Problem with Power Readings

[pahayden]

Hi.

I've been using my Dash + Power since earlier this year, and haven't had any issues with it until now. I started using it on my road bike, then switched it to my TT bike and had no problems with it at all, but since switching back to using it on my road bike I can't get reliable power output readings anymore. The unit and all sensors are set up properly, I've set the road bike profile up with tilt calibration and done a calibration ride, etc. but when out on a ride I will see the power output readings increase to levels that I know cannot be correct given the amount of effort I'm using when pedalling. I've done all sorts of things to try and fix this (using the handlebar mount instead of steerer tube mount, removed and then re-installed the iBike software on my iPhone, set up a new bike profile, used the "Check Calibration" feature in the new iBike software), but nothing has corrected the problem. What I've noticed now is that the power reading problems happen when I'm cycling on a road surface that is not very good and the vibrations coming up through my bike seem to be making the power output readings just climb and climb (they get up to around 800-900) and a lot of times this is happening when going down hill. As I say this is a new problem for me and it didn't happen on the same bike when I used it before or on my TT bike.

Does anyone have any advice for me about what I can do to fix this? I have tried using some foam under the Phone Booth in-between it and the stem with no success. At the moment this problem has really put a dent in my training because I can't trust the power numbers or use the ride files since having this problem.

I switched to using an iPhone 4S, but was having the problem with an iPhone 4.

PS. Anyone interested in the battery performance of the iPhone 4S... I've found it to be worse than my iPhone 4 and am charging the 4S more often.

Thanks

[Velocomp]

iOS 5 has caused some difficulties for the + Power units.

We have had to make changes both to our iBike software (release 6.0.2 is just out) and also to the iBike app (version 1.0.10 pending in the app store).

The update 6.0.2 to the iBike software corrects problems with the Check Calibration feature

We expect the new iBike app will be released very soon. Hopefully, these two updates will correct your problem.

[pahayden]

Thanks for the reply, but I was having this problem with a previous version of the iBike software (before the Check Calibration feature) and an on iPhone 4 before iOS 5.

[Velocomp]

Please post a ride file.

[pahayden]

Hi.

Here is a ride file. Hope you can see something significant in it.

Thanks

[Velocomp]

I looked at your ride file. You are seeing high wattages at very high road speeds; if the roads are extremely bumpy in the regions of high wattage then it may be that your case is vibrating too much.

Please post a side view photo showing how your case is attached to the bike.

[pahayden]

Here is the side view photo of how the case is attached to the bike. The odd thing is I didn't have this problem when it was on this bike in the same way in the past.

[Velocomp]

Please try moving the rectangular bracket of the steer tube mount to the rear position, closest to the stem bolt. Remove the two phillips head screws underneath the mount to do this.

This will make the system as stiff as possible and will improve vibration performance.

After doing this please try a ride and post the ride file.

[pahayden]

I've made the change and did a ride this morning. The attached ride file is from this morning. The power spikes are still there when going fast down hill and I wasn't using anywhere near the force to get readings like that.

[Velocomp]

We will look at this; in the mean time, please post a ride file from the same bike and an earlier ride, where you did not see this problem.

[pahayden]

Here is a ride file on the same bike, with all the sensors and mount, etc. setup as it is now.

[pahayden]

Any progress or solution to this problem yet?

[Velocomp]

Your problem is quite straightforward.

Between your 8/17 ride and your 10/27 ride, your profile changed.

Specifically, your Crr (rolling resistance) was correct on 8/17 (0.0055) but on 10/27 was a very high and incorrect value of 0.0138.

With your Crr being incorrectly high, your watts would be unable to balance on downhills, and your overall watt readings would be way too high.

To fix this problem, go into Options/Advanced Options/Check iDash Settings. Scroll down to Crr, and set it to the value 0.0055.

This should fix things nicely.

You can apply this same Crr fix to your recent ride files in the iBike software. Go to Tools/Tweak Crr. Set the value to 0.0055 and your watts should return to correct values.

Also note that your wind scaling changed between 8/17 and 10/27. Your 10/27 calibration looks better; the low wind scaling on your 8/17 ride would cause somewhat lower watts.

[pahayden]

Thanks, I will change the setting and see how it affects my next ride. Could the difference in Crr be because I'm now using a 25mm tyre instead of the 23mm tyre I was using before?

[Velocomp]

No. The Crr reading you are using would require that your bike be traveling through soft sand...

[pahayden]

ok, that's good to know. I haven't altered any of the settings for it to be like that, all I did was the tilt calibration and then the calibration ride do I don't know why it's like that. I will post a ride file with the new Crr setting when I've done a ride.

Thanks

[pahayden]

I did a ride to test the new Crr setting you gave me, but I still got a big power spike when going down hill at speed. I've attached the ride file.

[Velocomp]

Your watts look much more reasonable.

When pedaling downhill at high speed (33 mph) it is absolutely necessary for your iBike to have precise calibration. This requirement is doubly important when, as is the case for you, you are riding downhill during the first 5 minutes of your ride. For maximum calibration to occur, the iBike should be on the road for about 7 minutes.

Note that your power will be zero whenever cadence is zero. If you were soft pedaling, then any calibration issues would be displayed.

If you wish you can try doing a four mile long calibration ride. Go through the bike setup process, then do the cal ride afterwards. You'll want to double check to be sure the Crr is correct.

Or, you can just ride as is. From your file it looks like it took you 10 minutes to work up enough sweat to get your HR strap working. The effect of the watts spike difference on your overall ride stats is extremely small.

[pahayden]

How do I do a longer calibration ride? When I've done them in the past it has only done the ride for 1 mile and then ride back the same route 1 mile.

[Velocomp]

Just ignore the message and keep riding for a longer distance.

[pahayden]

I did a longer calibration ride yesterday as you suggested, and went out for a training ride this morning of my usual distance, but the problem isn't fixed. I'm giving up with the power meter now because apart from the big power spikes I get the wattage readings fluctuate beyond reason even when I'm using the same amount of force on a flat road. I've attached the ride file for your reference.

Thanks for all of your help in trying to solve the problem.

[Velocomp]

The poster and I have had a private email exchange over the past few days. I thought it might be helpful to share my overall analysis here.

Begin response:

I've looked at your ride files carefully. Your iBike is calibrated properly and 99% of the information is spot-on.

There are three or four places with watts spikes that are probably not right.

These spikes could be caused by minor variations in your riding position--something that cannot be "calibrated out". Or, it might be caused by a slightly high value of CdA. Or, it might be caused by a too-high wind scaling factor. Or, it might be caused by imperfect readings from the sensors. Or, it might be a combination of all these things.

Opposing wind forces increase as the cube of the speed. What this means is that, in doubling bike speed from 12 mph to 24 mph, wind forces increase by 8 times! So, any calibration or measurement errors go up exponentially at high speed, meaning that at high speeds (generally on downhills above 25 mph), watts spikes can occur.

To assess the significance of these spikes, I have taken your ride file and, with the iBike software command "File/Delete Selection"...surgically removed four watts spikes in the 43 mile ride file you sent.

Before spikes removal:

Ride stats before spikes removal

Before editing.png (29.25 KiB) Viewed 11805 times

You'll see in the "after spikes removal" stats that total ride time is less (due to removal of the spikes), and that other stats are affected comparably.

After 4 watts spikes removal

After editing.png (30.18 KiB) Viewed 11805 times

Your actual total ride (the "before" box) was two hours, 29 minutes, 53 seconds long. In this ride of 8,993 seconds, you had 4 areas of watts spikes. These spikes were removed and the "after" box shows a total ride time of 2:28:33. This is an 80 second difference, less than 1% of your total ride time.

Your average watts, the best indicator of your overall work output, is different by 1.5% between before and after. Again, this is within design spec.

One practical way to understand the significance of the watts spikes is in total calories: these spikes added about 45 calories to your "correct" total effort of 1885 calories. 45 calories is about the amount of energy contained in 1/5 of a candy bar...

There is no power measurement device in the world, at any price, that has 1% measurement accuracy. Competitive products (all of which are much more expensive than ours) have measurement errors comparable to ours, but those errors manifest themselves in a form that is insidious. Their primary issues are temperature drift and offset errors, either of which can cause variations in watts measurement. However, temperature drift is not visible in the ride file! So, with other devices the ride data can be "off" but for a far longer period of time, and can drift back and forth throughout the ride, invisibly.

And in a broader sense, what you are noticing is characteristic of all consumer electronics devices.

Consider a beautiful, high-definition TV image. You might notice that, when that image is changing quickly, instead of the crystal-clear picture that you see 99% of the time, the fast-moving images are blurry and even blocky. This is a design compromise that the TV people make so that high definition TV can be affordable.

Digital music is the same way. It can sometimes sound awful, due to the compression algorithms used to keep the file sizes from becoming prohibitive.

What you're noticing in the spikes is the equivalent of occasional blurry images in your TV, or sometimes scratchy-sounding music.

What can you do about the spikes?

A very simple way to prevent spikes from occurring is, during the ride, do not pedal on downhills. Watts will be forced to zero.

If zero cadence is contrary to your riding style, then after the ride you can edit out the spikes with the iBike software. It is easy and fast.

However, this analysis demonstrates something we've learned over the past 7 years: while spikes can be annoying when focused upon, they are almost always meaningless in terms of their significance.

So, the easiest thing of all to do is to ignore them...