Hydraulic wheel steering

[Nee-Wom]

I'm planning a Pypilot installation on my 41' ketch with hydraulic wheel steering. Willing to do without the complexities associated with installing an electric pump for that matter, I'm contemplating a wheel contraption based on a brushed motor for which I have two possible scenarios. The first one is the classical and proven "wheel within the wheel", the second one is to drive directly the shaft of the manual hydraulic pump in the console (which is coupled to the steering wheel). I have two questions:    

1. In any scenario, is it really necessary to disconnect the driven shaft (be it the wheel's or the pump's) from the motor? As I understand it, a brushed motor will generate electricity when its axle is turned. Where will this electricity go when the autopilot is not in use and the helmsman is happily steering away? In other words, when the motor controller is turned off, can the current generated by the wheel being manually operated damage Pypilot? I would very much prefer not to install a clutch mechanism and have the link be permanent between the motor and the shaft. 

2. In the second scenario, where the motor directly drives the pump's shaft, the purchase arrangement would be from smaller gears than in the first scenario. Wouldn't then the steering wheel, as dead weight being actuated, act as a flywheel and somehow affect the proper functioning of Pypilot? In this case it would take something to engage/disengage the wheel itself. 

Any thoughts?

[seandepagnier]

The issue with backdriving the motor is not so much power generated (which will actually flow into the battery if spun fast enough) but additional friction. The extra friction to spin the motor is not good when manually steering, and in extreme cases maybe it could even damage something although I have no reports of this.


Generally with a hydraulic arrangement there is a bypass valve. So when the pump is pushing the rudder, it does not spin the wheel. This is with a hydraulic pump. You are suggesting spinning the same wheel somehow, which is the same as the typical belt arrangement. Yes there is some inertia of the wheel, however it is generally not much compared to say the weight of the rudder. It really all depends on your system. You can expect the pilot to work if it can move the rudder in both directions by reversing polarity, but the less backlash, friction and inertia the better it will work meaning less power consumption and faster reactions which will allow it to work in more adverse conditions.

[Nee-Wom]

My hydraulic steering system does have check valves that effectively block the rudder when the steering wheel is not moving, great for not tending the wheel all the time but no physical rudder feedback. I still have to install a bypass valve – if I ever need one, in the case of a windvane or a secondary hydraulic pump for the autopilot. I'm just trying to figure out what would be the best minimalist intervention on my boat to have a good Pypilot setup.    

Sean, you said it well: backlash, friction, inertia, those are the enemies. Also, cost, energy requirements and complexity. So, brushed motor permanently linked to the shaft of the manual hydraulic pump appears as a good contender. Inertia is at a minimum. Friction, well, given the radius of the steering wheel and the likely purchase from the pump shaft to the motor, intuitively, I tend to estimate it would be negligeable. Especially considering that the boat will be manually steered only a fraction of the time. Backlash, that one I'm a bit more afraid of.    

The main unknown to me, difficult to test, is the electrical consequences of backdriving the motor. I had the time to talk it over with a friend who's more knowledgeable than me in electrical matters and he said the same thing as you: that the electricity generated will not harm the controller and find its way out (either in the battery or to the ground). But will it though? Probably a mix of ignorance on my part and a slight tendency to obsess over sometimes trivial things is responsible for my present doubt, but I imagine this uneven electricity being generated at the motor (when turning the steering wheel), travelling down the wire, hitting back on the "motor" terminal in the controller, going into the board… Will it really pass through the circuit board and all its components, and come out at the board's "Power" terminal without any possible damage? Isn't there something in there that will not readily agree (diodes, capacitors, etc.)?

[seandepagnier]

It will go back into the battery however, it will flow via the internal diodes of the motor controller mosfets. They are not designed for high currents. Another issue is, if the wire from the controller to the battery failed somehow but the controller is still attached and you turned the wheel very quickly it would have nowhere for the voltage to go and could eventually trip the TVS diode and either damage components or cause a short (brake) making it very difficult to turn the wheel. This is probably really difficult to actually do.


You could potentially also add protection circuitry to dissipate the power.. but this is very non-standard so I would leave it up to you to figure out.

Anyway, I suggest you consider the friction of backdriving the motor. Unless you have a gearless solution (which is very unlikely as the electric motor would have to be huge) the added friction will be a real problem to overcome in manual steering. So I suggest you reconsider a way to engage/disengage the motor from the wheel.

[Nee-Wom]

Anyway,  I suggest you consider the friction of backdriving the motor.   Unless you have a gearless solution (which is very unlikely as the electric motor would have to be huge) the added friction will be a real problem to overcome in manual steering.   So I suggest you reconsider a way to engage/disengage the motor from the wheel.

Absolutely right. I just pulled a windshield wiper motor from a 2010 Ford 150 and friction is indeed quite high. I severely underestimated the effect of the gearbox. A permanently linked motor wouldn't be wise, thank you for pointing it out. 

So what now? A dog&spline clutch seems appropriate, but it will have to be machined for the purpose. I wouldn't know where to find one ready made.

Or something else?

Also, an afterthought: the wiring on that wiper motor allows for different motor speeds. Is there anything to tap here? Faster speed in heavy weather but slower when its nice, in order to save energy?