Written by Ken Monday, 03 November 2008 00:54

ROV chronicles: Part 5

Motors and controllers

An ROV needs to be able to move around under water, and for that we need propellers (often called thrusters) which in turn need motors to drive them. Choice of motor, and how we are to control them, is key to the design of an ROV. Getting the motors and controllers working is also the last major design problem before we begin construction, and its been causing a few headaches at the DinkyKitty labs!

 

Types of propulsion – hydraulic

Hydraulic propulsion is popular on larger, high powered work class ROVs. A large electric motor powers a hydraulic pump which maintains pressure through a series of pipes, distributing hydraulic fluid to a number of thrusters, and normally also to manipulator arms and other tools (such as drills, water jet nozzles and saws).

 

The advantage of a hydraulic system is that it can deliver a lot of power – a single large electric motor, running at a constant speed (and thus requiring little in the way of electronics) can deliver several hundred horsepower which can easily be distributed to thrusters and manipulators. This hydraulic pressure is controlled using small electrically operated valves, which need very small amounts of power, making the electronic side of the ROV relatively simple.

 

The main drawback of hydraulic power is its weight and size – in addition to the thrusters, a hydraulic system needs a large motor, pump and a reservoir of hydraulic fluid (compensated for outside pressure). The other drawback of hydraulic power is that a single failure anywhere in the system usually means that the hydraulic fluid leaks out, crippling all the thrusters and manipulators at the same time. In contrast, failure of a single electric thrusters motor makes steering hard, but it’s still possible to maneuver. For this reason, many of the latest generation of deep ROVs are using electric thrusters instead of hydraulic because of improved reliability.

 

Brushed DC electric motors

Most smaller observation class ROVs using direct current electricity (often in the 500v to 1000v range) supplied by a surface transformer. The most common DC electric motor is the brushed motor. As the motor spins, two or more contacts (“brushes”) make contact with electromagnet coils, energising them and creating a magnetic field which works against permanent magnets. This causes a central spindle (or rotor) to move – as it turns, the brushes lose contact with one coil and make contact with another, energising a different coil and continuing the motion.

 

These motors are easy to build, and also easy to control. The more electricity you put into the motor, the more power it will delivery, so you can slow the motor down by reducing the input voltage, or more commonly by switching the voltage on and off and varying the gaps between the pulses.

 

Brushed DC motors have several drawbacks though. Firstly, because the brushes are making and breaking contact, they need to be protected from seawater. This means either pressure proofing the motor (which requires heavy and expensive outer casings) or filling them with oil at ambient pressure (which causes drag and reduces power) – both options require unreliable shaft seals to stop seawater coming in where the propeller is attached.

 

The other major drawback with brushed motors is that as the brushes make and break contact, they cause large amounts of electrical interference. This can cause problems for on board electronic control systems, and also interfere with some of the sensors an ROV carries (such as magnetometers – used for tracking buried pipelines).

 

Brushless DC motors

 

Advances in electronics have made a relatively new kind of motor popular. The brushless DC motor is basically an inside-out brushed motor. Instead of having the coils on the rotor working against fixed magnets, the coils are fixed and the magnets spin. Instead of using brushes to switch the electricity flowing through the coils, electronics do the job instead, supplying a varying (AC) voltage – the faster the current switches, the faster the motor spins.

 

The advantages of a brushless motor are that it can be run “open” to the sea (with a bit of protection from corrosion) since there’s no electrical switching going on, and it doesn’t generate the electrical noise of a brushed motor.

 

The main drawback is the complexity of the electronics needed to control the motors. In order to start, stop and reverse the direction of the motors the control electronics normally need some way of sensing the speed the motor is going at. Some motor speed controllers can control the speed without sensors, but they can have problems at very low speeds.

 

Other motors

Other types of electrical motor exist – universal motors (similar to brushed DC motors but using all electromagnets) can run off both alternating and direct current. Induction motors (such as “squirrel cage” 3-phase motors) work by inducing a current in a conductive material, and working against the magnetic fields that causes.

 

However, these more exotic motors are rarely used except in specialist applications (for instance the squirrel cage motor might be used to power the hydraulic pump on a large ROV).

 

Propelling the DinkyKitty ROV

At first glance, the decision for propulsion of our ROV seemed pretty simple. Hydraulics was too heavy (and the components too expensive) to use on a small prototype. Brushed DC motors cause too much electrical noise and engineering shaft seals for them seemed like a bad idea. The logical choice is brushless DC motors, since its possible to source them cheap from the radio controlled model industry.

 

Unfortunately, getting the motors to work with a computer interface is a whole different issue. Most small brushless DC motors use three-phase power, which is supplied by a dedicated electronic speed controller, or ESC. The speed controller takes in DC power (normally supplied by a battery pack, in the case of radio controlled models), along with a pulse width modulation (PWM) signal (normally generated by a radio control receiver). The ESC then converts these into three sine waves, out of step with each other by 120 degrees, and sends these down three wires to the coils in the motor.

 

The first problem we have encountered is that the commercially available ESCs are rather too smart when it comes to power – they detect battery type when first powered on, and attempts to power them off a workbench transformer have so far been unsuccessful. An old laptop battery is standing in as a power supply as a testing measure, but in the long term will need to be replaced by a transformer of some sort.

 

The second problem is that the computer interface board we are using doesn’t generate pulse width modulation signals at the right frequency. The pulses coming out of our board are at least 50% too close together and too short, and therefore the speed controllers are responding to them.

 

Possible solutions

In order to get the computer interface talking to the speed controller, we are first going to attempt to design and build a PWM generator. This will take an analogue signal from the USB interface board, and generate the correct frequency and timing of PWM signals to send to the speed controller.

 

If a PWM signal generator doesn’t work, it may be necessary to build a speed controller from scratch. In fact, this may be necessary anyway, if the commercial speed controllers prove impossible to power off a transformer. It may also be desirable to build a speed controller than uses sensors to detect the motor speed, to give better control at low speeds.

 

Both the PWM signal generator and the speed controller itself are fairly mature solutions – there are ready-made integrated circuits that do most of the hard work, and schematics are readily available. Hopefully we’ll have something positive to report within a week or so, since this represents the last major design hurdle for the ROV – once the motor and control situation is finalised, parts can be ordered and construction can begin.

 

Next time:  ROV safety systems