Personally I would measure air speed with a hot wire anemometer. You may be having problems with the pressure sensor due to the orientation of the diaphragm. If I had a choice I would orient the pressure transducer so that the diaphragm is on it's side that way you take gravity out of the system. You are probably measuring ~1" H2O and that's difficult to do effectively even in a lab environment.

Air density is really dependent on temperature so that is the most important measurement, next in the accuracy budget is barometric pressure then humidity.

After you have your current conditions you probably want to convert to standard units to compare one ride to another on days where conditions differ greatly. Most standard conditions are 1 atm, 68°F and 0% Humidity. I normally do that conversion with volume measurements and not linear speed so I would have to check my notes on the math behind it.

I've been doing airflow measurment for about 15 years so I have a lot of lab experience with it, not as much in an outdoor environment.

Thanks. The pressure sensor is now on its side. From he few runs I've done with it, it's better than expected for a cheap ratio metric sensor. I did spend a good bit of time on regulation and signal conditioning which helped a lot. I'm also trying to be careful with thermal design.

The stuff I'm using is used a lot in model aircraft so slightly higher speed. If this is 'useable' I'll look at better sensors. I'm not sure a hot wire would fit in the budget of a hobby project, but yes that would be ideal.

---

Developing aero, fit and other fun stuff at Red is Faster

I've seen inexpensive pressure transducers from Digi Key or Jameco that work pretty well. I would check zero on it often.

Biggest thing between those sensors and a high dollar one is temperature drift and durability so as long as those conditions are controlled you should be fairly consistent.

jaretj

Just to give you a ballpark idea of the pressure gradients you're working with (pic found on TTBikeFit, not my own work):



I imagine that with a less-aero position (the above looks pretty good aside from the head), different hand/forearm orientation, different helmet selection and head position the results would vary wildly from what you see above.

Consistency/repeatability is the most important part of any measurement. Accuracy is just a matter of calibration :) That said, I think you'd have a hard time getting repeatable measurements if your delta-p sensors are inside an area with a steep pressure gradient caused by the rider himself. You might be able to take care of that with some data smoothing but you're probably just better off putting the whole unit as far in front of the bike as you can without sacrificing stability.

Just out of curiosity, why did you choose a pitot tube and not a kiel probe? A kiel probe would have probably even been cheaper...

---

__________________________

I tweet!

Cheers.

I'll put up some pics on the blog of where the probe is going. It's ahead and above the front wheel and below the hands. Approx 650mm ahead of the steerer axis. According to that graphic, not a bad location. I'm also offset to one side by around 100mm.

The mini pitot tube I have is less than £10, or about $10. I understood that for small angles a pitot tube is not too affected, but I ought to check.

---

Developing aero, fit and other fun stuff at Red is Faster

Ah, that's not too bad as far as pitot tubes go. My experience has mostly been with lab-grade hardware--Dwyer pitots are in the $300+ range, United Sensor Kiel probes were under $80, so if you got a good Pitot for $10 then you're doing alright. The accuracy of the delta-p sensor is almost more important.

Pitot tube yaw tolerance is highly dependent on the tip shape. A blunt tip will "tolerate" up to about 15 degrees of misalignment, while an ogive tip will tolerate less (maybe 8-12), and kiel probes are good for ~40 deg.

Definitely an interesting project. Keep up the good work!

---

__________________________

I tweet!

It is tiny (4mm Diam.). The tip is flat, but at that size I'm not sure how much effect it'll have.

---

Developing aero, fit and other fun stuff at Red is Faster

While I'm thinking about it, does anyone have a good formula for air density (rho) based on temperature, air pressure, and relative humidity, please? I'm working through the various formulae for air density but I've not found one with these 3 inputs straight off.

---

Developing aero, fit and other fun stuff at Red is Faster

You can pull them out of the attached spreadsheet. The equations in the spreadsheet correct pressure according to altitude which you may not have to do depending on whether or not your instrumentation standardizes the pressure.

Thanks. I'm having a bit of trouble with throat link though (404).

---

Developing aero, fit and other fun stuff at Red is Faster

I can't really access my dropbox at work, so unable to do much. IM me your e-mail and I'll send it to you.

Actually wikipedia has a good formula for humid air so there is no sense in me repeating it

Look almost half way down the page

http://en.wikipedia.org/wiki/Density_of_air

Remember temps and pressures are in absolute so use R and K for temps and psia and KPa abs for pressure where appropriate.

jaretj

I do apologise for being thick here, but I'm struggling to get from RH to partial pressures required for the formulae above. I'm not normally bad with this sort of thing, but I keep going round in circles with this stuff getting the right values to feed in.

---

Developing aero, fit and other fun stuff at Red is Faster

1. Calc Psat (Psat = Es) and multiply by 100 to convert from hPa to Pa


2. Then Calc Pv (RH% x Psat)


3. Calc P, where Po is weather station pressure (if standardized for sea level). This is where the altitude and temperature adjustments come in. If P is not standardized (likely the case with direct measurement on the spot - i.e. your device) then you already have P at altitude and temperature and you can skip this step.


4. Calc Pd


5. Plug and chug
[13]

Thank you so much. There's a reason I don't get paid to do this stuff!

I think it was step 2 I was missing (somehow, don't ask).

---

Developing aero, fit and other fun stuff at Red is Faster

Someone will probably come in and tell me I'm wrong, but for now that's my version of the truth.

If you want to do a straight up calculation in Excel I've used the formulas from this site

http://www.brisbanehotairballooning.com.au/calculate-air-density/

or you can use their calculator

jaretj

Jaret's post reminded me to update step 1 to remain consistent with the spreadsheet I sent you

Now I feel very silly. I've been on the the ballooning site a number of times, but didn't think to view source and extract the functions from script.

Thanks for the help guys.

---

Developing aero, fit and other fun stuff at Red is Faster

That hPa thing threw me for a while back when I found the site.

Make sure you have the units correct.

jaretj

Check your messages

I sent a google doc link

I'm sure you have but I'll ask anyway... have you seen Robby Ketchell's BAT box?

I hadn't seen that article. I can see what they were trying to achieve. The problem they would have had will have been that position changes will have had a major effect on that sensor. They are using the same sensor as I am, but it doesn't look like they did much signal conditioning.

I now have 2 formulae to compare which I'll do as soon as I get a minute. I recoded the the JavaScript function into C for the arduino and posted it on the blog. It seems to work, but again it'll need validating.

---

Developing aero, fit and other fun stuff at Red is Faster

I did have a look at this one with a slight smile
http://flocycling.blogspot.co.uk/...part-5-wind.html?m=1

---

Developing aero, fit and other fun stuff at Red is Faster