points and report back. It might be better to *not* reset it during the ride and just note what it is.

Will do.