Open source speed sensor?

Need some advice? Trying out a new idea? Fancy a beer?
Join us for some general banter and good times.
mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Precission measurement of velocity

Post by mstrens »

Tempo wrote:@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".)
P1010292.JPG
Amazing application!

I understand the principle. But do you really need a 5 hole probe instead of a 3 hole probe?
I imagine that you automatically control the rudder but not the elevator.

Is your "autopilot" allowed in competition?

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

Re: Open source speed sensor?

Post by Tempo »

@mstrens
Yes, for yaw angle detection you only need 3 holes. In this way my control loop actuates rudder.
But I also detect nick angle as angle of attack. This is another option to measure velocity, because velocity is also a function of nick angle for a defined aeroplane ! ;) In future I will control the rudder by control loop of flaps.

My "autopilot" is not allowed for current competition. I make things for future. I´m interested in aerodynamic problems like
measuring drag polar (http://en.wikipedia.org/wiki/Drag_Polar) of whole aircraft in flight, optimized setting of flaps during thermal flights, flying with zero angle of fuselage, calibration of sensors in flight, ...
Carbo
Posts: 467
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo »

Hi Michael,
the weatherforecast was right, so here is a flight with a speed-probe on a second MS5611, logged as T1. GPS was on board for
comparison purposes.
Esprit_E-2014-07-10.zip
Test Speedsensor
(92.52 KiB) Downloaded 368 times
mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens »

Carbo wrote:Hi Michael,
the weatherforecast was right, so here is a flight with a speed-probe on a second MS5611, logged as T1. GPS was on board for
comparison purposes.
Hi Carbo,
Thanks for the log.
I had a look at it.

There is clearly a bug somewhere in XOS because Temp1 becomes negative whenthe value increase to much.
I will have to look at the code to fix it. Some one else had already notified me such an issue but now I have a better information to find the bug.

On the other side, I imported the log in XLS and I corrected the negative value adding 2^16.
I calculated the difference of altitude between the 2 sensors and I converted this difference in pressure (hectopascal) and finaly I converted this pressure it in airspeed (m/sec and km/h).
So I was able to compare the GPS speed and the (MS5611) Airspeed. I put it in a chart.
You clearly see a relation between the 2 but not always. This is not abnormal because I expect that GPS speed is not always ok and furthermore it is a ground speed and not an airspeed. If there where some wind, it can make a difference depending on the direction you flight.
I also put a chart to compare Altitude and Airspeed.
It seems that sometimes you rise a lot using the eletrical motor at full power and sometimes at mid power.
Do you recognise the phase of flight in those chart.

I also notice a drift between the 2 MS5611 sensors. This drift seems quite big because it is equivalent to nearly 10 km/h (at the begining the average airspeed is about 30km/h and at the end 40 km/h).

Here in attachement the xls with the charts.
Attachments
Esprit_E-2014-07-10 with airspeed.rar
(548.85 KiB) Downloaded 339 times
Carbo
Posts: 467
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo »

Meanwhile i changed the sensor-IDs in the arduino sketch and mounted a TEK-Probe to fly with dTE. The result was not as expected, when i flew loops and downturns. Only for a short period i noticed the influence of the TEK, then the vario behaved as usual. After the landing i checked the sensor: the connection to the tube was not leakproof. That explains the behavior, i fear, that the speed probe suffered the same fault. I have just sealed the connection with silicone and will do further tests, as soon as the weather permits.

Michael: would it be possible, using the ppm-connection, to change the sensor-IDs on the fly? Using eg. 0-100% for adjusting the sensitivity and when you have for 1 second -50% switch the IDs? So you could switch TEK on and off while flying.

mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens »

Carbo wrote:Meanwhile i changed the sensor-IDs in the arduino sketch and mounted a TEK-Probe to fly with dTE. The result was not as expected, when i flew loops and downturns. Only for a short period i noticed the influence of the TEK, then the vario behaved as usual. After the landing i checked the sensor: the connection to the tube was not leakproof. That explains the behavior, i fear, that the speed probe suffered the same fault. I have just sealed the connection with silicone and will do further tests, as soon as the weather permits.

Michael: would it be possible, using the ppm-connection, to change the sensor-IDs on the fly? Using eg. 0-100% for adjusting the sensitivity and when you have for 1 second -50% switch the IDs? So you could switch TEK on and off while flying.
It is not 100% clear what you mean by sensor id.
Do you mean that during the flight you would like to send sometime as VSpeed the value from sensor 1 and sometime the value from sensor2 (depending on PPM)
Carbo
Posts: 467
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo »

Yes, in openxsensor.ino you defined it. I dont know, if the code allows to change it on the fly. Also only the starting sensor is zeroed, so it possibly does not make sense, you wuld switch the altitude sensor too.
ABBC3_SPOILER_SHOW
mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens »

Carbo wrote:Yes, in openxsensor.ino you defined it. I dont know, if the code allows to change it on the fly. Also only the starting sensor is zeroed, so it possibly does not make sense, you wuld switch the altitude sensor too.
ABBC3_SPOILER_SHOW
What you named the sensor id is the I2C address of the sensor.
I expect that, in a test version, it should be possible to change it. I have to look forward.
Still for a little time (a few second), the alt and Vspeed would become wrong because OXS would merge the values from the 2 sensors.
Then the altidude would become wrong because there is a reset only at start up and this occurs on Tx (and not on XOS; XOS send always the altitude above sea and not the relative one). The fact that altitude become wrong, has no impact on the Vspeed because this is calculated in OXS by difference of altitude).

Note: about the issue for Temp1. I think it is not a bug. I think it occurs because OXS transmit a value that exceed the max value allowed for this field. The solution is to set a value = 10 (instead of 1) for the divider in the config file for the field TEMP1
Carbo
Posts: 467
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo »

Today i flew 5 different TEK-probes, the effect was from nearly nothing to quite acceptable compensation.

The second best:
TEK1.jpg
The best:
TEK5.jpg
The flight with this one was quite amazing, a youtube video with some clippings is on its way, i will add the link later. Nevertheless its far away from a total energy compensation (see the loops and waves). But at the beginning, at 5s, 7s, and 10s i heard someting like a disturbance in the vario-tone, after looking at the video, there were indeed "mini-thermals" that moved the prop upwards and were clearly noticed by the sensor. Although they are too small to use them, its cool to hear them. I felt, to have a better feeling for the situation in the air.

http://youtu.be/mU1d3Ccjnc8

To calculate the speed out of the TEK-Vario and a normal vario seems impossible, here is the log:
Esprit_E-2014-07-10.zip
(92.52 KiB) Downloaded 322 times
So i will make some more flights with the speed-probe to get more data. Maybe there is a chance to test the calculated energy compensation. The provisionally idea is to zero both sensors on the ground and to have a possibility to correct the drift of the sensor in flight with the ppm-input. This could maybe lead to a really total energy compensation.
Last edited by Carbo on Fri Jul 11, 2014 8:56 pm, edited 1 time in total.
mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens »

Carbo,
Thanks for the feedback.

I think then when you use a TEK probe, it is not possible anymore to calculate the airspeed afterwards.

Looking to you new log file, I notice again a drift between the altidudes reported by the 2 MS5611 (if we can assume that there was no wind at the beginning and at the end of your log).
This drift is about 3 meter. This would have an impact on a calculated airspeed based on 2 MS5611.

FYI, I worked to day on OXS in order to let it read a differential sensor (4525DO) and calculating/transmitting the airspeed.
Normally the job is done but I can't test it because I have no 4525DO sensor.
This would normally provide a more accurate airspeed.

Still, I presume that using one MS5611 connected to a good TEK remains the easiest and probably the best accurate way to get a total energy vario.

Please note that I have read on the web that a 100% well compensated vario will report vertical speed when you start applying elevator up or down. There is a good physical explanation therefore.
I also presume that the purpose of a compensate vario is to get better feeling while applying quite small up or down elevator but not to keep a zero "Vspeed" during a full loop.
User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Re: Open source speed sensor?

Post by Tempo »

Hallo Carbo,

bei deinen TEK-Düsenbildern fällt mir folgendes auf:

Die Nicksdüse im ersten Bild hat eine etwas zu große Bohrung für die Druckabnahme. Versuche sie mit etwa 0,5 mm Bohrungen zu machen. Zu große Bohrungen produzieren ihrerseits wieder unerwünschte Umströmungen am Düsenrohr.
Vermutlich hast du wegen der einfachen Verbindung zum Sensor die TEK-Düsen an der Stelle im Bild angebracht. Das ist aber relativ nahe im Strömungsfeld der Tragfläche und des Rumpfes. Die Beeinflussung der Strömung geht einige Zentimeter auch vor die Strömungskörper.
Deshalb wäre ein deutlich größerer Abstand besser, z.B. klassischerweise oben am Seitenleitwerk oder weit vor der Rumpfspitze (ginge auch mit Elektromotor, wenn dieser eine Hohlwelle besäße; aber empfindlich gegen Beschädigung bei Landung). Zur Not auch einfach ein längeres Röhrchen an der vorhandenen Position einsetzen.

Ist im zweiten Bild eine Zweiloch(Vierloch)-Nicksdüse oder eine Zweischlitzdüse zu sehen ? Reflexionen am Metallröhrchen machen das undeutlich. Gegebenenfalls ist der Einschnitt etwas zu tief, soweit auf dem Foto erkennbar.
Eine Zweilochnicksdüse, bei der beide Löcher knapp nebeneinander gebohrt sind , funktioniert auch sehr gut und ist leichter herzustellen als eine Zweischlitzdüse.

Für eine Gesamtdruckabnahme zur Fluggeschwindigkeismessung wäre ein Loch im Propellerspinner mit Hohlwelle ideal. Oder alternativ eine Gesamtdruckabnahme om oberen Ende des Seitenleitwerks. Auch hier gilt es, einen großen Abstand von benachbarten umströmten Flächen zu haben, wenn eine Geschwindigkeitsanzeige hinreichend genau werden soll.

Ach ja, ein absolut luftdichtes System von der Öffnung bis zum Sensor ist zwingend nötig, wie du bereits festgestellt hast.
Ich prüfe so etwas z.B. mithilfe einer Einwegspritze (vom Arzt), indem ich Druck auf die Leitungen gebe und dies dann abklemme.
Bei geringen Drücken reicht der Reibungswiderstand der Spritze aus, ohne Leitungen abzuklemmen. Dann 5 bis 10 Minuten kontrollieren, ob die Druckanzeige des Sensors konstant bleibt. Dann wäre das System dicht. Das bietet sich vor allem für die Schrumpfschlauchverklebung zum MS5611 an.

Ich baue übrigens gerade auch einen Schlauchanschluß an mein openxvario an. Dabei verwende ich ein 3mm Kunststoffröhrchen, das ich horizontal auf der Platine über dem MS5611 luftdicht einklebe. Dann kommt ein Silikonschlauchübergang für meine harten, aber "lange" Leitungen zum Seitenleitwerk dran. Das wäre ein weiterer Punkt, der zu beachten ist: Lange Leitungen müssen hart sein, sonst verfälschen Schlauchdeformationen bei Beschleunigungen z.B. in Turbulenz oder in unruhiger Thermik die Anzeige.
mstrens
Posts: 1435
Joined: Fri Dec 27, 2013 7:49 pm
Country: -

Re: Open source speed sensor?

Post by mstrens »

Tempo wrote:Hallo Carbo,

bei deinen TEK-Düsenbildern fällt mir folgendes auf:

Die Nicksdüse im ersten Bild hat eine etwas zu große Bohrung für die Druckabnahme. Versuche sie mit etwa 0,5 mm Bohrungen zu machen. Zu große Bohrungen produzieren ihrerseits wieder unerwünschte Umströmungen am Düsenrohr.
Vermutlich hast du wegen der einfachen Verbindung zum Sensor die TEK-Düsen an der Stelle im Bild angebracht. Das ist aber relativ nahe im Strömungsfeld der Tragfläche und des Rumpfes. Die Beeinflussung der Strömung geht einige Zentimeter auch vor die Strömungskörper.
Deshalb wäre ein deutlich größerer Abstand besser, z.B. klassischerweise oben am Seitenleitwerk oder weit vor der Rumpfspitze (ginge auch mit Elektromotor, wenn dieser eine Hohlwelle besäße; aber empfindlich gegen Beschädigung bei Landung). Zur Not auch einfach ein längeres Röhrchen an der vorhandenen Position einsetzen.

Ist im zweiten Bild eine Zweiloch(Vierloch)-Nicksdüse oder eine Zweischlitzdüse zu sehen ? Reflexionen am Metallröhrchen machen das undeutlich. Gegebenenfalls ist der Einschnitt etwas zu tief, soweit auf dem Foto erkennbar.
Eine Zweilochnicksdüse, bei der beide Löcher knapp nebeneinander gebohrt sind , funktioniert auch sehr gut und ist leichter herzustellen als eine Zweischlitzdüse.

Für eine Gesamtdruckabnahme zur Fluggeschwindigkeismessung wäre ein Loch im Propellerspinner mit Hohlwelle ideal. Oder alternativ eine Gesamtdruckabnahme om oberen Ende des Seitenleitwerks. Auch hier gilt es, einen großen Abstand von benachbarten umströmten Flächen zu haben, wenn eine Geschwindigkeitsanzeige hinreichend genau werden soll.

Ach ja, ein absolut luftdichtes System von der Öffnung bis zum Sensor ist zwingend nötig, wie du bereits festgestellt hast.
Ich prüfe so etwas z.B. mithilfe einer Einwegspritze (vom Arzt), indem ich Druck auf die Leitungen gebe und dies dann abklemme.
Bei geringen Drücken reicht der Reibungswiderstand der Spritze aus, ohne Leitungen abzuklemmen. Dann 5 bis 10 Minuten kontrollieren, ob die Druckanzeige des Sensors konstant bleibt. Dann wäre das System dicht. Das bietet sich vor allem für die Schrumpfschlauchverklebung zum MS5611 an.

Ich baue übrigens gerade auch einen Schlauchanschluß an mein openxvario an. Dabei verwende ich ein 3mm Kunststoffröhrchen, das ich horizontal auf der Platine über dem MS5611 luftdicht einklebe. Dann kommt ein Silikonschlauchübergang für meine harten, aber "lange" Leitungen zum Seitenleitwerk dran. Das wäre ein weiterer Punkt, der zu beachten ist: Lange Leitungen müssen hart sein, sonst verfälschen Schlauchdeformationen bei Beschleunigungen z.B. in Turbulenz oder in unruhiger Thermik die Anzeige.
Tempo,
Thanks for all the practical explanations.
I used google to get a translation to english. I understand the main part but the automatic translation is not always good.
If you have time, I think it would be good writing a translation in english .
It could be usefull for other users.
User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Re: Open source speed sensor?

Post by Tempo »

Further informations on TEK-pipes:

viewtopic.php?f=86&t=5856
User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Re: Open source speed sensor?

Post by Tempo »

mstrens wrote:
Tempo wrote:Hallo Carbo,

bei deinen TEK-Düsenbildern fällt mir folgendes auf: ...
Tempo,
Thanks for all the practical explanations.
I used google to get a translation to english. I understand the main part but the automatic translation is not always good.
If you have time, I think it would be good writing a translation in english .
It could be usefull for other users.
You get significant details there viewtopic.php?f=86&t=5856&p=83775#p83775
paulj
Posts: 22
Joined: Sun Jun 29, 2014 8:20 am
Country: -

Re: meassurement of pressure for velocity, altitude and TEK

Post by paulj »

Tempo wrote: 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
Hey Tempo - could you please give me the book reference where you have scanned this information? I would be interested to read more about this topic! I flew full size for a number of years, but took some of the technology for granted - shame on me...

Many thanks,
Paul

PS: I understand it is in German- it will be a good chance for me to practice more :)
User avatar
Tempo
Posts: 83
Joined: Tue Feb 04, 2014 4:04 pm
Country: -

Re: Open source speed sensor?

Post by Tempo »

Hi paulj,

yes, you are right it is "Helmut Reichmann: Streckensegelflug; Motorbuchverlag Stuttgart 1982; ISBN 3-87943-371-2"
You can also get an English translation:
Reichmann, H. Streckensegelflug. Motorbuch Vlg., Stuttgart, 1975. English translation as Cross-Country Soaring published in 1978 by the Soaring Society of America, ISBN 1-883813-01-8.

(Further Information: Who is Helmut Reichmann ? http://en.wikipedia.org/wiki/Helmut_Reichmann )
paulj
Posts: 22
Joined: Sun Jun 29, 2014 8:20 am
Country: -

Re: Open source speed sensor?

Post by paulj »

Thanks Tempo. Helmut Reichmann is clearly a legend! I wish I had come across this when I was actively involved in gliding (full size). I have tracked a copy down on eBay, and look forward to reading it.
Carbo
Posts: 467
Joined: Fri Aug 02, 2013 6:55 pm
Country: Germany
Location: Freinsheim RP

Re: Open source speed sensor?

Post by Carbo »

Meanwhile the open source speed sensor is working:

viewtopic.php?f=86&t=6080

Post Reply

Return to “The Pickled Gnu (The Pub)”