Crazy Power Reading

[David.C.Anderson]

All,

I have completed my Coast Downs and 4-mile ride, and while I am not excited about the derived values I do believe that they are probably fairly accurate (I guess it is really true that big old heavy guys have much poorer values for CdA and crr).

I went out for a ride today over lunch (not too long - but long enough to observe some strange anomolies), and was surprised to see my peak power generated going downhill with the wind at my back and pedaling very leisurely (still pedaling however).

PowerSpikeGoingDownhill.jpg (60.75 KiB) Viewed 13085 times

Later in the ride, I attempted two efforts of roughly 5 minutes each with the wind in my face and on a road that is uphill or flat (no downhill portion), and found that my power often showed 0 (presumably because Cadence fell to 0 frequently). The strange thing though is that I never stopped pedalling during these efforts, and the cadence held constant through other parts of the ride.

Power_and_Cadence_Anomolies.jpg (83.72 KiB) Viewed 13085 times

Expected Cadence Output:

Steady_Cadence.jpg (65.41 KiB) Viewed 13087 times

Questions:
------------
Why would Power be so high going downhill and with the wind at my back (I presume this has nothing to do with the strange cadence anomolies)?
What could cause such extreme wierdness with the cadence readings?
Does cadence have to be correct for Power to be correct?
As long as cadence is non-zero, should the power captured be accurate?
If the answer to the last item above is yes, then why would power be so low (hard effort - definitely above VO2) compared with the downhill readings?

Clarification:
---------------
Batteries on power unit are good (285 after ride).
Wind offset was performed prior to ride, and temperatures outside and where iBike is stored are +/- only a few degrees.

Thanks in advance for any insight you may have.

-David

[marshr]

Make sure you did not have the trainer mode on, i did that once and had some odd powers and did windoff set and tilt on road then when got home noticed the trainer mode was on, turned off and everything back to normal power.

[David.C.Anderson]

Yep - Trainer mode was off.

I was bit by that once as well. Good idea, but does not fit this scenario.

thanks.
-David

[Site_Admin]

David - the images are good, but send me the files as well if you could.

[David.C.Anderson]

See attached (ride file and exported profile)

[Velocomp]

I looked at your ride file and the primary issue is the unstable reading you're getting from your cadence sensor. Remember that when cadence goes to zero (as it does far too many times in your ride file) then watts will be forced to zero too.

I suspect that your cadence magnet and sensor are too far apart, resulting in the unstable reading. Make sure to follow the directions to get a good and solid cadence reading.

Until your cadence signal is stable it's pretty difficult to figure out what else, if anything, is going on.

[David.C.Anderson]

I understand that if Cadence = 0, then Power = 0. That makes complete sense. What I do not understand though is if/how the cadence value factors into the power calculation for the times that Cadence is non-zero.

So, I think that a couple of my original questions are still valid here:

1. Why would Power be so high going downhill and with the wind at my back (I presume this has nothing to do with the strange cadence anomolies)?
If you look at the 1:00 minute period of time on my ride (at around the 4:30 to 5:30 minute mark - roughly), you will see that Cadence is non-zero during this time. At this time, my Avg Power output is around 300 watts and I am doing nothing more than spinning the pedals. The elevation consistently decreases during this time and the slope is negative as well. In addition, I was headed east at this point and the wind was strong out of the west. This period of time is clearly not impacted by zero cadence values, yet I can be certain that the power here is too high - far too high - for the effort exerted.

2. What could cause such extreme wierdness with the cadence readings?
It was mentioned that the sensors are probably too far apart. What is too far apart? My sensors are probably around 4 or 5 mm apart (this is rough as I do not have precision equipment to measure these distances accurately). They are definitely closer together than my speed sensors and there appear to be no obvious anomolies with speed. In addition, I get consistent Cadence readings when on the trainer and have not had any problems detecting Cadence at all since my mount was replaced a month or so ago. This seems to me to indicate that there is something else affecting Cadence here, but I have no way of being certain (just looking at the behavior and symptoms of this issue).

3. As long as cadence is non-zero, should the power captured be accurate?
This is really related to Q1 above.

4. If the answer to #3 is yes, then why would power be so low (hard effort - definitely above VO2) compared with the downhill readings?

[travispape]

I understand that if Cadence = 0, then Power = 0. That makes complete sense. What I do not understand though is if/how the cadence value factors into the power calculation for the times that Cadence is non-zero.

So, I think that a couple of my original questions are still valid here:

1. Why would Power be so high going downhill and with the wind at my back (I presume this has nothing to do with the strange cadence anomolies)?
If you look at the 1:00 minute period of time on my ride (at around the 4:30 to 5:30 minute mark - roughly), you will see that Cadence is non-zero during this time. At this time, my Avg Power output is around 300 watts and I am doing nothing more than spinning the pedals. The elevation consistently decreases during this time and the slope is negative as well. In addition, I was headed east at this point and the wind was strong out of the west. This period of time is clearly not impacted by zero cadence values, yet I can be certain that the power here is too high - far too high - for the effort exerted.

2. What could cause such extreme wierdness with the cadence readings?
It was mentioned that the sensors are probably too far apart. What is too far apart? My sensors are probably around 4 or 5 mm apart (this is rough as I do not have precision equipment to measure these distances accurately). They are definitely closer together than my speed sensors and there appear to be no obvious anomolies with speed. In addition, I get consistent Cadence readings when on the trainer and have not had any problems detecting Cadence at all since my mount was replaced a month or so ago. This seems to me to indicate that there is something else affecting Cadence here, but I have no way of being certain (just looking at the behavior and symptoms of this issue).

3. As long as cadence is non-zero, should the power captured be accurate?
This is really related to Q1 above.

4. If the answer to #3 is yes, then why would power be so low (hard effort - definitely above VO2) compared with the downhill readings?

David,

Let's get your cadence detection working before we make too many conclutions on this particular ride. It might be that there are more cadence drop-outs than shown in the cadence data and that they are having an effect on the power data. Usually the cause the the bad cadence data is that the magnet and detector are too far apart. I've had cases where the magnet slips along the crank and also situations where the sensor rotates on the chain stay, so check both. If you saved a file from the day that you did your calibration ride, you can also check to see if cadence was a problem on that day in which case the you need to redo it. A bad calibration due to cadence drop-outs would definitely hurt all of the analysis based on it.

Another thing that needed to be repaired in this ride is the tilt, and this is something that would not be messed up by the flakey cadence signal. For most of the ride your unit started pointing up a little too much meaning that the iBike over-reported power for most of the ride. Correcting this brings down the power at the downhill, but it is still more than "just spinning".

Also, I looked at that downhill section in your data and your unit was on the verge of being overloaded with road vibration. It was not marked as an area of high vibration, but it almost was and had it been it would likely have been repaired by the iBike2 algorithm for doing so.

Another thing to keep in mind is that going into a more aerodynamic position on a downhill than the position that you used when you did you calibration ride would cause positive power. For example, if you did your coast-downs on the brake hoods with straight elbows, but tucked during that downhill at 4:30 to 5:30 while pedaling, your power would be over-reported.

Again, before we spend too much time debugging this particular ride file, let's get your cadence signal working right and make sure that your profile is good. There is no doubt that the iBike will give you very accurate power numbers once it is set up correctly.

Travis

[David.C.Anderson]

The cadence sensors are solid - cannot move them in current position (without providing significant force), so I guess it is possible that they are slipping but I would doubt it.

So, let's go with the possibility that my calibration is off. I have saved and attached all the coast-downs and the 4-mile ride that was done immediately after. Are you able to determine whether the calibration is good from this data? I should mention thought that the first 5 or so coast-downs were done whilst in "Trainer Mode". After realizing this, I switched the "Trainer Mode" off and did several more along with the 4-mile ride. I am also including the profile that I used after excluding those rides so that you can look at that as well.

As far as the more aero position, I do not think this was contributing. I performed all of my coast-downs on the hoods. During my rides, I ride on a combination of the hoods and the bars (the bars being less aero). I do not recall whether I was on the hoods or the bars at this point, but I know that I was not more aero than the hoods. So, the worst case here would be that Power should be under reported. Correct?

At any rate, the files are attached. Let me know if there is anything interesting or conclusive that is found.

DavAnder_Coast-downs_and_4mileride.rar

(50.01 KiB) Downloaded 346 times

Thanks.
-David

[travispape]

I have saved and attached all the coast-downs and the 4-mile ride that was done immediately after. Are you able to determine whether the calibration is good from this data?

People often assume that the short ride files that you see in the download window after doing a set of coastdowns are the actual coastdown rides, but that is not the case. Those short files are the rides between the coastdowns. The coastdowns do not get saved as normal ride files. However, with the new version of the iBike2 (2.0.6), there is a button to save a support file including information from the coastdowns. This option is available in the window for processing the calibration ride.

As far as I can tell from looking at your 4 mi ride, it looks like the coasting you did was probably actual coasting and so that is good; however, most people end up with a negative riding tilt, especially riders your size riding on the hoods. Maybe the profile is ok, but I wouldn't expect the positive riding tilt.

I should mention thought that the first 5 or so coast-downs were done whilst in "Trainer Mode". After realizing this, I switched the "Trainer Mode" off and did several more along with the 4-mile ride.

I haven't tried to do coastdowns with trainer mode enabled, so I'm not sure how that would work but I think there is a chance that the coastdowns were ok.

As far as the more aero position, I do not think this was contributing. I performed all of my coast-downs on the hoods. During my rides, I ride on a combination of the hoods and the bars (the bars being less aero). I do not recall whether I was on the hoods or the bars at this point, but I know that I was not more aero than the hoods. So, the worst case here would be that Power should be under reported. Correct?

That's correct if you are less aero on the tops; however, there are some riders who are more aero on the tops than when they are on the hoods. While moving from the hoods to the tops will sit you up more making you taller, it also makes you skinnier. There's a chance that you catch less wind on the tops.

Regarding the cadence sensor, often the sensors pick up better if the magnet passes near the end of the sensor--off center. The wireless sensors have a mark.

Travis

[David.C.Anderson]

Travis,

I went back and performed a quick test on the tilt calibration, and as you had mentioned I believe the tilt was off (I had a reading of +2.1 in one direction and a reading of +0.1 after turning 180 degrees). I readjusted my bar extender, re-positioned my wireless mount, and then cranked everything down about as hard as I could (not sure that I will be able to remove these without some sort of bicycle surgery).

At any rate, I did a set of coast-downs and a 4-plus O&B ride this evening in hopes of ensuring that my profile is now correct. Is there any way that I can tell whether these settings are valid (I do not have the luxury of having a DFPM to use a baseline for comparitive analysis). I am somewhat re-assured by the relative consistency of the Crr variable (Crr is now .0087 and was .0083 the first time). The other values are not quite as consistent:

Aero is .475 and was .425 the first time
CdA is .384 and was .435 the first time
Fric is 19.921 and was 9.135 the first time (OK - this is no where close)

Please let me know if there is any way to verify or validate that I am now using a correct (or at least a more accurate) Profile.

Thanks.
-David

BTW, I did not appear to have any Cadence drop-out issues on this ride. I changed nothing related to the cadence sensors. I am still confused on this one.

[Ccummings]

i had the same thing happen today, watts were 1600 on a descent, i also noticed that the current % elevation was 7% , so something with the tilt and the position must be a big factor

[David.C.Anderson]

I am far from an expert at this, but you might try to verify your tilt settings:

CHECKING THE ACCURACY OF YOUR TILT CALIBRATION
Once your tilt calibration is complete you can check the accuracy of your calibration very
simply.
1) Look at the slope number in the top of the main tilt screen. This is the slope of
the road where your bike is now pointing (1.2% in this example):

tilt_calibration1.jpg (5.38 KiB) Viewed 12565 times

2) Turn the bike 180 degrees, being careful to place the wheels on exactly the
same marks in the pavement where they were resting prior to being turned.
3) Look at the number in the tilt display; it should be the negative of the number
you had prior to turning the bike:

tilt_calibration2.jpg (5.58 KiB) Viewed 12565 times

4) If the magnitude of two numbers differs by 0.2% or less then your tilt calibration
is good. If it is greater than 0.2% you should repeat the tilt calibration.
5) IF YOU NEED TO REPEAT THE CALIBRATION MAKE SURE YOU DON’T MOVE THE
BIKE DURING CALIBRATION , THAT YOU TURN THE BIKE 180 DEGREES BEFORE
TAKING THE SECOND READING, THAT YOU PUT THE TIRES PRECISELY IN THE
SAME SPOTS ON THE PAVEMENT, AND THAT YOUR HANDLEBAR MOUNT CAN’T
WIGGLE ON YOUR HANDLEBARS.

===========================================================================

I found that my tilt calibration was off, which meant that I needed to redo the coast-downs and 4 mile O&B ride.
Before doing this though, I went back and re-positioned my mount and ensured that everything was securely attached (no room for movement).

As far as other possibilities.....
It is always recommended to do the following prior to each ride:
Provide some time for the iBike to acclimate to the outside weather (10-15 minutes - so have it out with you as you are preparing to ride)
Do a Wind Offset Adjustment (in a place out of the wind)

The other thing I noticed while riding (after correcting my tilt issue and re-doing my profile) is that the Power numbers for several downhill sections reported high wattage (which I knew was not correct). These were evident in looking at the ride data later on in the iBike software. However, after correcting for wind and power they came down to a more expected value (so the software was able to adjust for this particular anomoly) - but I imagine that this will only work if you have a good profile and a good pre-ride routine.

If you want to share/send your ride file and profile, I would be willing to take a look (no guarantees that I will see anything enlightening - but others on the forum may).

HTH.
-David

[Ccummings]

i have had this thing for a week and i'm ready to toss it, i have done at least 20 coast's, i think once it seemed correct, i have done everything over and over, and just when i think the watts are good the next day its off. as far as the data for the coast calibration whats the most important?

[Lorraine]

I have also had some bizarre power readings. In fact, since upgrading to the iAero I am not happy with the readings, and can't get any decent profiles on my TT bike or normal road bike. In fact last Wednesday afternoon while going downhill I had very high power readings and all I was doing was pedalling to keep my legs turning over. Same thing happened yesterday, was coming back from a ride, stopped to change the front tube after a puncture and suddenly I had power readings of over 800w!!when I started again. Once over the traffic lights I stopped, and found the offset was way off and giving readings of 88 plus. I reset this, the tilt had not changed when I checked it! The bike was standing up while I changed the tube and I have sufficient clearance when leaning over it to lock the wheel , so didn't lean on it inadvertently. The iAero is wireless on the road bike, the mount is rock solid. Even after resetting the offset, I know the power readings were not correct, as they should have been in the vicinity of 160-190, the wind was all of about 4-11km/hr as it was very calm all day. I also lost cadence/power for no apparent reason on the return leg for a couple of minutes, yet I went out and back the same way!

Today I was on my TT bike which is a Giant and the iAero (wired version) is attached below the bars well away from my hands and any obstructions. The bike was sitting on the verandah for an hour before I warmed up on the rollers, and the offset and tilt were perfect before I got on for 35mins or so before a 20km TT. I checked the offset and tilt on the start line and both were still perfect. At about 14.5kms into the TT I have totally lost speed/power and it has not returned until about 18.5kms. However, prior to this, from 6kms the power readings are way off, they have progressively climbed, peaking at 1276! The 20.5km TT was held on a circuit of 10.3km, and the only time my cadence dropped was on the corners as the road was damp from earlier showers, and one half way through my first lap, but not enough for me to get wet. The weather bureau put the wind at 4km/h and 15 degrees when I was riding, so very calm. I had no minute man, went off towards the end of the field, passed one person soon after I started before I got to the first corner and then one other after the second corner and after that didn't see a soul, so can't blame any other riders

All of my calibration rides in the last fortnight have been done when its been reasonably calm, (4-8km/hr) but a couple mornings it was exceptionally cold, but the iAero had more than enough time to 'acclimatise'. Some have been done in the afternoon, and in fact one was done in almost identical conditions to one I did on the iBike, but the readings are way off on the iAero?

Can't work out how to attach the files??
I have run out of ideas?
Lorraine

[Velocomp]

Lorraine:

Please contact me at jhamann@velocomp.com and include a ride file with your email. Your iAero should not be having any of the problems you're reporting.

If you don't know how to attach files call me at 406 682 4988 and I will help you.

[Lorraine]

Back up and running thanks to John's help and very impressed with level of service.

Appears my iAero crept out of the factory without some important calibration being completed.

Lorraine