DIY Autopilot/controls for Smaller/auxiliary Outboards
Hope all is well.
I know there is already a thread on Autopilots. But I wanted to start one specifically for smaller auxiliary motors or trolling motors for you fishing types :-)
I'm thinking in the 20HP and under range.
There is an excellent product already on the market for sometime called the TR1 Autopilot that Garmin now sells. Many people I know have been extremely happy with this autopilot.
That being said, I'd like to start a project of building a similar product and get it to be an opensource community developed project. I think a project like this would be fun and provide an opportunity to build something that the opensource community can tweak and use for potentially other projects.
I know there is the Ardupilot, Openpilot and other projects that seem to be geared towards RC based applications; which might work for this. But one of the fundamental challenges with any of these is starting with the hardware and actuators.
The commercially available product, the TR1, uses a hydraulic ram with a stroke of 2". It has a hydraulic pump and of course a flux gate compass and controller. Additionally, it has a sensor attached to the spark plug to read the RPMs of the engine.
Given the cost of the hydraulic components, I'd like to look at using a 12V DC Linear Actuator that is (waterproof) IP65 or higher rating. Such as this
Additionally, for position feedback, these Linear Actuators can be ordered with a potentiometer embedded to provide fairly accurate position.
The throttle control can be a servo motor tied to the carburetor linkage (such as used on the Troll Master products).
As far as controllers...well I'm thinking either a Ardunio based board or potentially an ARM processor board with embedded Linux (such as the Beagleboard, Beaglebone
It just depends on what all makes sense from an expandability perspective.
I have a 9.9HP Mercury Bigfoot I'd like to try to build this on.
But first, I'm curious to see if anyone else has tried to do this?
The tiller Autopilots I see out there don't really work well on Aluminum fishing boat type of applications where the 'kicker'/trolling motor is mounted off the stern with tight space (sometimes with a bracket attached to the transom). That is why the TR1 has been so popular. It mounts the hydraulic ram to the actual bracket/swivel point of the engine and a proprietary bracket attached to the lower vibration isolation mounting shroud (above the lower unit).
For example, here is a mount on a Mercury 9.9 Bigfoot 4stroke outboard: https://picasaweb.google.com/lh/phot...eat=directlink
As you can see, the value of having a actuator attached in this location provides several things:
1. You don't have to have a device attached and in the way of the tiller.
2. The actuator stroke (distance from fully retracted to extended) doesn't have to be a long as something attached to the tiller handle.
3. Allows you to easy trim, tilt / raise the motor without having to disconnect a tiller handle based autopilot. The bracket is mounted to the swivel point of the engine and moves up/down with the engine (such as when you trailer the boat and want the motor up or you're running your big (Main) motor(s) and you want the trolling motor up).
With this design, there is indeed some benefits. But just brainstorming here...the hydraulic ram has the support of allowing the user (captain) to manually steer the motor if he deactivates the Autopilot. I believe there is a bypass valve to allow for easy manual steering. If I go with a DC Linear Actuator, the static holding pressure is pretty high and I have not seen any 'cheaper' models that have a clutch system to provide a similar feature. I think you'd have to use pins to remove the Linear actuator when you wanted to steer manually. Not sure how practical this is - or if someone has a workaround for this.
But this is one thing that comes to mind.
Additionally, where this is located, do you think that an IP65 rated actuator will be adequate? Its not going to be submerged in water, but it will get splashed with river/ocean water and of course rain water as its exposed to the elements.
At any rate, I've been at the brainstorming phase of this project.
I want to do this in two phases:
Phase I: Physically mounting and just using a simple rocker switch control to steer port/starboard and a simple PWM control for the throttle.
The reason for this, why work on the autopilot navigation piece before I get the physical mounting, water intrusion concerns, bracket design, etc.etc all figured out.
With phase one, I can spend some time on the water testing general behavior, steering, throttle and make sure that the actuator and the servos purchased are going to work. If they don't then I need to go back to the drawing board and investigate other designs (maybe going back to using a hydraulic which is going to drive the cost up).
If it does work well, then move on to phase 2.
Phase 2: Incorporate a controller (again either Ardunio or a embedded Linux board) to provide control (such as PWM which can be used to change the speed of the actuator). Also using the serial & general I/O ports, I can add a fluxgate compass, and take NMEA 0183 feeds from the GPS.
Phase 2 would be simple 'stay on heading'.
Phase 3 - take waypoints provided by GPS for full auto nav functions.
Phase 4 - Trolling pattern options (such as zig zag'ing, etc)
Personally, I'd be very happy just to get to phase 1 & phase 2. Phase 3 & 4 would be cool...but as they say, you have to walk before you can run.
So what do you think? Is there anyone out there that has tried this?
Do you think the Linear Actuators will work? Is there another manufacture that provides a better IP67 rated Actuator that is reasonably priced?
I'm leaning towards the Beaglebone because I already have been working on other projects using Beaglebone and it comes with the Angstrom Distribution of Linux which supports perl and python right out of the box. I'm writing other projects in python...so I was thinking that I'd just use Python for all of the control software.
The benefit of using the Beaglebone is expandability and if I wanted to really geek out later, you can easily add a touch screen or other items (heck its got a USB interface and controller built in including an ethernet interface).
Like I said, I'm just in the brainstorming phase of this so if you have any insight; would like to be involved in this too, please let me know.
Last edited by jeb : 03-01-2012 at 01:04 PM. Reason: added clarifying words.
Never thought about it. Can you use the standard Baystar type hydraulic steering ram system ? Use an electric servo motor to turn the helm pump to supply oil pressure ?
Possibly. How do you get position feedback? That is the part that seems a little tricky. Plus what is the cost on that? The Linear Actuator is around $150 and includes a built-in potentiometer for position feedback. I haven't looked at the price of the Baystar hydraulic steering. I'll do a little Googling to see.
I thought about even using the trim tab actuators - but I'm not sure of the duty cycle on those or the pump. Steering would be used constantly near 100% duty cycle. I'm not sure if trim tab motors can handle running that long.
If the Linear Actuators can withstand constant splashing of water, then it looks like the cheapest option. The only probably is when you want to manually override and manually steer from the tiller. They have a static hold pressure of around 500LBs. So you'd have to use pull pins to remove actuator to manually steer; unless someone knows of an actuator that has an override or clutch system. That would be the ticket if they made one.
Somehow you have to steer..turn the motor and this mechanism must be compact and attractive. Even a outboard push pull system could work. An electric acuator could push pull the steering cable. This actuator could be located out of the weather in a steering console. How to sense rudder angle is a good question.
If the electric actuator will drive fast enough, then you could just use the wheel (or tiller) to provide a potentiometer feed to your autopilot. Effectively going full fly-by-wire. Some force-feedback in there would be useful in that case; and I'm not sure how that would work.
Keep the mechanical design as simple as possible. Some electric actuators will give a position feed and that will be a godsend; it saves you so much hassle. IP65 rating should be adequate, but be careful of the materials, as corrosion will be more of a problem than water ingress (IP65 is tested at 1 meter immersion for 30 mins).
For software, I have code for a PWM servo driver in the openPilot project, which might be useful, likewise, code for parsing NMEA data and various other stuff. Since you'll be doing a fair bit of low-level interfacing I'd avoid python and similar languages, and use C or C++, which are more often used for this.
Having an autopilot system in openPilot would be great, and it provides a lot of the framework you'll need. It does require a slightly more powerful computer, though.
Open Source Marine Charting - openpilot.sourceforge.net
Open Source Vessel Dynamics opendynamics.engineering.selfip.org
Do you know if anyone has compiled openpilot for ARM processors? The reason I ask is, I was looking at using the Beaglebone potentially since its a full release of Linux and the expandability would be very easy. I can feed the potentiometer position feedback from the actuator directly to the the analog inputs on the Beaglebone. I then can easily use those readings for position. At that point, if openpilot already has some support for this, it would seem to be an easy tweak to pull from the analog inputs.
Also, what about error deviation on heading? Does openpilot already have some algorithms defined to deal with getting whipped around by wind, current ,etc?
I've run bits of openPilot on a TS7550 ARM SBC, so it will work, but you need to compile it in an emulator as the QT libs it depends on are big. For ADC support, there is code for a MAX186 on a parallel port, but other port and chip drivers could be written easily enough.
Openpilot currently has no "autopilot" capability, but it has all the "surrounding" code and architecture. So no, at the moment, heading is treated as heading, with no filtering. If you wanted filtering, you could copy the demux widget and make a "filter" widget. Then just connect your feed to the filter input, and the filter output to the autopilot.
Open Source Marine Charting - openpilot.sourceforge.net
Open Source Vessel Dynamics opendynamics.engineering.selfip.org
I looked into this for controlling my 26ft sailboat, but never built it.
Anything electronic tends to have a high failure rate on boats.
At the very least, as you said, you must have a system that is fail-safe, i.e. allows easy and quick manual control with the failure of the electronics. You want to leave the electronics in a safer place that on the motor, hydraulic makes more sense I'd think.
Haven't been around for a while, but I've just come over from 'that other thread' on autopilots, and to be honest my eyes glazed over after reading one or two pages there. Why replicate an already existing commercial system ? To my mind it makes no sense whatsoever. If you can't afford one, then buy s/h.
Let's start afresh, from a blank sheet of paper ...
What is the most expensive, and the most problem-prone part of any autopilot ? Surely, the linear actuator. I've forgotten how many times I've heard people complain of their Tiller-Pilot packing-up in mid-ocean - and it's always that part of the kit that fails. It's vulnerable, and has a very hard life.
So, is there another way ? YES.
Remember Wind-Vane Self-Steering ? Especially the Swinging Pendulum sort ?
As you know, the swinging pendulum generates massive amounts of power, enough to steer very large boats, and as a bonus, without using any of those precious Amps from your battery.
Remember Walt Murray and his swinging pendulum systems ? For those who don't, Walt used an angled crank to convert a push-pull error signal from the wind-vane to rotate the pendulum oar. And it didn't take very much rotation for the oar to be driven sideways by the force of water passing over it, and then for that movement to be harnessed to steer the boat, invariably by moving the tiller - which of course could equally be a rudder tiller, or an outboard engine tiller ...
As I see it, one solution then presents itself by disconnecting the wind-vane push-pull rod, and rotating the angled crank by means of a stepper motor instead. I would have thought that the size of motor as used in printer paper-feeds should be about right. Of course it would need weather-proofing, but spares would be absurdly cheap, and several spare motors could always be carried.
So - for those with sailboats - build the full Walt Murray wind vane gear, and disconnect the crank as required when motoring under autopilot. And for the motorised fraternity with their small fishing boats (like me), simply build the lower half of the system, i.e. without the mast and vane.
I'd bet money that there's the kernel of a workable idea here - so now it's over to you blokes ...
Any progress on this? I've been looking into becoming active with the arduino projects and am currently looking for an autopilot for my Merc 9.9 FS as well. The TR-1 autopilot looks like a great project, but at 2400+, it's warranting a bit of research into DIY solutions.
I'd love to see what you've come up with so far.
I have built an AP from the ground up and it is now on several sail boats. It sounds like your interest is in the actual mechanical drive, replacing the hydraulics with electrics. If that is the case you can easily purchase the control electronics from several suppliers and then experiment with various drive solutions, just base them all on 12V motors. They will interface readily to the motor output of any one of the AP controllers out there.
You don't need to monitor rudder position if you choose a system that monitors the motor current. The Ray Marine systems do this and most certainly so does NKE and B&G, and likely the other big players. Since your application is on an outboard the RM X5 controller might be a really good fit as you need not worry about a clutch arrangement.
If you are rolling your own electronics than the compass and creating electronics that handles the dynamics of same in a real world sea way is the challenge. ONce that is tackled the challenge will be feedback algorithms that handle the engine and hull dynamics without going out of control.
Very interested in this thread.
I actually believe I have many of the components that may make up a system like this.
On my boat today, I have a bracket made by an outlet called Remote Troll. This unit combines a trolling motor mounting bracket that has a pivoting faceplate controlled by a 12v linear actuator (see http://remotetroll.com/index.php/models). The unit has a wireless option which I don't have -- I simply use the wired remote which comes with a spring loaded toggle that steers the boat left and right.
I've also got a Lowrance HDS-10 Chartplotter unit that talks NMEA 2000 and 0183. It has the built in GPS and obviously the ability to plot a course.
What I would need to complete a DIY autopilot would be a device that can connect to the 12V actuator and automate the steering (left right), plus take waypoint details from the chartplotter and adjust the boat's heading. Today, that middle part is me watching the GPS and steering with my thumb.
Can anyone point me in the right direction? I've read over the Ardupilot site but just can't figure out how the unit would connect to the linear actuator and how I'd take over manual control.
|Thread Tools||Search this Thread|
|Thread||Thread Starter||Forum||Replies||Last Post|
|homemade autopilot||bertho||OnBoard Electronics & Controls||326||Yesterday 04:17 PM|
|Smaller Roger Simpson designs||jamez||Multihulls||3||10-19-2009 01:49 AM|
|Parts for AP45 Autopilot||stuartyh||Services & Employment||0||08-19-2008 05:21 PM|
|smaller dimension sizes Rhino, help Pl.||lazeyjack||Software||5||08-08-2007 03:57 PM|
|electric controls for outboards||naturewaterboy||OnBoard Electronics & Controls||5||07-01-2007 01:16 AM|