PyPilot

[bcn]

That one went into a wrong thread earlier today:
--------------------------------------------------------------------
Looking into the translations at Crowdin a question:

"Displays lines on the chart showing the autopilot's configured headings"

Is heading here a synonym for course or is it really heading, i.e. orientation?
There are some other sentences raising the same doubts.

And another one:
Servo Period
Please explain..

Hubert

[seandepagnier]

Originally Posted by bcn

That one went into a wrong thread earlier today:
--------------------------------------------------------------------
Looking into the translations at Crowdin a question:

"Displays lines on the chart showing the autopilot's configured headings"

Is heading here a synonym for course or is it really heading, i.e. orientation?
There are some other sentences raising the same doubts.

It's a little ambiguous because it really is "heading" if steering a compass course, and it really is "course" if steering a gps course. If steering to wind, it is not really either.

And another one:
Servo Period
Please explain..

The servo period is the minimum time the motor moving the rudder can be either on, or off.

[verkerkbr]

Dear BA,


Pypilot is an interesting project. But I have perhaps a simple question. I see window-wiper motors etc.



I have an RPI3B+ with the latest OpenPlotter version 1.0.0 running.


I also have an IMU module and an tillerpilot with an nmea183 connection.


What do I need to get PyPilot running. Can the Pypilot instructions to the AP send via NMEA183 or do I need a controller board to switch the tillerpilot motor direct ?


Regards,


Bram

[seandepagnier]

You would need a motor controller to switch the tillerpilot motor directly bypassing all of the circuit boards on the tiller pilot.

nmea0183 does not have sufficient sentences for autopilot control. The best you can do is send it wind data, or route commands, and this is used by the internal controller. Not sure which tillerpilot you have, but all the ones I have seen do not have gyroscopes, and have gimbaled compass that can literally swing around. This gives poor performance and you cannot easily tune them or control them remotely.

As far as I know, there are no provisions for motor control over nmea0183. If it were possible to move the motor directly it would be possible to use pypilot with no additional hardware.

I have motor controllers available on pypilot.org, and they can work with any dc motor. I am also planning to design a new controller for half the cost in the next few months, but I do not know when it will be ready and tested.

[rgleason]

Sean, I'll be interested in the new control too, as once my Raymarine AP Control fails, I'll need a good control for my RM wheelpilot. I suppose it could be controlled by a PC using OpenCPN and the pypilot_pi plugin via nmea? or even an RPI running Openplotter.



https://github.com/pypilot/pypilot/wiki/controller

I have motor controllers available on pypilot.org, and they can work with any dc motor. I am also planning to design a new controller for half the cost in the next few months, but I do not know when it will be ready and tested.

[verkerkbr]

Originally Posted by boat_alexandra

You would need a motor controller to switch the tillerpilot motor directly bypassing all of the circuit boards on the tiller pilot.

nmea0183 does not have sufficient sentences for autopilot control. The best you can do is send it wind data, or route commands, and this is used by the internal controller. Not sure which tillerpilot you have, but all the ones I have seen do not have gyroscopes, and have gimbaled compass that can literally swing around. This gives poor performance and you cannot easily tune them or control them remotely.

As far as I know, there are no provisions for motor control over nmea0183. If it were possible to move the motor directly it would be possible to use pypilot with no additional hardware.

I have motor controllers available on pypilot.org, and they can work with any dc motor. I am also planning to design a new controller for half the cost in the next few months, but I do not know when it will be ready and tested.

Dear BA, thanks for the reply.

My Tillerpilot is a Simrad TP30. I have never used the NMEA183 connector.

Any idea what instructions this tillerpilot get via NMEA. Is it simply a bit left or right movement of the ram ?

What NMEA sentences use such an autopilot ?

I'am sure a lot of autopilots work on the same princible. I would be nice of the control can be done with the Pypilot via NMEA.

Regards,

Bram

[rgleason]

My RM control for wheelpilot 4000 has nmea0183 in for fluxgate compass and perhaps gps. I have no idea what nmea out would be other than all the nmea sentences on the ST bus.

[seandepagnier]

Originally Posted by verkerkbr

Dear BA, thanks for the reply.

My Tillerpilot is a Simrad TP30. I have never used the NMEA183 connector.

Any idea what instructions this tillerpilot get via NMEA. Is it simply a bit left or right movement of the ram ?

What NMEA sentences use such an autopilot ?

I'am sure a lot of autopilots work on the same princible. I would be nice of the control can be done with the Pypilot via NMEA.

Regards,

Bram

As I said, it is not possible with standard nmea sentences. If you can figure out a way to manually control it over nmea let me know. Normally the nmea is used only for wind input, or for autopilot route following.

You would need a different controller, and then wire it directly to the dc motor and bypass all the obsolete electronics.

I have done this with success on:
simrad tp22
autohelm 1000
tillermaster
navico 1600
other units I built from 3d printer screw drives and dc motor.
wheel drive units and

[verkerkbr]

Originally Posted by boat_alexandra

As I said, it is not possible with standard nmea sentences. If you can figure out a way to manually control it over nmea let me know. Normally the nmea is used only for wind input, or for autopilot route following.

You would need a different controller, and then wire it directly to the dc motor and bypass all the obsolete electronics.

I have done this with success on:
simrad tp22
autohelm 1000
tillermaster
navico 1600
other units I built from 3d printer screw drives and dc motor.
wheel drive units and

Dear BA,

according to the service manual the motor of the Simrad TP30 has only two wires.

Is the steering simply a matter of switching the polarity ?

What are the dimensions of the controller board ?

Can the board placed inside the tillerpilot, instead of the present controller board.

How is the controller board connected to the RPI ? It needs 12 volt and wires for the commands instructions.




Regards,




Bram

[seandepagnier]

Originally Posted by verkerkbr

Dear BA,

according to the service manual the motor of the Simrad TP30 has only two wires.

Is the steering simply a matter of switching the polarity ?

If you can power the two wires and the ram moves in, and reversing them, the motor moves out, then it will work. Otherwise you must open the tillerpilot and run the wires directly to the dc electric motor which is simple to do, bypassing all the electronics inside.

You would need to do this for the tp32, but I never heard of a tp30.

What are the dimensions of the controller board ?

Can the board placed inside the tillerpilot, instead of the present controller board.

It fits easily in 85x55x38mm box. It can fit inside the tillerpilot but I don't recommend it unless you also integrated rudder feedback or end of travel switches which are optional.

How is the controller board connected to the RPI ? It needs 12 volt and wires for the commands instructions.

The controller needs 12 volt power on larger wires. I recommend 14 gauge for efficiency and reliability, though they could be smaller. Then it also needs 4 smaller wires of 18-22 gauge although I got away with using phone cables (24 gauge) for up to a few meters. These carry the serial port, of 3.3v commands from the raspberry computer to the motor controllery. They are optically isolated from the motor controller so the ground is maybe not shared, and the 3.3v power is needed this is why it is 4 wires. You can use ttl to rs232 converters on each end if you want to make runs longer, but it's not needed for 3-5 meters.

I don't really recommend putting the controller inside of the tiller pilot. It is simpler to just run the two motor wires to the controller which is nearby the autopilot computer, but it's all up to you how to set things up, and it can work either way.

[verkerkbr]

Ok, it is clear how to rig the cables. But one question is left. How are the end stops of the tillerpilot done ?


For the RPI I only need the controller board ?


Regards,




Bram

[seandepagnier]

The ram detects end of travel from from current measurement. This rarely happens in normal autopilot operation, but it works really well when it's needed since the current draw is much higher when the motor stalls. It reacts in 1/20th of a second and won't continue in that direction anymore until it is commanded to move back a certain amount.

Some drive units are difficult to detect this, if they do not have a significant stall current above normal drive current, but tiller pilots and other drive units do generally work well. You can also rig in optional end of travel switches, or rudder feedback via a potentiometer as well to prevent end of travel before the current detection.. these are optional.

[sailorchic34]

Well I decided to use Sean's Pypilot magic to make a new autopilot for the Rose. I'm using a off the shelf, 10Amp H-bridge pwm motor controller (30 amp intermittent).

I was going to write the code in C as that's what I knew, but decided to do a quick review of Python and found that Sean already had the hard wired code included in one of his Pypilot files. Soooo Sweet.

Also found the GPIO pins for the buttons so it's a bit of soldering and it will get done rather quickly.

Really impressed with pypilot, it's quite a nice bit of coding. It's probably better then any other marine autopilot out there for not a lot of money.

This is to replace my still working really old Autohelm 3000 (analog version). It works ok, but only when the weather is nice. It's too slow when the seas are up.

Initially I will be using the autohelm motor / wheel drive connected to the H-bridge to work the rudder. I may add a linear drive later as well as the rudder position indicator.

The linear drive needs to be a ball screw type with low back drive. That is when the motor is unpowered there is low resistance which allows the wheel to turn by hand. That is how ray marine does their linear drives.

I'm going to make an openplotter too.

[rgleason]

sailorchic, Sean may respond more knowledgeably, but did you see his use of a windshield wiper motor? There is a video on the Opencpn.org website, Opencpn Manual then find Plugins and finally Pypilot wiki In that write up I believe there is a link to a video showing pypilot working with a windshield wiper motor. They maybe a bit stronger than the ball screw type, depending on what you get. It's worth asking Sean about it before starting down that road.


Please keep us informed and add some photos as you build this. Perhaps we can add your experience to the wiki.

[sailorchic34]

Well for the beginning I'll just use the old autohelm motor as it's already mounted to the wheel and boat and works well enough.

The Ball Screw is a type of linear drive to drive the quadrant below deck. That would be a way in the future modification. A wiper motor would work, as well as some gear reduction motors.

I know C and did some Z80 control stuff back in the 90's, just for fun. Sean's pypilot is a very nicely done bit of software.

I was going to roll my own code, but a quick 3 day study of python and I can see where using Sean's code would save a ton of coding. I'm using a different H bridge with 2 control pins, but that's an easy add to the pypilot servo code.

Just need to do a bit of prototyping and soldering. I'm thinking of added a beep board so that keypresses get an audio feedback.

Parts are ordered. Alas it's hard to get UPS to deliver to an anchorage. It's my august project anyway.