Pypilot working on my Nicholson 44

[xfactor99]

Hi,

just want to report my setup. We did the first testing at sea today ....... happy ....

Boat
Nicholson 44 from 1977,

Former Autopilot:
Robertson/Simrad x300, with two Controls/Displays. The LCD's where getting bad ....
(I guess before that there was a Neco-Autopilot/Curse-computer)


Drive:
Neco Marine 17DR8-24Volt
Spocket with chain to the cardanic stearing.
It is a brushed compound-wound DC-Motor with electromagnetic clutch, I guess from the 70's.
It changes diriection by reversing the polarity to the amature and keeping the polaity to the series- and shunt-coil.

Motor Driver
Chinese Arduino-Nano clone, close to the Raspberry Pi with Pypilot Motor.ino software and a Pololu G2 1367
Motor-Driver inside the Neco-Drive (former watertight Relay-Box on top of the drive). They are connected btw. eachother with a shielded cable (only) 3-wires (common GND, RC, Clutch).
The clutch is activated by a 5v/24V , (chinese/relay) the relay gets activeated by the Nano (Pin11) and the Relay is powerd by the Pololu 1367. (I will change this to an SSR-Relay later).
For safety reasons ther is a cable to the helm witch interupts/activates the clutch.
There are two Reed-Contacts (Normaly Open) to tel the limits of the stearing. They go to the Pololu Driver and worke fine.
I tried PWM  and for now I have the setting at Pypilot Min-Speed=50% Max-Speed=100%

Course Computer
Raspberry Pi with Openplotter/Pypilot,  GPS and IMU 9255. Later I will add the windsensor data. 
The system for now is headless. Access is via VNC-Viewer from the Mobile phone or tablet or laptopp.

Sea trials today:
Setting: "Simple Autopilot" and Compass, Gains: I=0.00690 D=0.16750 P=0.00330
On compass mode (under engine) it works good also I have do a lot of more testing on this ...



To thanks to Sean and all who have contributed to this Project.

Andreas

[Sailoog]

Thanks for sharing!!! Do you have any picture?

[xfactor99]

Thanks for sharing!!! Do you have any picture?

I will do some this week and post them .....

[Saqqara]

Great write-up! As our boats and needs are all so different, the more detail we can get about configurations that work, the better!

My own efforts have been delayed as my old Autohelm wheel drive was slipping, and some repairs I attempted didn't work. I ordered a newer Raymarine Wheel Drive on ebay - I just don't know where to start cobbling together a drive system, yet.

[xfactor99]

As prommised here some photos:

first and most important ..... the boat:
   

then the first (working!) setting with simple relay-h-bridge (2xrelay) the two others are for the clutch (engage and safety interupt):
   

Then test with the pololu G2 1367 (top left in the photo) testing with red/green LED and PWM (see Osziloscope)
   

The Neco-Drive with the Pololu and Clutch relay inside the top (I will re-paint the drive .... later ;-) ...) :
   

And at latest a phone-screenshot from the first testsail working arround the gain's:
   

Save sailing to all
Andreas

[xfactor99]

Hi to all

Here are two videos of the pypilot sailing today.
The conditions today were App-Wind 15-25 kt and the Seastate moderate.
The pypilot just works fine as you can see at the ruddermovement, also the waves pushing .....
The power consumtion is dificult to determine for me as I have also the fridge running the solar charging ect.
Overall is not more then the old autoppilot and depends verry much on how acurat you want your coures, your gain setting and
the sail trimming. The solarpannel produced all what was needed so I arrived after 63NM winth full batteries.

Right now I use the Tinypilot img 30.11.2019 on a Rpi3b and keys and nokia display at the helm and also control through OpenCPN-Plugin on an other RPI3b with Openplotter 1.2

The sailing today was on Compass mode and also True-Wind and GPS-Waypoint.
All modes worked realy well.



Video outside:
https://www.dropbox.com/s/mfddyqrqi5n811...2.mp4?dl=0

Video inside:
https://www.dropbox.com/s/ii92u01h2ktfi8...0.mp4?dl=0




I still working on the best gain-setting:
On flat water and light wind I use only D=0,1 and P=0,005  FF=0,73(..... P=D*5%....)
Today I Used: D=0,12 P=0,006 PR=0,006 DD=0,2 FF=0,73

So the conclusion of the day was, the pypilot is doing reay fine and I´m sure it can do still better when I have more understandig of gain setting ;-)

Thanks again to Sean for that great projekt !!!!

Save sailing
Andreas

[seandepagnier]

Overall is not more then the old autoppilot and depends verry much on how acurat you want your coures, your gain setting and
the sail trimming. The solarpannel produced all what was needed so I arrived after 63NM winth full batteries.

pypilot tracks amp hours and also calculates the current average power consumption in watts.     This calculation is not perfectly accurate for a few reasons but close enough to be useful and usually less than other autopilots!

On flat water and light wind I use only D=0,1 and P=0,005  FF=0,73(..... P=D*5%....)

What do you mean P=D*5% ?   Would it be useful to lock it this way or something?   I found they should nearly be locked but not exactly, but using different dimensions that are transformed to these gains could be better.

Today I Used: D=0,12 P=0,006 PR=0,006 DD=0,2 FF=0,73

So the conclusion of the day was, the pypilot is doing reay fine and I´m sure it can do still better when I have more understandig of gain setting ;-)

Maybe it won't do much better than "doing fine" for a basic pilot since it's limited to these gains.

[nick.k]

Hello Andreas very interested to come across this thread, thank you.

I have a 74 Nicholson 48 with a Nico autopilot. A similar setup to yours looking at the photos. The motor and gear unit is fine but something is wrong with the control. Reading through correspondence from previous owners in the boat file the autopilot never really worked satisfactorily so rather than try to fix forty year old electronics the option to build a new control unit sounds good!

On my Nico motor unit the rudder angle feedback and the travel limit switches are mounted on a unit driven by a worm screw attached to the primary socket. This part seems to be a well known weakness of the Nico.
You mentioned "There are two Reed-Contacts (Normaly Open) to tel the limits of the stearing." are these the existing switches on the motor unit or did you make up a new assembly? Does the pilot have a rudder angle sensor or it's not necessary?




Did we meet in Xabia in November last year?

Nick

[xfactor99]

Hello Andreas very interested to come across this thread, thank you.

I have a 74 Nicholson 48 with a Nico autopilot. A similar setup to yours looking at the photos. The motor and gear unit is fine but something is wrong with the control. Reading through correspondence from previous owners in the boat file the autopilot never really worked satisfactorily so rather than try to fix forty year old electronics the option to build a new control unit sounds good!

On my Nico motor unit the rudder angle feedback and the travel limit switches are mounted on a unit driven by a worm screw attached to the primary socket. This part seems to be a well known weakness of the Nico.
You mentioned "There are two Reed-Contacts (Normaly Open) to tel the limits of the stearing." are these the existing switches on the motor unit or did you make up a new assembly? Does the pilot have a rudder angle sensor or it's not necessary?




Did we meet in Xabia in November last year?

Nick

Hi Nick,

nice to hear from you and yes we met in November, I remeber you were on the way north ....

The NECO drive ist a masterpeace as your boat is as well. The E-Motor is a Wound-Compound-DC-Motor (Googele it !) and has a series Coil and a Shunt Coil. This gives the drive some nice charateristict but require special atention when it comes to drive it by an controler. Mine was modified as it was already with the second electronic from Roberston/Simrad. The drive also has an Clutch and is geared down so it has gear oil.

My drive has no end of travel switches so the housing you are speaking in mine is emty. 
I use end of travel switches at the rudder-schaft  .... 
For now I don´t use an rudder-angle sensor but you can use one (you don´t need)
Both right now are just to prevent drive-action when at limit ....
Maybe an an further version of the software they might use the Rudder-Angel for performance of stering or for autolearning.

The Pypilot/Tinypilot is a great project, but not "plug and play" !!!!
If you want to start, you can buy a readdy asembled Tynipilot and Motrocontroler from Sean at "www.pypilot.org" or you
buy an Raspberry Pi3b+, an IMU (Compass/Pitch/Roll) and an Ardino and some LED to make it work.
Once you through that you loock for an Motorcontroler (I have the Pololu G2) To connect the Drive to the electronic.


Take care if you have a 12/24V drive when it comes to the Motorcontroler ...

In all this will be about 150€ and a lot of learning, fun, headdache .....
Or you buy a ready Autopilot electronic to attach to your drive (wich might still need some modifications)

You can run the pypilot alone or in Openplotter.
I use Openplotter and Pypilot/Tinypilot on two diferent Rasperry Pi.

For controling it you can use the OpenCPN-Plugin, the webapp(mobilphone) or make an display with buttens or all together.

Good fun and save sailing
Andreas

[seandepagnier]

How did you drive the shunt coil of the motor? To make it reverse, current must be the same direction through one coil but not the other... Did you use a bridge rectifier, or is the shunt powered through the clutch?