The autopilot control head is an Autohelm 3000. It currently steers the boat via a newer Raymarine 4000 wheel pilot but although the boat is only 10 metres it cannot cope once the sea state gets up a bit. By not cope I mean the belt starts slipping despite being new and correctly tensioned. Plus from my previous experience of these wheel pilots they are poorly made, very weak pieces of kit that will break if overloaded even slightly. I do a lot of single-handed sailing and when the sea gets up is when I need the autopilot to cope and keep some sort of course even if it isn't the perfect course. What I have at the moment lacks the power or robustness to do that and so I am just investigating options for an improvement on the current set up that will not break the bank or cost 1/5th the value of the boat.
My second option is a tiller pilot attached to my wind vane steering which is very low budget but they are also unreliable and it is a bit of a mickey mouse set up that is vulnerable to the weather and is it is a set up that is no good for motoring as the prop wash rattles the wind vane to death.
I may try the Autohelm 3000 and relays option first and if the results are not satisfactory in the course holding department then look for a used autopilot control set up on eBay to replace the Autohelm 3000. It would be silly not to try the relay option first for the price of a couple of relays and some wiring and terminals.
The ram I am considering has a hard over time of 13 seconds and when on minimum 'rudder response' the Autohelm 3000 would take forever to achieve this as it only gives out very short pulses of power but on maximum 'rudder response' it could achieve it very quickly.

 

I have received information from the linear ram manufacturers that has put paid to the plan to use relays:

thank you for your enquiry, unfortunately you can not use relays to provide for a greater power output from your ST3000. The drive output is PWM, that is pulsed 12 volts. It is like this to control the output power and speed of the motor. The longer the pulse the faster the motor goes. The pulses are between 10 and 200 a second depending on the autopilot. Relays would just chatter or simply close and run the motor at full power all the time giving the autopilot no control over the drive. Unfortunately the only solution is to upgrade the autopilot.

 

The MOSFET circuit in post #14 by crutschow would still do the job.

 

This is the Circuit for the Auto pilot. It could be modified to drive Hydraulic solenoids if you used a hydraulic pump & ram system. Which ive seen on plenty of boats. The red X is the port/starb drives for the H bridge.

 

Thank you for pointing that out AlbertHall.

Also thank you debe - but I do not want to go down the hydraulic pump and ram route due to the high electrical demand of a constantly running hydraulic pump.

 

Why not use an electric driven ram, that's what the old Auto helm Tiller pilot used that I had & it was quite strong. You could then just use it in place of the wheel drive. & the original H drive on the auto pilot. Theres heaps of them on Ebay all diferent sizes. I would be picking one with a suitable stroke & push/pull loading. They are cheap enough to carry a spare.

 

Sorry debe but I'm not quite sure I follow what you are saying. Tiller pilots have to be connected and disconnected from the tiller as they have no clutch in them and so are not suitable to be attached to a steering quadrant. I've been searching for electric rams on eBay for some time now and I've not seen anything suitable. The ram would require a clutch so that normal manual steering could happen with it still connected to the steering quadrant. Do you have an example of the 'heaps of them on eBay'?

 

Probably not practical when you have a Quadrant on the rudder shaft, & probably not all that acessable. Ime used to yachts with tiller steering, where the actuator is easily fitted & removed.

 

I'm not a sailor, nor do I have any boat experience. That said, I'm seeing your solution giving you full left rudder or full right rudder. Question in my mind is "How do you return the rudder to zero?" If your auto pilot decides 2˚ left turn is required to stay on track, your relay rig will drive your motor full left rudder and begin turning the boat however hard the boat is capable of turning. Once you reach the desired direction your full left rudder will oversteer you, causing the auto pilot to redirect right rudder to compensate. Your boat will now be in a full right turn until you cross your desired heading and now the auto pilot has to demand full left rudder again. So how are you going to return the rudder to zero position once you're on course? I see your system zig-zagging you across the lake leaving a lot of boaters in stitches watching you try to handle a boat that goes from full left to full right and back again repeatedly. Without feedback your autopilot won't know where the rudder is, only the heading of the boat.

Of course, I could be completely wrong.

 

Autopilots don't go full left then full right - they make small adjustments or big adjustments depending on how far off course you are. The present autopilot controller that I have has no way of knowing what position the rudder is in and simply tries to keep a compass heading using the small movements and bigger ones if needed and then slowly returning the rudder to where it started as you get closer to the desired course. Generally rudder position sensors are used on more complicated systems that try to learn the boats characteristics and improve the autopilot response and also to stop the ram from extending or retracting too far and damaging the steering quadrant with too much force past the rudder stops.

 

OK. As I said, I have no experience. I just had to play devil's advocate.

So you're telling me that if you need a 2˚ course correction the AP will give slight (lets say) right rudder, and then when course is achieved the AP will reverse the rudder position the same amount, returning it to zero? If so - OK. How does your ram compare with the original drive mechanism? If the old motor gave 1/10˚ per second and the ram gives 1˚ per second of movement, your AP is still going to be wildly steering your boat. AGAIN, I have no experience and honestly don't know what I'm talking about. But it's worth considering - I think.

Maybe I'm wrong. If so - - - .

[edit] Hydraulic rams typically have a cylinder with a single rod, meaning only one sealing surface. What that REALLY means is that a 1 inch ram with a half inch rod will have more surface area on one side of the piston than on the other. So 1 second of extend command may result in 5 inches of movement whereas 1 second of retract command may result in 7 inches. I'd be worried your AP will not know this and may extend right rudder 5 inches then THINKING it's retracting 5 inches it could over correct. Then when the boat is off course the AP will again extend 5 inches. Once the boat is on course again, the AP will retract the ram, again something other than 5 inches.

A 1 inch piston has 0.79" square area for the pump to push against. Given a specific flow rate and a specific duration the piston will move that specific distance. With the back side of the piston having a half inch shaft, the surface area for the pump to act against is only 0.77" square area. The piston will move further for the same amount of fluid delivered by the same pump driving for the same period of time. The rudder is not going to return to zero.

The devil's in the details.

 

So what are your thoughts about using a MOSFET H-bridge circuit to drive the actuator since that should work fine with the pulses?

 

Most steering quadrants turn 36 to 40 degrees each side of centre. Normally rams with a total of 8" of movement are used. Whatever the ram length used is it can be mounted at a further radius point from the rudder shaft or closer in - obviously the loads will change due to the change in leverage.The ram I am considering is designed for autopilot use and so should have none of the issues that you have raised. It has a hard over time of 13 seconds which is suitable for a sailing boat and anyway the speed of response can be adjusted on the autopilot if it is found to be making adjustments too slowly. The Autohelm 3000 controller is designed to turn a wheel and not all boats have the same number of turns lock to lock so the 'rudder response' adjustment is how one wheel pilot can be made to work with a boat with 2 turns lock to lock and also on one with three turns lock to lock.

 

So what are your thoughts about using a MOSFET H-bridge circuit to drive the actuator since that should work fine with the pulses?

I like the idea and I have posted an image of your idea to the ram manufacturers for their thoughts. The ram is over £1000 so I want to be sure these ideas have a good chance of working before I order!
I am not really an electronics guru and so I have been trying to understand how your diagram works and how I could turn it into an actual product. Are the part numbers listed on their suitable for the purpose of driving a 20A ram?

 

The autopilot control head is an Autohelm 3000. It currently steers the boat via a newer Raymarine 4000 wheel pilot but although the boat is only 10 metres it cannot cope once the sea state gets up a bit. By not cope I mean the belt starts slipping despite being new and correctly tensioned. Plus from my previous experience of these wheel pilots they are poorly made, very weak pieces of kit that will break if overloaded even slightly. I do a lot of single-handed sailing and when the sea gets up is when I need the autopilot to cope and keep some sort of course even if it isn't the perfect course. What I have at the moment lacks the power or robustness to do that and so I am just investigating options for an improvement on the current set up that will not break the bank or cost 1/5th the value of the boat.
My second option is a tiller pilot attached to my wind vane steering which is very low budget but they are also unreliable and it is a bit of a mickey mouse set up that is vulnerable to the weather and is it is a set up that is no good for motoring as the prop wash rattles the wind vane to death.
I may try the Autohelm 3000 and relays option first and if the results are not satisfactory in the course holding department then look for a used autopilot control set up on eBay to replace the Autohelm 3000. It would be silly not to try the relay option first for the price of a couple of relays and some wiring and terminals.
The ram I am considering has a hard over time of 13 seconds and when on minimum 'rudder response' the Autohelm 3000 would take forever to achieve this as it only gives out very short pulses of power but on maximum 'rudder response' it could achieve it very quickly.

I know the Autohelm 3000 very very well. It is quite old c. 1985/6 . I have the full service manual here , It drove the wheel with a toothed rubber belt like a sewing machine. The ST 4000 wheel drive was a little later and used more or less the same electronics, in fact the control box was almost identical. Success or failure depended a lot on the effort required to drive the rudder.The early type drive motor was a maxon motor which was tiny, only a little bigger than found in a child's toy. The epicyclic reduction gear was a common point of failure, The lever clutch also failed easily if overloaded What might work for you, is if you could drive your ram from a servo system i.e. putting either a position or torque sensor on the wheel and using the ram to reduce the load on the steering wheel - like power steering on a car. Your ST4000 wheel pilot will then be more suited to the lighter load.

 

I know the Autohelm 3000 very very well. It is quite old c. 1985/6 . I have the full service manual here , It drove the wheel with a toothed rubber belt like a sewing machine. The ST 4000 wheel drive was a little later and used more or less the same electronics, in fact the control box was almost identical. Success or failure depended a lot on the effort required to drive the rudder.The early type drive motor was a maxon motor which was tiny, only a little bigger than found in a child's toy. The epicyclic reduction gear was a common point of failure, The lever clutch also failed easily if overloaded What might work for you, is if you could drive your ram from a servo system i.e. putting either a position or torque sensor on the wheel and using the ram to reduce the load on the steering wheel - like power steering on a car. Your ST4000 wheel pilot will then be more suited to the lighter load.

Thank you for your idea but I feel it adds an extra level of complexity - and I feel I might be the mk1 tester for this idea!
I think my boat has gone through the mk1 and mk2 and mk3 wheelpilots and the only part of the original system that has survived is the Autohelm 3000 control head which I think is of early 80's vintage. I would like to just go over to a below deck more powerful ram but all the autopilot manufacturers price their products so high these days that for older boats it is prohibitive. I also prefer the simplicity of the Autohelm 3000 1980's electronics rather than the all singing all dancing modern computer autopilots. It has lasted this long for a reason whereas I know of plenty of £2000 modern autopilot computers on friends boats that have not lasted a year. Hopefully with the help here I might come up with a working solution - if not it is back to hanging on to the wheel and missing out on lunch if the sea gets up!

 

Are the part numbers listed on their suitable for the purpose of driving a 20A ram?

No, they have too high an ON resistance and will dissipate too much power.
If you go with the circuit I can look for some that have a lower resistance.

 

No, they have too high an ON resistance and will dissipate too much power.
If you go with the circuit I can look for some that have a lower resistance.

OK - I thought so as I looked up the specifications for them and it didn't seem quite right. I'll see what the ram manufacturers have to say before progressing any further. This is going to be a winter out of the water project anyway so there is no great rush - I'm just doing the research in advance. It may not go ahead at all if she who controls the purse strings can't be persuaded!

Thanks for your help and knowledge.

 

I've been wondering if there is not something off the shelf already created for the robotics world that would accept a 12v PWM signal to a pair of inputs and convert it to power output left or right to an electric motor?

 

I have had a reply from the ram manufacturers that does not fill me with confidence that they know what they are talking about:

Thank you for your enquiry which was passed to our technical manager, the proposed circuit will not work with your unit, it is a basic "totem pole" circuit taken from a basic electronics for beginners book. It is used to demonstrate in college courses basic motor control. It is not a practical solution.