First, let me explain that I know very little of electronics. Thus, please, forgive me if I am asking something very obvious or abstract. I have tried to Google this but I am obviously not using the right words as I find very little.

The problem: I have an auto-pilot that is supposed to control a hydraulic pump for a ram which will actuate a rudder. Well, I would like to use an electric 12 vDC brushed motor to control such rudder instead of the ram. The motor has a reduction and is connected directly to the output power for the hydraulic pump, which is 12 vDC, up to 25 amp and reverses the polarity when it needs to actuate in different directions. It works more or less and I would like to get rid of the 'less'.

Is it possible to have an interface between the output of the autopilot and the motor? ...so that:

  1. It fades the voltage of the output. When the autopilot decides to apply power to the motor, it does it straight away. I would like to have some dumping on it so that the motor starts and stops smoothly. I have put some capacitors attached to the motor but am not sure is doing much; probably just absorbing peaks and filtering noise?

  2. It brakes the motor when there is no voltage applied (by shorting it?) So, when the autopilot does think that the rudder is in the right position and does not provide power to the motor, the motor cannot spin freely.

In short, the power is 12VDC and reverses on demand of the autopilot. The motor is 12 VDC and there is no need to alter voltage or amp in the power.

I have seen many motor drivers but I believe there must be something simpler? This one can do what I want but I would then need to transform the -12/+12 output from the autopilot to 0/2.5/5 in order to use the analogue input of the driver. I am sure it must be simpler than this.

I obviously appreciate very much a full solution but I am very much willing to do my homework. Please, any advice or heading so that I can research further in the right direction?

Edit: It seems I didn't explain well. The autopilot is a controller and is fed by many sensors -heading, rate of turn, wind, rudder angle, GPS. It basically can follow a track or maintain a heading compensating for any forces. The problem is that it controls the rudder with a +/- 12 VDC, which is actually activating a DC motor is both directions. What I would like is to improve that signal and brake the motor when there is no current out of the autopilot, by shorting the motor I guess, and to apply those +/- 12VDC progressively, let's say with a period of 1/4 of a second, so that the motor starts and stops smoothly.


For a soft start. Could I use a simple circuit with a resistor, a capacitor and a MOSFET? Initially, the MOSFET is off and the current goes through the resistor, once the capacitor is charged, the MOSFET turns on and the resistor is bypassed.

And for braking it. Could I use a high power resistor in parallel with the motor controlled by a MOSFET?

  • \$\begingroup\$ One caution : hydraulics are often used because they can generate high forces and withstand high external forces : have you made sure your motor and reduction are capable of the same? \$\endgroup\$ – Brian Drummond Dec 19 '13 at 20:27
  • \$\begingroup\$ @Brian -Thanks Brian. Yes, the Motor has a gearbox and it is driving the rudder by a chain and sprocket. It works well apart from the desired improvements... \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:12
  • \$\begingroup\$ Just to mention that the motor has also an electromagnetic clutch controlled by the autopilot \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:29

You first have to decide what signal exactly the autopilot is producing. Is it something that says go more right or go more left, or is it saying go to this rudder position? Note that one is the derivative of the other.

In either case, it sounds like you need a real control system, not just something that applies power to the motor sometimes. You should probably implement a rudder position controller, then either use the signal from the autopilot directly or integrate it, depending on what it actually is.

The position controller will apply whatever current to the motor to maintain the required position. If it doesn't take much torque, then little current will be applied. If the rudder is at one end of travel and there is a lot of back force on it, then the system will automatically apply more torque. The point is, what you really care about ultimately is rudder position. So control that, then use whatever signal you get to adjust the desired rudder position.

On a separate topic, someone who "knows very little about electronics" shouldn't be messing with a flight control system. You have no business being in there. If you want to kill youself, that's your business. But it's my business if you want to do it in a airplane because I might be underneath when it comes crashing down.

| improve this answer | |
  • 2
    \$\begingroup\$ Who said airplane? Maybe it's a ship. Also, are you aware there are entire clubs dedicated to people who know little to nothing about electronics flying airplanes they built themselves? You better wear a helmet! \$\endgroup\$ – Phil Frost Dec 19 '13 at 20:01
  • 1
    \$\begingroup\$ Indeed, it is a vessel; it is not going to fall on top of anybody. There is a control system: the autopilot. It is complemented by many sensors -heading, rate of turn, actual angle of the rudder, wind and GPS. The autopilot decides what angle must set to the rudder, according to a given track or heading and the forces that are affecting the vessel. Kalman filters process the input and provide the course to steer. The autopilot activates the rudder by some degrees sending a voltage signal to the rudder drive until the desired angle is achieved and then it just stops the voltage. \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:31
  • \$\begingroup\$ It works reasonably well apart from the desired improvements. \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:32
  • \$\begingroup\$ No matter how precisely we engineer it , if it hit a undiscovered flaw like 'subsynchronous resonance' like what had happened in Nevada power station, things fail and disaster could happen. I like the idea of helmet. More than that a complete spacesuit or something. \$\endgroup\$ – Standard Sandun Jan 19 '14 at 12:25

I would then need to transform the -12/+12 output from the autopilot to 0/2.5/5 in order to use the analogue input of the driver. I am sure it must be simpler than this.

Nope, that's what it appears you'll have to do. You'll need a small circuit interfacing the +/-12V signal and producing an output that is 2.5V+/-2.5V. This is probably best not done with resistors alone - use a rail-to-rail op-amp (possibly AD8605) powered from the auxiliary 5V supplied by the proposed motor controller and devise a circuit that takes the +/-12V from the autopilot and produces the required output.

Alternatively you can use a couple of small 12V relays - one driven by the positive 12V and one driven when it goes negative to do the business. This might be easier if you are not confident about using op-amps.

| improve this answer | |
  • \$\begingroup\$ -Hi Andy, thank you for your input. I'll certainly have a look at this route. I thought it would be easier to just smooth the output of the autopilot and short the motor when no power is present. Using the driver seemed too long a chain of devices. \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:15
  • \$\begingroup\$ @user1266225 I thought that using your proposed method caused problems (according to your question). Maybe I misunderstood? \$\endgroup\$ – Andy aka Dec 19 '13 at 23:20
  • \$\begingroup\$ Well, the autopilot sends a +12 and the motor turns in one direction; it sends a -12 and it turns in the opposite direction. I just would like the motor to be shorted when there is no current out of the autopilot and, if possible, to apply the + or - 12 progressively. \$\endgroup\$ – Aznarepse Dec 20 '13 at 10:31
  • \$\begingroup\$ Can I use PWM to smooth the signal and a transistor to do the shorting? \$\endgroup\$ – Aznarepse Dec 20 '13 at 10:48
  • \$\begingroup\$ @user1266225 That's probably not a bad idea but I'd use a relay for the shorting because it's probably simpler. I'd also consider fusing the relay circuit or having a low resistance high power resistor in series to limit the peak current from damaging the contacts. \$\endgroup\$ – Andy aka Dec 20 '13 at 10:54

What you seem to be describing is a servomotor. You need a goal (rudder at 5 degrees) and a way to measure the current state (rudder at 2 degrees) so you can calculate the error (3 degrees). Then through some algorithm, frequently a PID controller, generates control inputs (motor voltage) based on the error.

You can make a servomotor yourself with an ordinary motor and a microcontroller or discrete electronics. Or, you can buy motors with the sensors and feedback mechanism included.

Another option without feedback is a stepper motor. These are motors that move in predictable, discrete steps. You know where the rudder is based on knowing where it was in the past, and how many steps you have commanded it to move sense them.

| improve this answer | |
  • \$\begingroup\$ I didn't explain myself well... Yes, there is a sensor that is referencing the angle of the rudder continuously to the autopilot. There is also a heading and rate of turn sensor feeding data and a GPS and wind sensor. Thus, when the autopilot sets a course, activates the rudder n degrees by running the rudder drive, a motor in this case, until the desired angle is reached. \$\endgroup\$ – Aznarepse Dec 19 '13 at 23:19
  • \$\begingroup\$ @user1266225 so then, what's your problem? \$\endgroup\$ – Phil Frost Dec 20 '13 at 1:44
  • \$\begingroup\$ -The problem is that the signal sent by the autopilot is a +/- 12 VDC and that is all. I need the motor to be braked when the autopilot does not send power. Basically, short the motor is no current is detected. Also, it would be very good to have the +/- 12VDC progressively applied so that the motor is smooth in starting and stopping. \$\endgroup\$ – Aznarepse Dec 20 '13 at 10:25
  • \$\begingroup\$ @user1266225 right...that's what the PID controller does, only it does it better. \$\endgroup\$ – Phil Frost Dec 20 '13 at 13:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.