Airspeed sensor for openXsensor

[Tempo]

Let´s concentrate information and discussion to this topic airspeed sensor for openXsensor.

You will find previous debate there
viewtopic.php?f=48&t=5632
and there
viewtopic.php?f=86&t=5920
and elsewhere ...

[Tempo]

@mstrens

you develop openXsensor enhancement for differential pressure sensor MS4525DO, is that right ?

I integrated TEK-probe in thermal flying machines and it is very easy to add simple Pitot-tube (without static source).

TEK-probe gives pressure = p - q
Pitot-tube gives pressure = p + q
You get dynamic pressure = q = {(p - q) - (p + q)} / (-2) if you connect TEK- and Pitot-tube to differential pressure sensor
You get static pressure = p = {(p - q) + (p + q)} / 2

Separate static source is redundant.
openXsensor-Variometer/Altimeter is connected with T-connector to TEK-probe.

What do you think ?

(30.07.2014: correction of + in formula for static pressure)

[mstrens]

FYI, a version of OXS that allows to measure the airspeed is currently in test.
It uses a sensor 4525DO-DS5AI001DP that has to be connected on the I2C bus (just like - in parallel - to a baro sensor MS5611).

If somebody wants to collaborate to the tests, just ask for and I will give the code.
This code is currently still changing every day because I try with Carbo to use the airspeed in order to make an electronic compensated vario and this requires some set up.

I hope that a version will be put on google quite soon.

[Tempo]

precise distinction:

(A) electronic compensated vario
Needs airspeed information and uncompensated variometer information, therefore MS4525DO and MS5611.
You need Pitot-tube anyway. And correct static pressure for airspeed and vario.
Static pressure is source of error, because it is very difficult to get correct static pressure !

(B) TEK compensated vario
You only need TEK-probe, therefor only MS5611 with tube connector.

(C) airspeed sensor
You need Pitot-tube anyway and correct static pressure for airspeed. Therefor MS4525DO (alternatively 2x MS5611)
Static pressure is source of error, because it is very difficult to get correct static pressure !

(D) TEK compensated vario and airspeed sensor
You only need TEK-probe (easy to make Nicksdüse) and easy to make Pitot-tube. No extra static source is needed.
(See suggestion above !)
Therefor MS4525DO and MS5611.

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

I prefer version (D).

[mstrens]

@mstrens

you develop openXsensor enhancement for differential pressure sensor MS4525DO, is that right ?

I integrated TEK-probe in thermal flying machines and it is very easy to add simple Pitot-tube (without static source).

TEK-probe gives pressure = p - q
Pitot-tube gives pressure = p + q
You get dynamic pressure = q = {(p - q) - (p + q)} / (-2) if you connect TEK- and Pitot-tube to differential pressure sensor
You get static pressure = p = {(p - q) - (p + q)} / 2

Separate static source is redundant.
openXsensor-Variometer/Altimeter is connected with T-connector to TEK-probe.

What do you think ?

This is one option but currently Carbo is making test with an OXS having 2 x MS5611 and 1 x 4525DO.
One MS5611 measures the static pressure in the clasic way (no probe).
The second measures p-q with a TEK probe.
The 4525DO is connected to a Prandtl and is used to measure airspeed.
Based on airspeed, OXS calculates the compensation to add (substract) to the vario signal from the first MS5611 (static pressure).
This process allows to compare the results with the measurements from the second MS5611 (TEK probe).

This is just for testing.
Later on, this allows many options:
- OXS measures only the airspeed with a 4525DO + Prandtl probe (or perhaps a Pitot if static in the fuse is good enough) when vario/altitude is not required
- OXS works only as vario (without 1 MS5611 and no 4525DO) with or witout TEK probe
- OXS works with one MS5611 (no probe) and one 4525DO (prandtl probe) ; you have altitude, airspeed, uncompensated vario + electronic comppensated vario based on airspeed (if tests are successful)
- OXS works with 2 MS5611 (1 without probe and 1 with TEK); you have altitude, uncompensated vario + TEK compensated vario + airspeed (based on difference between the 2 MS5611)
- OXS works with the 2 MS5611 (one connected to TEK) and 4525D0 (connected to Prandtl - or pehaps Pitot probe); you can select between uncompesated + 2 compensated vario and 2 airspeeds (this is probably too much)

What you propose is eventually an option more (with 1 MS5611 and one 4525DO, that would probably gives the best results:
User get:
- compensated vario from TEK (I presume it will be the most accurate)
- air speed from Prandtl (I presume it will be more accurate than just a pitot and electronic calclation)
- altitude calculated based on TEK and airspeed.

Please note that having several source for one information when using the 3 sensors, allows in therory , to apply some Kalman filter in order to increase accuracy but this is another debate.

[mstrens]

precise distinction:

(A) electronic compensated vario
Needs airspeed information and uncompensated variometer information, therefore MS4525DO and MS5611.
You need Pitot-tube anyway. And correct static pressure for airspeed and vario.
Static pressure is source of error, because it is very difficult to get correct static pressure !

(B) TEK compensated vario
You only need TEK-probe, therefor only MS5611 with tube connector.

(C) airspeed sensor
You need Pitot-tube anyway and correct static pressure for airspeed. Therefor MS4525DO (alternatively 2x MS5611)
Static pressure is source of error, because it is very difficult to get correct static pressure !

(D) TEK compensated vario and airspeed sensor
You only need TEK-probe (easy to make Nicksdüse) and easy to make Pitot-tube. No extra static source is needed.
(See suggestion above !)
Therefor MS4525DO and MS5611.

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

I prefer version (D).

Do you think that a Pitot + TEK will gives better results than Prandtl probe for airspeed?
I presume that a Prandlt probe is also easy to build (there is one at 10$ at Eagletree)

[Tempo]

Yes, you get it. Let´s continue good job ! It´s an interesting experiment.

My prefered solution is without extra static source. You can get also errors with bad Prandtl-probes.
Generating only values with sensor and microcontroller is easy. Making real measurements is pretentiously.The same holds true for probes.