No problem. I do my best to present what I do fairly, Marc and other folks in this topic are still the experts. But I'm getting better at it. Learning more the more I use the tools.

Quick trick :

You can ignore this part, it's just for understanding and you probably already know

aero watts = 0.5 * airDensity * CDA *airSpeed^2 * airSpeed.
Assume no wind so it becomes 0.5 * AirDensity * CDA * airSpeed ^3

So at 36km/h (10m/s) 0.005 difference means 0.5 * 1.16 * 0.005 * 10^3, so

0.005 * 580 = 2.9watts

At 46.8km/h (13m/s) 0.005 means 0.5 * 1.16 * 0.005 * 13^3 so

0.005 * 1274 = 6.37watts


So the part to remember

At 36km/h about .6 watts per .001
At 48km/h 1.3 watts per 0.001

Indeed big differences for the helmets alone, or did the head position alter with the helmets?

 

The report you did in the other thread is great. I am so fed up of winter and I want to get testing.

It would be really cool if we could come up with some kind of template for recording/reporting results. I may propose something when I get out there but ideas would be cool.

I was looking at the pic and spent 30 seconds looking for your aerometer :-). Then it hit me. I need to do some testing about it's location. I am very curious on the impact of yaw in that position. I saw things in the tunnel but never with the device at that "extreme". I think there is value especially when doing cockpit changes but at high yaw....not so sure. Only one way to find out

What order of magnitude of factor are you getting ? In the 1.4 or in the 1.2 range ?

Depends on the wind that day and what clothes I wore. Like skinsuit or flappy winter jersey.

I've seen from 1.38 to 1.295. But never anything outside of that. If talking statistics talk, the "most" would certainly be 1.38. I see that factor a LOT.

I put the sensor way over there after seeing photos of it on some of the UK TT forum's favorite pro and amateur riders out testing in a velodrome. It hit me "why move the sensor with the cockpit all the time?". I just felt like I didn't want the sensor near that area and putting it out there was in better air. I may have it wrong, but it seems to work well to me.

I keep forgetting to put an old scale in my car to record weight at run time. I did weigh myself plus bike plus kit I usually am using on tests though, and use that in the GC Notio field.

At 0 yaw I would have no concern.
At yaw, I am almost certain it will be off.
However, on a "out and back" the "offs" may average themselves out.

As for the 1.38 to 1.295 that is a big big delta.
From one run to another not surprising, so you need to zero in on a value and stick to it.
I have seen people using a different value for each test which makes their data completely meaningless.

I went back, those factors were from long ago. They're lower end now and always very close.

Next up, today did a try on the Garmin positional laps. I'm an idiot not doing this before. Start the lap position once the live display has "evened out" a good bit, then hit the lap position again about a minute from your turn around. This gave me pretty nice laps.

Also, need to get home and view in GC, but looks like from lap data I might be under .200 now. Without a fresh leg shave and without the skinsuit on. And with the .006 slower tt helmet. That would be nice. But, "I'm from Missouri, show me". I'll have to see it in GC first.

Also, despite it still hurting. 300w today for a few runs felt nicer than it used to. There is hope for 2022.

The lower, the better. But consistency of numbers trumps all


Try Strava segments. They appear in automagically in GC.

Wuuut. Even better. Thanks! Do they need to be public Strava segments, or can they be private?

I have private ones setup in this area I rode today, but haven't ever seen them popup in GC.

You can thank DCRainmaker for that. It was his idea.

The process for me is to "star" the segments in Strava and they get automatically uploaded to the Garmin. Then when riding the Garmin gives you "segment in 250 meters......go". They are logged in the FIT file and you can get to them in GC.

What Garmin are you using ? I use this on my 935 watch. Not sure my 130 has it.

BTW, my testing of Garmin positional laps was that they were quite imprecise and would degrade as the ride went on, but maybe they fixed that

Using a 530. On the ride today, it was well close enough each lap location to be worthy. I think the stars will work from Strava to the Garmin. I'll go in and star the segments of interest and try again.

So confirmed in GC, without skinsuit was seeing 0.197 and .0036 combo. The road I'm on isn't super smooth but using TT tires.

Also have a "sin" of too big a tire on the rear on the HED disc for training comfort.

Sweet. Progress.

Nice. What does it look like in "classic" Aerolab without the wind component ?

This is more related to aero testing, so I'll bring the discussion over here

When analyzing data from a fellow STer in South Africa, I noticed something that can impact us when aero testing. Curious what you guys think

He rides with a set of Vector 3 pedals and a powertap. The vectors are logged by the standard Garmin mechanism, the Powertap is logged by a ConnectIQ datafield so he has both data values in the fit file.

I plotted 3s average of PT over 3s average of Vectors, expecting something like a 0.97 due to drive train innefficiency.

What I see is the calculated ETA is very close to 1.0 and stays pretty consistent. However on a "coast", it will drop down to 0.92 and stay there for maybe 2 or 3 laps, then pop back to a 1.0 for a few laps....

My guess is this is the auto-zero kicking in when he coasts (during turn arounds) and throwing one PM off.

This can really screw up testing.

Anyone else experience this ?

I've had three PTs during the years. The first one I never did (much) aero testing with. The two others are a track G3 and a road G3. The track one seems to be fine when track testing. Quite stable, no funny values. The road one seemed to jump quite a bit around, like +/-7W or so (so two equal setups tested minutes apart could show a 14W difference) - likely due to auto zero shenanigans - and I gave up on using that for testing. FWIW, YMMV, etc.

I put some logging of auto zero calibrations in my aerogizmo. Maybe I can identify things.

Maybe P2Max auto calibrations are flawless but I am going to track it for a while.

My PTs, the older ones and the G3, often made troubles with their auto zero. In some out and back TTs it turned out crucial when a new zero was set during coasting to the turning point. The problem was, that a small change in the zero value resulted in a rather big change for the power measurement.

For aero testing with a PT I think it is important to always do manual zeros until the zero is the same as in the run before. If the conditions or the temperature respectively do not change much between the runs there is no reason why the zero should change.

I think what is often not considered is that the zero calibration of a power meter itself has a measurement error which adds to the error of the power meter.

I don't own one, so don't know, but could you all with the PT intentionally pedal through the turn around? Even if the power is like just 20w or something? Would that keep it from doing that?

Either way I turned off this junk in the SRAM app for the Quarq.

Zero stuff is just weird to me anyway. If you haven't uninstalled and reinstalled hardware touching the measuring device and retorqued that, why do we manual zero anyway? The meters all supposedly temperature compensate, built in. So no need to do it for that reason either.

So, if this is so..........why are we all zeroing these for every ride? Aero testing or not?

In the facility I work in, the only time we "zero" an instrument in a similar way is with optical density because it's a comparison in the process from what the process water is versus the product stream. If the process water changes and you measurement is product concentration, you need to know the delta to get the true concentration.

I don't feel like that should be a thing for a mechanical strain gage. If it's the same reason, knowing the delta from the stationary strain to the applied strain........if you haven't touched the thing with a tool or crashed........why?

 

What's interesting is most of these device auto-zero by default. You can turn it off on some, probably all, but I suspect most leave it on.

In the aero test I was analyzing it would definitely throw numbers off significantly

My 1st gen P2M you can't turn off auto zero. Don't know about newer ones. Shimano PM you can't turn it off either - and worse yet, you can't even see what the zero value is. I wouldn't be surprised that there's a lot of currrent PM's where you can't turn off auto zero. And likely also many where you can't see the zero value. IIRC you can't see the zero value on Garmin Vector either?

I use a Track P2M Type S with a 144BCD Narrow wide chainring.
No auto zero.

Thanks Marc, worth noting this is a pretty old powertap model. After I thought about it, seem to recall Tom A an Rchung suggesting to slowly pedal and not coast in their protocols (I think in the platypus thread) although not sure if that was down to auto calibration or something else that messed up data.

It's been an interesting investigation. I suspect it's more common than we think.

I found a really helpful engineer at one of the companies of a brand I own (I own 4brands). They told me logging this would be useful because they see cases when power drops for unexpected periods of time and come back.

it's funny, Garmin tells you when they update your speed sensor wheel circumference based on GPS, but they don't tell you when your powermeter changed it's zero offset.

Aero testing is very particular. If you lose 4watts for 5 minutes on a 2hour ride, who cares. If you lose 4 watts on the 3rd B of ABBAABB and it comes back for the 4th B, you scratch your head.

I've actually wondered with aero testing like this if a tool of the trade someday instead of using a power meter connected to legs and cranks or pedals and hubs...........it moves to a rear hub based e-bike assist motor.

You instead measure amps/milliamps and command an output. You couldn't aero test the rear wheel itself in that situation.........but everything else would be the same.

It would eliminate rider fatigue. It would eliminate zero cal issues on strain gage power meters. Etc....

Not sure the road going classification of a throttle based bicycle legally speaking if it's under X power and XX speed.

But it would effectively totally eliminate a HUGE variable set.

There is a track federation that looked at this for a simple reason, to be able to test at race speeds (~70km/h) for longer periods of time (to get stable results) and in a way the cyclist could repeat multiple times in a session without being wiped out.