Open source speed sensor?

Need some advice? Trying out a new idea? Fancy a beer?
Join us for some general banter and good times.
User avatar
thomas9x
Posts: 230
Joined: Wed Feb 29, 2012 3:12 am
Country: -
Location: USA

Re: Open source speed sensor?

Post by thomas9x » Tue Jul 01, 2014 8:09 pm

The advantage of re-purposing a ready-to-use sensor like EagleTree's Airspeed V3 module is that it it will have the necessary low noise construction to ensure accurate/stable results. Plus integration would be as easy as soldering four wires and some very simple firmware. Keep in mind that EagleTree will set it to the third party mode if you order directly from their site and request it.

But I've got another idea for this project. A couple weeks ago EagleTree formally released their new Vector Flight controller. One of the optional modules is a $39 airspeed sensor with a Pitot tube.
Imagehttp://www.eagletreesystems.com/index.p ... uct_id=145

But this solution will require a bit of reverse engineering in order to write the I/O code. I'll start the ball rolling and publish what I know. :)

Sensor:
The Vector's airspeed module is a "dumb" sensor (all processing is on the host and access is via i2c). It uses the inexpensive MP3V5004DP dual port pressure sensor.
Link to sensor data sheet

Glue Parts:
The PCB has a 3.3V LDO Vreg and a I2C ADC. There's very little else to it -- just some EMI filters and Vreg stabilization caps. Here's a photo of everything (pressure sensor is on the other side, so it is not seen):
vec-pitot.jpg
Vector Air Speed Sensor
The SOT-23-5 that is marked L00A is the Vreg. My best guess is that it is a LP2980A 3.3V, but it could be something else. We really don't need to know the exact part number, so there's no reason to spend any time researching it.
http://www.mouser.com/ds/2/405/snos733m-317056.pdf

The other SOT-23-5 (marked S14W) is the i2c ADC. My gut feeling is that it is a 12-bit, but it could be 10-bit. For example, maybe it is something similar to a AD7476?
http://www.analog.com/en/analog-to-digi ... oduct.html

The Goal:
To use the Vector's airspeed sensor the only missing link is to identify the ADC. Its data sheet will have the details to the i2c interface. Beyond that, its only code. To infinity... and beyond!

- Thomas


User avatar
thomas9x
Posts: 230
Joined: Wed Feb 29, 2012 3:12 am
Country: -
Location: USA

Re: Open source speed sensor?

Post by thomas9x » Tue Jul 01, 2014 9:10 pm

ADC:
I think I have identified the ADC used in the Vector's airspeed sensor. The guys at EagleTree like to use Microchip's parts so my search started at microchip.com. I found they had a 12-bit ADC with i2c that was in a SOT-23-5 package. And to my glee, the data sheet confirmed that its part marking would begin with S1 (the full number is S14W).
http://ww1.microchip.com/downloads/en/D ... 21732D.pdf
http://ww1.microchip.com/downloads/en/A ... 00845a.pdf

So assuming the detective work is reliable, I think the loose ends are all tidy now. Next step is for someone to create a bit of i2c code and the pressure-to-speed conversion calculations. How hard could that be? :)

- Thomas

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Wed Jul 02, 2014 8:03 am

Thomas9x,

Thanks for the information about the Vector's airspeed sensor.
Still I expect that a solution around the 4525DO-DS5AI001DP would be better because:
- the ADC is build in the sensor and so the sensor should have less noise
- the sensor should be more sensitive because ADC is 14 bits (instead of 12 bits)
- temperature compensation should be better because it is digital
_ there should be no issue about an offset
- as extra, you can get the temperature
- the sensor runs at 5 V (the same as arduino pro mini) and there should be no issue for I2C interface. I do not know if the Vector ADC (and I2C) is 3.3 or 5 volt but I expect it is 3.3 volt so there could perhaps be some issue with I2C interface.

The connection with OXS is in all cases similar (4 wires).

As drawback, I only see that the price for the kit from 3DRobotics is a little more expensive than the Vector.
Still if you buy only the sensor from digikey then it become cheaper (or similar if adding only the pitot tube from Eagletree which is available for 10$).

User avatar
thomas9x
Posts: 230
Joined: Wed Feb 29, 2012 3:12 am
Country: -
Location: USA

Re: Open source speed sensor?

Post by thomas9x » Wed Jul 02, 2014 2:13 pm

I do not know if the Vector ADC (and I2C) is 3.3 or 5 volt but I expect it is 3.3 volt so there could perhaps be some issue with I2C interface.
It is a 3.3V i2c interface.

The eLogger V3 airspeed sensor operating in third party mode will be the least amount of effort. The firmware impact is minimal and the hardware is ready to plug in. But in the spirit of DiY, raw parts like the 4525DO-DS5AI001DP are worthy. But in the end, the best sensor choice is the one that the volunteer code writer is willing to support.
If someone provide me a 4525DO-DS5AI001DP sensor, I can try to write the code in order to let OXS measure airspeed and dTE (when associated to a MS5611 vario sensor).
Because you are willing to develop the necessary firmware the 4525DO-DS5AI001DP sensor is the best choice.

- Thomas

User avatar
kaos
Posts: 3247
Joined: Wed Dec 28, 2011 1:15 am
Country: United States

Re: Open source speed sensor?

Post by kaos » Thu Jul 03, 2014 4:19 am

$42 with 3rd party mode shipped sounds better now. ;) IF the code writer is willing. ;)
I figure to buy the raw material to make the pitot tube would add quite a bit to the cost (shipping needs to be figured in) and labor, except for those who already have those material readily available.
I usually think if DIY would reduce the 'total cost' by 40-50% with increased precision/ functionality would be a good candidate for DIY, otherwise the plug and play would be the way to go. ;)


Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Mon Jul 07, 2014 8:53 am

In openxsensor we have routines for measuring altitude with 0,1 m resolution. Why not use this routines with a second ms5611 sensor (two i2c adresses are possible) that is connected to a probe? The speed can be calculated with the delta altitude: V=sqrt(2gh). The energy of 10 m/s correspond to a delta altitude of 5m, 5 m/s to 1,25 m.

The probe can be just a tube in flight direction, i think, a special pitot-tube is not needed, because of the second ms5611. Fixing a pressure-connection to the ms5611 is possible.

In a next step, the measured speed could be used for an energy compensated vario.

@ Michael (mstrens): are you interested in making an openxsensor version for this setup? I would need month for it.

i am already preparing a ms5611 for using a TEK-probe:

TEK-probe

User avatar
jhsa
Posts: 19382
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: Open source speed sensor?

Post by jhsa » Mon Jul 07, 2014 9:12 am

what about the pressure not being always the same? for example we always have to reset the altitude to zero before flight..
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Mon Jul 07, 2014 9:20 am

Both ms5611 have to be resetted before flight - then you have one ms5611 to calculathe the altitude as usual and the second one to measure the dynamic pressure to calculate the speed.

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Mon Jul 07, 2014 9:51 am

Carbo wrote:In openxsensor we have routines for measuring altitude with 0,1 m resolution. Why not use this routines with a second ms5611 sensor (two i2c adresses are possible) that is connected to a probe? The speed can be calculated with the delta altitude: V=sqrt(2gh). The energy of 10 m/s correspond to a delta altitude of 5m, 5 m/s to 1,25 m.

The probe can be just a tube in flight direction, i think, a special pitot-tube is not needed, because of the second ms5611. Fixing a pressure-connection to the ms5611 is possible.

In a next step, the measured speed could be used for an energy compensated vario.

@ Michael (mstrens): are you interested in making an openxsensor version for this setup? I would need month for it.

i am already preparing a ms5611 for using a TEK-probe:

TEK-probe
It would be quite easy to add a second MS5611 to OXS and to calculate an airspeed (and afterwards an dTE).
Still I don't think it would be the best solution because:
- I noticed that the altitude reported by MS5611 drift a lot during a flight. It is "normal" getting a difference of several meters between begin and end of a flight. I presume that this is cause by a bad temperature compensation inside the MS5611.
So, I am afraid that the airspeed calculated by the difference of pressures reported by the 2 MS5611 would become wrong when time is going on.
- you can not connect directly a probe to a MS5611. You have to build a kind of small box to put the MS5611 inside a closed volume connected to the probe.
- I am not sure that the static pressure reported by the first MS5611(the one used for altitude inside the glider) would be the same as the static pressure provided by a pito tube.
- a MS5611 has a range of pressure that is much wider than some sensors (like the the 4525DO which has 1 a range of 1 psi).So it would normally be less accurate for low speed (low difference of pressures).

For those reasons, I believe that a solution build around a sensor 4525DO and a pitot tube would be better to get both airspeed and dTE.

I have no experience by I presume that if you only want having dTE, then the solution of only one MS5611 (=normal OXS software) and a TEK probe would be the easiest solution.

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Mon Jul 07, 2014 10:16 am

Hi Michael - that was fast!

- if the two ms5611 drift in the same direction - the fault will compensate. mounting them together (same temperature) will help
- i "stole" an idea, sticking an preformed heatschrink tube on one side to the sensor, on the other side to the tube. So the small box is obsolete
- your concerns about the static pressure inside the fuselage should be considered, but you have it also when you use a vario, or an altimeter, so think, it can be compensated
- the resolution should be enough, because 5 m/s (corresponds 1,25 m delta-alt)

I will try the TEK-probe but - and here the circle closes - how can i measure my actual altitude? With a second ms5611? Thats how my consideration started :-))

Maybe its worth a try?

Edit: And i have some ms5611 lying around!

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Mon Jul 07, 2014 11:12 am

Carbo wrote:Hi Michael - that was fast!

- if the two ms5611 drift in the same direction - the fault will compensate. mounting them together (same temperature) will help
- i "stole" an idea, sticking an preformed heatschrink tube on one side to the sensor, on the other side to the tube. So the small box is obsolete
- your concerns about the static pressure inside the fuselage should be considered, but you have it also when you use a vario, or an altimeter, so think, it can be compensated
- the resolution should be enough, because 5 m/s (corresponds 1,25 m delta-alt)

I will try the TEK-probe but - and here the circle closes - how can i measure my actual altitude? With a second ms5611? Thats how my consideration started :-))

Maybe its worth a try?

Edit: And i have some ms5611 lying around!
If you use a TEK-probe, the vario will give you dTE but not anymore a correct altitude.
As you said, adding a second MS5611 can be a solution in order to get the altitude.
If you want, I can adapt OXS in order to support 2 MS5611. Each MS5611 would calculate altitude and vertical speed.
It would be possible to setup how the telemetry fields are transmitted.
E.g. the sensor connected to TEK probe could transmit his vertical speed and the other sensor the altitude.
It would also be possible that one sensor transmit his values as altitude and verticalspeed and that the other sensor send his measurements (altitude and vertical speed) as T1 and T2. This would allow to compare the 2 sensors and check if drift goes in the way for both.

Are you interested?
Do you expect more?

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Mon Jul 07, 2014 11:33 am

That would be perfect! So i can try both variants:

1. TEK probe, have dTE, measure 2 altitudes (pressures), calculate speed
2. Pressure probe (pitot), measure 2 altitudes (pressures), calculate speed and dTE

When the weather permits, i can make some tests this week (vacation), logging also gps, to have some comparison values and see if it works.

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Tue Jul 08, 2014 7:38 am

Carbo,

please find in attachment a version of OXS that you can use in order to test 2 MS5611 connected to an Arduino OXS.

To activate the second MS5611, you must leave the line "#define VARIO2" uncommented (in the config.h file).
You can then add lines to define which fields are transmitted. The config file allows now 4 new fields (Altimeter_2, VERTICAL_SPEED_2, ...) that you can use.

This version does not calculate an airspeed and a dTE but it is possible to connect one of the 2 MS5611 to a TEK probe.

Please note that I do not have an OXS with 2 MS5611 and so, I could not test this version.
I just hope it will work. If you get some issues, just let me know and I will try if I can find the reason(s).

Do not forget that you have to connect the CE pin to 3.3 volt on the second MS5611 in order to let him react on a different I2C address.
Notes: I added some more comments in the config file (in the section 4). Here an extract:
// 4.1 Connecting 1 or 2 MS5611 barometric sensor is an optional feature
// Each can calculate the relative altitude (in meter) and the vertical speed (in meter/sec with 2 decimals)
// Uncomment the line "#define VARIO" to enable first MS5611.
// Uncomment the line "#define VARIO2" to enable second MS5611.
// Both MS5611 are connected in parallel (using same arduino pins A4 and A5)
// First MS5611 is the one having the MS5611 CE (chip enable) pin connected to ground. This is the normal set up for e.g. a GY63 module. It react on an I2C addess 0x77.
// Second MS5611 must have the MS5611 CE (chip enable) pin connected to VDD. So it get an I2C addess 0x76.
//
// Note : it is not required to comment the sensitivity, hysteresis and OutputClimbRateMin/Max parameters when a vario is not used (those parameters are just discarded)
// The vario setup parameters (sensitivity, hysteresis, PPM, ...) are shared by both MS5611
// Analog vertical speed is generated (if selected) only by first MS5611.
// EEPROM saves /loads only data from the first MS5611.
// Important : When only one MS5611 is used, it must be connected as first MS5611 (CE connected to ground)
Attachments
openxsensor_2014_07_08_09am.rar
(51.05 KiB) Downloaded 113 times

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Tue Jul 08, 2014 1:36 pm

Hi Michael, thanks a lot for this sketch, it was plug and play!

Here a some pictures of the Vario with two MS5611, i hope they are selfexplaining. The idea with the tube connection is from a german forum:
Anschluss1.jpg
Anschluss2.jpg
Anschluss3.jpg
Anschluss4.jpg
Vario2MSoffen.jpg
Vario2MSfertig.jpg
THERMIK_XL-2014-07-08.csv
Log-File with Alt,VSpd,T1,T2
(32.13 KiB) Downloaded 294 times
Blowing from 40 cm slowly in the direction of the tube gave clearly reactions of T1 and T2. The vario worked as usual. I´m optimistic, that it will work in the plane too. Unfortunately here is rain and wind, so the maiden has to wait.

P.S. What Michael called CE (chip enabled) is labeled CSB on my breakout boards.

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Tue Jul 08, 2014 3:20 pm

Fine that it seems to work from the first try.

Let me know the results of your tests. If OK, I can publish this code on Google.

After you try with a TEK probe, you could make some tests with a normal pitot probe running at some fix speed with a car and logging the data. If the data are reliable we could then try to calculate an airspeed and a dTE using a pitot probe instead of a TEK probe. Still I expect that the precision would be less than with a TEK probe or with a 4525D0 sensor.

User avatar
jhsa
Posts: 19382
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: Open source speed sensor?

Post by jhsa » Tue Jul 08, 2014 3:24 pm

so, now this second sensor will only read the pressure (altitude) to give the speed?

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Tue Jul 08, 2014 4:37 pm

jhsa wrote:so, now this second sensor will only read the pressure (altitude) to give the speed?

João
Currently the sencond sensor performs exactly the same calculations as the first one.
This allows :
- to make tests to compare 2 sensors
- to connect one sensor to a TEK probe, using this sensor to get a (pneumatic) compensated vertical speed; the other sensor can then be used to still measure the altitude.

The current version does not allow to calculate the Airspeed (based on a difference of pressure between a static pressure and a dynamic provided by a pitot probe). So it does not allow currently to calculate a dTE using a pitot probe.

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Tue Jul 08, 2014 4:42 pm

@ mstrens:

To stay on topic :-) i will start with a pitot tube (and speed-probe sounds sexier than TEK). In opentx there is an "airspeed" telemetrie field, can openxsensor adress this field? To calculate the airspeed you subtract Alt (Vario) from Alt (pitot) = delta Alt. Speed ist then sqrt (2 g delta Alt) = sgrt (2 x 9,81 x delta Alt) = sqrt (19,62 x delta Alt). Could you integrate this formula? In my setup is sensor 1 = Vario and sensor 2 = pitot.

Thinking about dTE: If your speed increases then sensor 2 (pitot) will give you negative V-Speed, because pressure increases (= sinking), so when you subtract VSpeed sensor 2 from VSpeed sensor 1 you have an energycompensated Vario? Too easy to be true? I can not yet verify it mathematically, but i would like to test it, again beeing barefaced: could you integrate this formula and send it as dTE? Otherwise it can be sent as VSpeed.

@jhsa:

yes, with the two sensors, one inside the fuselage, the other outside with a tube, you can calculate speed and an energycompensated vario - if it works with the actual setup, needs to be tested. Flying loopings, your vario (nearly) doesn´t change tone. Becoming faster in a thermal without changing altitude, your hear the climb-tone. The vario reports the energydifference of your plane.

Edit: dTE calculation wrong

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Tue Jul 08, 2014 5:39 pm

Before making all those changes to OXS, I would like to check if the results would be acceptable.
So, I would like to perform the calculation in XLS based on the telemetry data logged by Taranis on the SD card.

Can you make some tests (real fly) in neutral air (no lift, no sink) where you increase/decrease the speed playing with the elevator.
During those speed up/down, you should activate the log.

If I get logs with both Altitude and VSpeed, I can then perform some XLS calculation.

Note: I think that, at least, your formula to calculate airspeed is not 100% correct because it must take care of the density of the air

User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Speed sensor for low velocity

Post by Tempo » Tue Jul 08, 2014 6:36 pm

@mstrens
@carbo

My experience:

(1) Using one MS5611 for TEK-variometer and a second MS5611 for altitude:
It´s only interesting for high speed. By low speed-thermal-circling with TEK the difference to an extra altitude sensor without TEK is low.
Changes in weather/atmospheric pressure during flying make more differences in altitude values.
(2) I use another measurement principle for measuring low aerodynamic speed: thermal-flow-sensor.
I use this sensor-family
Sensirion_Differential_Pressure_SDP600series_Datasheet_V1.7.pdf
thermal-flow-sensor
(1.01 MiB) Downloaded 218 times
Speed range goes down to about 15km/h and up to 110km/h. I´m measuring total pressure at the nose of fuselage (on a f3j-machine it´s about 40cm in front of wing) and static pressure inside closed fuselage.
(3) I prefer TEK-probe for compensating variometer. You get good compensation with "Braunschweiger Düse" .
If you use electronic compensation with speed information, you need a good static pressure measurement, which does not depend on air speed. It´s not easy to get it.

Carbo
Posts: 443
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo » Tue Jul 08, 2014 6:44 pm

Hi Michael, i can do that, when the weather will be better, maybe thursday (only 50% rain forecast, tomorrow 90%). Nevertheless it would be more fun to see (hear) the calculated airspeed. Maybe you find time to add the airspeed-calculation ( i thought in my rashness, this is peanuts, against adding a second sensor in the code).

I know that the accuracy of the measurement depends on a lot of factors, for example the pressure -ratio inside the fuselage is one, temperature and density are others. But espacially the airspeed calculation is less problematic, because many errors are the same in both sensors and we calculate the difference between them. By the way, when thinking about it, Pitot and TEK is at least the same, only with different key signature in the calculation.

@tempo:

give it a try, the routines are there, the sensors are wellknown and cheap. I am really optimistic.

Edit: As we know, the ms5611 with the oxs-routines has a resolution of 0,1 m at least. This corresponds to a speed of 1,4 m/s = pedestrian:-)

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens » Tue Jul 08, 2014 8:29 pm

Carbo,
I am not so optimistic as you.
The accuracy of MS5611 is much less than 0.1 m.
If you look at your log (see previous post), you will see that the sensor filling the altitude has a drift of about 75 cm in a few min.
The second sensor has not the same drift (about 47 cm based on T1).

Using the difference of altitude to calculate the airspeed will generate significant errors.

As Tempo says, I think you need to use a TEK probe or a good diferential pressure sensor in order to have a good compensated vario.

I will wait for some logs from real flights (one altitude being reported by a pitot, the other being a static one) before making further changes to OXS.

User avatar
jhsa
Posts: 19382
Joined: Tue Dec 27, 2011 5:13 pm
Country: Germany

Re: Open source speed sensor?

Post by jhsa » Tue Jul 08, 2014 8:33 pm

I think a wind tunnel would be a nice tool to have at the moment :)

João
My er9x/Ersky9x/eepskye Video Tutorials
https://www.youtube.com/playlist?list=PL5uJhoD7sAKidZmkhMpYpp_qcuIqJXhb9

Donate to Er9x/Ersky9x:
https://www.paypal.com/cgi-bin/webscr?cmd=_s-xclick&hosted_button_id=YHX43JR3J7XGW

mpjf01
Posts: 367
Joined: Thu Jul 26, 2012 3:07 am
Country: -
Location: Adelaide, Australia

Re: Open source speed sensor?

Post by mpjf01 » Wed Jul 09, 2014 1:01 am

mstrens wrote:I would suggest using a 4525DO-DS5AI001DP sensor.
I think that this sensor can provide a higher precision for small speeds. This would be useful when airspeed is used too to calculate dTE (for a vario with total energy).
If someone requires very high speed measurements, it would be possible to use another sensor from the same serie with the same software.

If someone provide me a 4525DO-DS5AI001DP sensor, I can try to write the code in order to let OXS measure airspeed and dTE (when associated to a MS5611 vario sensor).
My preference would be to use a sensor that is judged accurate enough along with a proper pitot. If you wish to proceed with the use of the 4525 I am prepared to supply it. Probably the most efficient way, given our relative locations, would be either for me to purchase it from your local supplier and get it delivered to you direct or for you to purchase it and I will reimburse you using Paypal. PM me if interested.

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Speed sensor for low velocity

Post by mstrens » Wed Jul 09, 2014 9:24 am

Tempo wrote: (1) Using one MS5611 for TEK-variometer and a second MS5611 for altitude:
It´s only interesting for high speed. By low speed-thermal-circling with TEK the difference to an extra altitude sensor without TEK is low.
Changes in weather/atmospheric pressure during flying make more differences in altitude values.
Tempo,
You I right.
Just to quantify it, I made some calculations.
Imagine a glider flying at see level with a airpressure = 1000 hpa and 36 km/h (10m/s).
Static pressure would report 1000 hpa and altitude would be 0 meter (if reset just before landing)
When flying at 36 km/h pressure reported by a TEK probe would be about (1.2 * 10 *10 /2) = 0.6 hpa less than static pressure.
So the altitude would be reported as a little more than 5m.
This is based on the normalised air pressure saying that at sea level pressure is 1013.25 hpa and at 500 m it is 954.61 (so 0.01 hpa => 0.085 m)
If the flying speed is 90 km/h (25 m/s), then the difference of altitude becomes nearly 32 m (<=3,75 hpa.)
Tempo wrote: (2) I use another measurement principle for measuring low aerodynamic speed: thermal-flow-sensor.
I use this sensor-family :
Sensirion_Differential_Pressure_SDP600series_Datasheet_V1.7.pdf.

Speed range goes down to about 15km/h and up to 110km/h. I´m measuring total pressure at the nose of fuselage (on a f3j-machine it´s about 40cm in front of wing) and static pressure inside closed fuselage.
Where do find the sensor you are using?
I had a look on Google but I do not find a web supplier?
At Farnell, I can find some sensors SPD600 but they do not have the right range of pressure (and the right air connections).
Furthermore it seems quite expensive (about 100 $).
Tempo wrote: (3) I prefer TEK-probe for compensating variometer. You get good compensation with "Braunschweiger Düse" .
If you use electronic compensation with speed information, you need a good static pressure measurement, which does not depend on air speed. It´s not easy to get it.
I agree that using a TEK probe is probably the easiest way to get a compensated vario.

Still if some user want the airspeed, there is need for a good low pressure differential sensor.
I expect that calculating the airspeed based on such a differential sensor should be much better than by difference of 2 MS5611 foreseen to measure much higher (about 1000hpa ) absolute pressures.

You say that it is difficult to get a good static pressure.
I expect it is so if you use the static pressure inside the fuselage (combined with just a normal Pitot probe for the dynamic pressure).
But do you think that it is difficult to get correct airspeed values when using a Prandtl probe set in a proper location?
Building a Prandtl probe seems quite easy and the design seems less critical than an TEK-probe.

User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Re: Speed sensor for low velocity

Post by Tempo » Wed Jul 09, 2014 12:31 pm

mstrens wrote:
Tempo wrote: (2) I use another measurement principle for measuring low aerodynamic speed: thermal-flow-sensor.
I use this sensor-family :
Sensirion_Differential_Pressure_SDP600series_Datasheet_V1.7.pdf.
...
Where do find the sensor you are using?
I had a look on Google but I do not find a web supplier?
At Farnell, I can find some sensors SPD600 but they do not have the right range of pressure (and the right air connections).
Furthermore it seems quite expensive (about 100 $).
This link:
http://be.farnell.com/sensirion/sdp611/ ... 2-acce-sen
http://be.farnell.com/sensirion/sdp610- ... 2-acce-sen

Yes its expensive but adequate for aerodymanic experiments in range of low flow velocity as our RC-gliders.

User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

meassurement of pressure for velocity, altitude and TEK

Post by Tempo » Wed Jul 09, 2014 1:18 pm

mstrens wrote:...
You say that it is difficult to get a good static pressure.
I expect it is so if you use the static pressure inside the fuselage (combined with just a normal Pitot probe for the dynamic pressure).
But do you think that it is difficult to get correct airspeed values when using a Prandtl probe set in a proper location?
...
There are two situations:
You have a robust and easy mechanic solution by measurement of static pressure and total pressure at different locations (see picture A and B). In this case it´s difficult to get correct static pressure. Inside closed fuselage yields good approximation.
You have best aerodynamic solution by measurement of static pressure and total pressure at same location (see picture C and D).
But location had to be outside of fuselage- and wing-flow-field.
pressure_tubes_comparison.jpg
( 26.07.2014: picture with corrected pressure p and q )

Please note dimensions of Prandtl-Rohr. You need 4 to 8 holes round diameter for static pressure.
Prandtl_Rohr_Abmesssung.jpg
mstrens wrote:...
Building a Prandtl probe seems quite easy and the design seems less critical than an TEK-probe.
No, TEK-probe is easy to built. You only need aluminium tube with outer diameter 3mm, bend it to 72 degree, close it, bore a little hole (about 0,5mm diameter) and you have got "Nicksdüse". You can adjust compensation by distance of hole and pipe end.
The difficulty is to find position outside of fuselage- and wing-flow-field.
Reichmann_Motorbuchverlag_1982_Abmessungen.jpg
Reichmann_Motorbuchverlag_1982.jpg
Last edited by Tempo on Sat Jul 26, 2014 10:58 am, edited 1 time in total.

User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Precission measurement of velocity

Post by Tempo » Wed Jul 09, 2014 1:54 pm

jhsa wrote:I think a wind tunnel would be a nice tool to have at the moment :)

João
Better than wind tunnel is a "flying wind tunnel" like this 8-) :
2013_08_01_Schuß_FHT_HS_Foto.jpg
2013_08_01_Schuß_FHT_HS_Foto.jpg (28.31 KiB) Viewed 8396 times
At top end you spot a yellow flag, which detects angle of attack. In this example only one direction of flow is measured, but i will add a second one at 90 degree position ...
Right side you spot a "Wirbelzähler" (vortex-flow-meter) for high precission measurement of flow velocity. It is a completly different way to get velocity than with pressure measurements. No static pressure is needed, no temperature compensation is needed, no drifts ;) . "Wirbelzähler" is onetime calibrated in professional windtunnel.

Among other things this equipment makes possible a calibration of other sensors in flight !

mstrens
Posts: 1411
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Precission measurement of velocity

Post by mstrens » Wed Jul 09, 2014 3:43 pm

Tempo wrote:
jhsa wrote:I think a wind tunnel would be a nice tool to have at the moment :)

João
Better than wind tunnel is a "flying wind tunnel" like this 8-) :
2013_08_01_Schuß_FHT_HS_Foto.jpg
At top end you spot a yellow flag, which detects angle of attack. In this example only one direction of flow is measured, but i will add a second one at 90 degree position ...
Right side you spot a "Wirbelzähler" (vortex-flow-meter) for high precission measurement of flow velocity. It is a completly different way to get velocity than with pressure measurements. No static pressure is needed, no temperature compensation is needed, no drifts ;) . "Wirbelzähler" is onetime calibrated in professional windtunnel.

Among other things this equipment makes possible a calibration of other sensors in flight !
Thanks for this information.
I did not know about this kind of flowmeter.

Did you build such a sensor for an RC model?
If so, did you use a micro in order to collect the vortex frequency?
I presume that it requires then some good analog amplificator with filters.
Is it not to complex to build compare to a already made differential pressure sensor?

User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Precission measurement of velocity

Post by Tempo » Wed Jul 09, 2014 4:43 pm

@mstrens
In picture above and below you see RC-model of my own (wingspan 3,3m F3J-glider) "Wirbelzähler" fitted in picture above.
Yes, I use a micro with analog circuit (frequency-filters and amplifier with dynamic compressor) designed for my "Wirbelzähler".
Velocity is calculated by an optimized fourier-transform implemented on Atmega328.
Correct catching of vortex is important. Look here for functional principle : http://en.wikipedia.org/wiki/K%C3%A1rm% ... tex_street

(Inside my F3j-glider is microcontroller for datalogging and control loop. For example actual control loop brings fuselage to zero-angle of attack automatically. Measurement of flow angle by differential pressure with "5-hole-probe". I used apparatus in pictures for calibrating flow angle of "5-hole-probe" http://www.grc.nasa.gov/WWW/k-12/airplane/tunp5h.html. I have integrated "5-hole-probe" in fuselage nose of F3J-glider. Flow velocity is measured permanent with differential pressure calibrated by "Wirbelzähler" in flight. Yaw angle of "5-hole-probe" calibrated by yellow-flag-angle-reference in flight. In daily use you can watch nothing special on my F3J-glider except 5 holes at fuselage nose ... ;) )
P1010292.JPG
Last edited by Tempo on Wed Jul 09, 2014 5:46 pm, edited 1 time in total.


Post Reply

Return to “The Pickled Gnu (The Pub)”