A functional 360° Servo for Schottels/Thrusters

[Guest MdG-RC]

Hello Guy's,

 

Here a special topic about my functional 360° servo....

 

 

I will tell a small story first, how it starts with this 360° servo development...

 

Two years ago, there was a question on a Dutch Forum...

In this discussion there was one main question... "Is it possible, to make a servo with endless rotation control"

Most answers were Yes it's possible... but you have to use a microprocessor...

 

In that period, I didn't even now how a microprocessor works.

So I didn't get a answer...

After a small year, I start learning to program a microprocessor...

My teacher was (and still is) "Google" and I starts reading,reading and reading...

Some months later, I did get enough info to make a start with this "nice" project/challenge.

With some "Trial and Errors" I write finally a good working program for a functional endless servo control.

Here a video from a test-setup:

This servo will always take the "shorted way"

 

The hardware for this servo is a very easy and basic setup...

Just two potentiometers (endless), a H-bridge to control the servo-motor and a microprocessor.

The only thing is, the potentiometers are expensive...

I hope I can design my own potentiometers, to reduce the price and the size !

 

In the future (hopefully this summer) I will upgrade some parts...

On the moment I'm using a 8-pin microprocessor... for a new concept, I will use a other microprocessor with more I/O.

The 8-pin microprocessor is now running on his internal clock/resonator and with the bigger one, I can change this to a external resonator...

This makes the servo faster and better.

 

I have test this servo with a Graupner MC-19 radio, a Futaba F-14 and a FC-16...

I suggest to use it with a digital radio, just to make it easier to swap the direction.

With a F-14 radio, the only option to swap the direction is mechanical.

(in the future I hope to make this easier)

A good thing is, it's working with the most used radio systems

 

In the transceiver it's only a potentiometer and is mostly "plug and Play"

On the servo-side there is one PCB (Printed Circuit Board) with all components together...

With one PCB I can control two servo's (each servo used his own microprocessor and H-bridge)

This servo will work from 6V to 12V without any problems.

The RPM from the Schottel/Thruster is around the 15/RPM.

This is a nice speed and gives a smooth reaction.

(the motor on my video, is running a bit to fast)

 

If there are any questions, just let me now

(I understand German language)

 

For anyone how go's to the meeting today... Have Fun and enjoy your time

[Guest HBose]

Hallo Marten

 

schöne Lösung hast du da gebaut.

 

Ich hätte eine Frage zu deiner Lösung.

 

Sowie es aussieht hast du 360° Potis verwendet?

Die Bauteile haben aber leider systembedingt zwar mechanisch einen Drehwinkel von 360° Endlos aber elektrisch liegt der nutzbare Bereich bei 325°-340°. Die Schleifbahn des Potis darf halt nicht durchverbunden sein, dann ist es kein Poti mehr.

 

Diese Eigenschaft führt zu einem "elektrischen toten Winkel" von 20°-35°. Man kann das zwar etwas kompensieren durch einen Widerstand zwischen Schleifer und Potiende. Den Widerstand sieht man auch auf deinem Video, der direkt am Poti angelötet ist. Das Verhindert aber nur unsaubere Zustände und kann nicht zur Auswertung des toten Winkels benutzt werden.

 

Wie kompensierst du den nicht auswertbaren Bereich des Potis? Oder hat das verwendete Poti keine toten Winkel? 20°-35° toter Bereich sind schon störend beim steuern.

 

Salu Hans

[Guest MdG-RC]

Hello Hans,

 

The potentiometers have a full 360° range...

They are working with a magnet-sensor !

This works great and a other good this is, they are contactless

 

Later this evening, I will post my idea for the radio-setup... (Iḿ now not at home)

[Fabian F.]

Dear Marten,

 

given the fact taht the sensors are quite expensive and you use larger working range than 360°, i think you are using something that is called a rotary encoder (Inkremental Drehgeber). Some time ago i also had some ideas on how to make such a Schottel servo and there i found some interesting encoders. E.g. the ones of POSIC. Have you seen them before?

 

I am very excited hearing from you soon.

 

Greetings from hot Munich,

Fabian

[Guest HBose]

Hallo Marten

 

danke für die Antwort.

 

Diese Spezialpotis sind natürlich sehr teuer . Sie werden primär in der Industrie für Steuerungsaufgaben eingesetzt. Man findet sie vom Bagger bis zur CNC-Fräse. Das sind natürlich die absoluten Edelbauteile.

 

Das erklärt auch die hohen Kosten.Megatron ist da eine Quelle. Hans (Whisper) und ich hatten uns diese Lösung auch angesehen und als zu teuer verworfen. Wenn man zwei Antriebe in einem Schiff verbauen muss wird es richtig teuer für die Bastelkasse.

 

Salu Hans

[Guest MdG-RC]

Hello Fabian,

 

Thanks for your link and info

 

About the sensors...

No the sensors are not the same as the one you show to me.

The problem with that sort sensors is, you don't get a real-position !

It only gives a pulse on his output.

So when the microprocessor lose his count... you don't no where the Schottel exactly is.

Whit the potentiometers I always get my exact position feedback.

When I switch off my radio and I turn the potentiometer a bit...

I can be sure that when I put the radio back on again, that the servo go's to his correct position.

(When my radio is off and the servo is that moment on, the servo go's to his preprogrammed Home-Position )

It's possible to record a new Home-Position whit my servo-system.

 

The sensor in the potentiometer, they give a analog output from 0V to 5V (linear).

The resistor on my video, that is to Pull-down the output from the sensor.

When the position is 0° the output is 0Volt, when the position is 10° the output is 0,14Volt..... and the position 360° is a output from almost 5Volt.

 

And yeah, the potentiometers are to expensive for our goal ...

I hope to make them (cheaper) by myself...

 

@ Hans,

 

Yeah that's correct and I'm using Megatron potentiometers.

They cost €50,- each ! (if you buy more, then they will give you more discount)

[hoppppla]

Hallo,

 

würden sich für diesen Zweck nicht auch die hier eigenen?

Das wäre wohl ein ganze Stück günstiger, die Mechanik müßte man sich halt selber bauen. Und die Magneten sauber zentrieren.

Hätte auch den Vorteil, das man den aktuellen Winkel digital auslesen könnten, und damit einen weitere Fehlerquelle (den AD-Wandler) aus dem Spiel nimmt.

 

viele Grüße,

Hermann

[Guest HBose]

Hallo Hermann

 

das ist ein Teil des Innenlebens der eingestzten Drehgeber. Ihre Handhabung erfordert sehr viel Programmier KnowHow.

Aber es gibt auch andere Ansätze als den von Marten. Siehe diesen Thread : http://www.schiffsmodell.net/showthread.php?t=36688&referrerid=6292

 

Salu Hans

[Guest MdG-RC]

Hello Herman,

 

I have ordered some IC from them to give it a try.

In Juli I have more time to play with this.

The IC's are already here and the magnets are coming (hopefully) this week to.

 

Whit the new microprocessor it is possible to use the digital (serial) connection.

A good thing whit the analog output, I only need one input

It works fine but it can be much better...

 

@ All,

 

Here my setup for in the radio, to control the Schottels:

Full Resolution

 

For the speed-controller, I will use also a magnet-sensor...

It has the same analog output like the potentiometers.

When the magnet is close to the sensor, the output is 0Volt ore 5Volt (it's dependent with the North ore South side from the magnet)

Farther the magnet, the voltage will change until 2,5Volt.

(when the magnet is removed, the output is 2,5V)

 

This setup is totally contact less (for the servo and for the speed-controller)

The magnet for the speed-controller is going Up and Down when you will move the handle.

The servo-potentiometer is mechanical connected with a gear.

The exact sizes I don't now, this is just a basic plan.

 

Any other ideas ore suggestions are always welcome

[hoppppla]

Hallo,

das ist ein Teil des Innenlebens der eingestzten Drehgeber. Ihre Handhabung erfordert sehr viel Programmier KnowHow.

hm, der Programieraufwand hält sich meines Erachtens in Grenzen, das verwendetet serielle Protokoll sollte eigentlich kein Problem sein. Mehr Probleme seh ich bei der Mechanik, da die Drehgeber sehr genau zentriet werden müssen. Auch für die Kalibrierung beim Start müßte man sich noch was überlegen, ich hab noch keine Ahnung, wie sehr die Dinger driften.

Aber es gibt auch andere Ansätze als den von Marten. Siehe diesen Thread : http://www.schiffsmodell.net/showthread.php?t=36688&referrerid=6292

Jup, den kenn ich schon, leider steht dort nichts über die verwendeten Sensoren, insofern ist das für mich bis jetzt noch nicht so interessant

 

viele Grüße,

Hermann

[Guest MdG-RC]

Hello Hermann,

 

I hope that I can build them self, including a good mechanical setup.

It's a hard job I think, to let it work properly.

I will come back later about this...

 

@ All,

 

Here some info about my testing model:

http://www.mobilemarinemodels.com/acatalog/Afon_Alaw.html

I only get the casco from the Aflon Alaw !

I now it's not a original Schottel Tug, but for a test it will be a nice one I think.

It has enough space for the Schottels...

The Schottels, I will place them on the front-side (like a Tractor-Tug setup)

For this test I will use the Graupner Schottels type II.

 

When this test is Ok, I will build one set (two Schottels) in my Karl model.

The Karl is now active in Holland and I hope I will get a live demonstration on this Tug.

I now the captains and I live very close to them

Here a PDF from the Arion: www.iskestugs.nl/assets/Uploads/Iskes-ARION.pdf

 

If there are any questions about this... feel free and ask

[Guest MdG-RC]

Hello @ All,

 

Here a picture from my Servo driver-board:

Edit: The sizes are 75mm X 45mm

 

With one Servo driver-board, it's possible to control two Schottels separate.

Each Schottel gets his own microcontroller and H-bridge.

(on the picture, I have made one single servo functional for a test)

 

Edited July 2, 2010 by MdG-RC

[Admiral von Schneider]

epic...

 

Looks really nice!

If I ever build a ship with a thruster i will contact you.

[Guest MdG-RC]

epic...

 

Looks really nice!

If I ever build a ship with a thruster i will contact you.

 

Thanks for your interest

 

@ All,

 

Here some pictures from my test-model:

 

 

 

I'm still waiting for my Graupner Schottels...

I don't no why it takes so long...

 

I hope to put here soon a update from a functional test-model.

(including some video's)

[hoppppla]

Hallo,

 

hübsche Platine

ist das eine L298? Bzw. was für ein Mikrocontroller? Überwachst Du den Motorstrom?

So langsam juckts mich echt in den Fingern, das auch mal auszuprobieren

 

viele Grüße,

Hermann

[Guest MdG-RC]

Hello Hermann and All,

 

Yes, it's a L298 and on this moment I don't check the current from the servo-motor.

There are no more inputs on the PIC microprocessor !

In the future, I will make this option functional.

 

It's still a trial/error-process...

My favorite microprocessor for RC is a PIC 12F683

For this 360° servo he is a bit to slow (without a external resonator)

It's runs now on his internal resonator.

But when the servo is running under max 10/RPM he is OK. (see next video)

 

http://www.youtube.com/watch?v=wOaCrZGRDFk

 

During this record, I used my Futaba FC-16 radio including a normal servo.

This normal servo is running together with the 360° servo by a 1/1 Mix.

(Channel 5 is the 360° servo and channel 6 is the normal servo)

With the normal servo, it's good to see where the breakpoint is between 359° and 0°.

(The servo makes a Full move Max<-->Min.)

 

The 360° servo has a Failsafe function and when this is activated, the 360° go's to his Home-position.

It's possible to change the Home-position all the time .

 

The Failsafe is activated when the radio-signal is failed ore when the transceiver is still off.

(see last part from the video)

 

The next video (not yet reordered) I will show this same 360° servo...

Only than it's running with a Stepper-motor

[Guest MdG-RC]

Hello @ All,

 

Here the video with a Stepper-Motor

http://www.youtube.com/watch?v=Clfwo5yjt4w

It works great and this servo is fast, direct and more precise.

I hope they are strong enough to drive a Schottel... when the Schottel gives his full output.

 

On this video I reduced it with 3:1 (15T teeth for the Motor and a 45T on the potentiometer)

 

Comments are welcome...

[Guest Marc-Hamburg]

Hallo Marten,

 

um auf Schrittmotoren zu antworten, möchte ich vorher etwas fragen. Da Du auch den Karl baust und auch die Schottel II von Graupner einbauen möchtest, wie möchtest Du die Propeller drehen lassen und mit welcher Drehzahl ?. So wie Graupner es vorschlägt ( mit dem Kegelrad ), oder die Motoren oben raufstellen ( wie Wedico / Hegi es in dem Bauplan vorschlägt ) ?.

 

schöne Grüße Marc

[hoppppla]

Hallo Marten,

 

der Schrittmotor ersetzt bei diesem Aufbau einfach den Richtungsantrieb durch einen "normalen" E-Motor, also ist im Boot weiterhin ein 360°-Poti nötig, wenn ich das richtig verstehe, oder?

Lohnt sich dann der zusätzliche Aufwand bei der Ansteuerung?

Würde mich mal interessieren, wie der Aufbau im Vergleich zur Ansteuerung durch einen "normalen" E-Motor abschneidet.

Vielen Dank übrigens für die ausführliche Dokumentation Deines Ansatzes und die ausführliche Beantwortung von Fragen

 

viele Grüße,

Hermann

[Guest MdG-RC]

Hello Marc,

 

I will build them like Wedico/Hegi it suggest...

This gives less noise and the plastic gears from the Graupner Schottels are not OK.

It looks they are not centered.

 

@ Hermann,

 

During my test with a "Schrittmotor" I'm still using the 360°-Poti.

A Schrittmotor don't now his direct position by himself...

This means (without the 360°-Poti), the processor must first make a good calculation from the total steps, what the Schrittmotor must make forward/backwards...

The Schrittmotor Driver-board in my test has a MicroStep function....

One step with a Schrittmotor is now --> 1.8° : 16 = 0,1125°

For a complete 360° move, the Schrittmotor must make 3200 steps.

This gives a better resolution and is very precise, also the speed can be higher.

But....

When this servo is running without the 360°-Poti and with a Schrittmotor....

First I have to work with a Home-position with a light-sensor.

This sensor gives the Zero-Position from the Servo.

When the Schrittmotor mist some steps, the count is incorrect and the position is not okay.

And if I turn the Servo by hand, the the servo will not correct himself !

This is my reason, why I'm still using the 360°-Poti to give a correct feedback.

 

After this test....

I suggest to use a DC-motor, because it's smoother and correct enough and we don't need the full-speed from the Schrittmotor.

 

I hope this answered your question...

[Guest Marc-Hamburg]

Hallo Marten,

 

danke für die schnelle Antwort.

Für die Version und einer Übersetzung von 3:1 ( 3 Schrittmotor, 1 Antrieb) und ca 2500 U/min am Propeller, braucht man, meine ich, schon eine hohe Stellkraft / Haltekraft.

Bei meinem Test, hatte ich schon Probleme die Antriebe mit der Hand bei Vollast zu drehen.

Ich hatte aber auch eine 1:1 Übersetzung, die ich aber noch in 4:1 bzw. 6:1 ändern werde, wenn ich bei einem Servomotor bleibe.

 

schöne Grüße Marc

[Guest MdG-RC]

Hello Marc,

 

For my DC-motor servo, I'm using a RB-35 (conrad) motor with a RPM from 30 without last and 26/RPM with last.

It's already hard to turn this by hand, without any power on the motor.

This is a good thing and they are hopefully strong enough for the Graupner Schottels.

 

With a normal servo including that 3:1 ratio, I'm sure they can't hold it easy.

If you make the ratio bigger, then you need a very strong servo.

The current will be very high to !

 

I think/hope that my servo's will use a maximum 500Ma with last

[Guest Marc-Hamburg]

Hallo Marten,

 

ich meinte das mit meinem Servo : zum Beispiel : 60 Zähne an dem Antrieb und 10 Zähne auf dem Servo. Mein Fehler. Der Servo läuft nur als Motor zum drehen.

Wie bei Dir ist das Poti ausgelagert.

 

schöne Grüße Marc

[Guest MdG-RC]

Hello Marc,

 

Here a picture from the 360°-Poti:

Copyright @ MegaTron

 

Here the data-sheet link: http://www.megatron.de/export/DB_MAB25A_engl.pdf

 

I hope this gives you a idea how it looks.

Otherwise... please wait until I finished my test-model.

 

Edit: Here some pictures from my Karl (Arion):

 

The old drive-units...

 

Edited July 12, 2010 by MdG-RC

[Guest Marc-Hamburg]

Hallo Marten,

 

vielen Dank für Deine Information.

Ich bin sehr neugierig darauf, wie Du die Lösung über 360° hinbekommen hast. Also von 359° auf 1° ohne das der Antrieb den ganzen Weg über 180° zurück geht.

Wie ich auf Deinen Video´s geshen habe, dreht der Antrieb sehr gut und hat einen schönen Nachlauf, wenn man schnell von 0° auf 180° dreht.

 

Dein Test-Modell sieht ja fast so aus wie meins

 

schöne Grüße Marc