Good Wiring Practise
Our controllers get used in a wide variety of applications often switching large currents (several hundred amps perhaps) at around 20kHz. This makes the wiring used with them more critical, and this page tries to explain some good wiring practise and things to avoid.
The author spent about 12 years working for a welding equipment company. We did not use high current motor controllers there, but exactly the same principles applied. Indeed, that is where I learnt much of the following. The principles of good layout and wiring practise are pretty universal.
The controller
In the diagram below, the controller draws current from the battery positive, feeding it through the motor and returning it back to the battery negative. It controls the motor current by chopping on and off this battery current at around 20,000 times per second.
Now wires are not perfect conductors. They have resistance and they have inductance. These two properties mean that there is a voltage drop across any wires carrying these currents. The fact that the current is being chopped makes the inductance of the wires concerned particularly important, but whereas resistance is controlled by the length and diameter of the wire, the inductance is controlled by its length and its routing. The wire thickness has little effect on the inductance.
A wire also has other properties: it has an ability to radiate and to receive electromagnetic signals. This is interference reception and generation. You need to keep this to minimum, and the same principles apply to this as to other possible problems.
By way of this radiation/reception ability, the signal in any wire can also interact with the signal in any nearby wire. That is another potential cause of problems.
A page elsewhere on this site discusses PWM motor speed control and the nature of the current and voltage transients that occur.
Battery loop inductance/resistance
The battery loop is a wire from A to B, the battery itself and the wire from C to E. All of these will have resistance and inductance, they will all drop voltage when the current is being switched. However, the controller can cope with this, and the controller cannot tell whether the wire AB is long and CE is short, or whether the CE wire is long and the AB wire is short. All the controller can know about is the combined inductance/resistance of this loop. The page on PWM motor speed control explains why this loop impedance should be generally kept as low as possible overall. The battery loop impedance directly affects the work done by the on-board decoupling capacitors so if it is not low, these may fail unless the controller’s performance (current limit) is reduced.
To minimise the loop inductance, the wires AB and ED should be as close together as possible, preferably twisted (though that’s often impractical). The loop inductance is actually a function of the area inside the loop. Running the positive wire close to the negative minimises this area, minimising the inductance and also minimising any noise transmitted from this loop. Since this loop is carrying high currents, it’s potentially a good generator of noise – which could be picked up in other wiring.
Motor Loop
In the same way that the battery has a loop inductance and a loop area, so too does the motor wiring. So keep motor wiring as a pair, close together, preferably twisted and keep this also away from sensitive wires that might pick up noise.
Common earth impedance
Consider the input signal. In a machine of the type discussed here, this is often from a microprocessor or PLC. The source doesn’t matter. As designed, it’s a pot and is fed into a connector on the controller, shown here as F and G. G is the reference input, usually connected to battery negative (E on our drawing) inside the controller.
In machines, the earth return connection quite easily gets connected back to somewhere other than the pot reference. In out drawing, consider this return being connected back to point D, somewhere near the battery.
But we have already seen that the battery wire C-E has a voltage drop in it. Some of this voltage drop is of course present in the wire D-E. But as the input signal is being referenced to point D, the voltage drop in the wire D-E is actually in series with the input, so is being fed back into the controller’s own input! It will of course affect the controller, therefore the controllers output, therefore it will affect itself! An unpredictable feedback loop exists.
Every signal source has a return wire and every time you connect an input signal, you must consider where the return is and make sure the return is clean and uncontaminated by any high current signal.
Earth loop
There is another page explaining Earth track fuse and Earth loops on motor controllers.
So you are feeding the input correctly into the input connector, between pins F and G. But the PLC is drawing its power from the same (battery) supply as the controller. The PLC’s power earth may, for instance, be fed back to point D. But is the power earth connected internally in the PLC to the signal earth? If so, then the voltage drop in the wire D-E is in parallel with the wire D-L-G and, through the controller, so to E. A large circulating current can flow as a result. This current is flowing in the earth path of the controller, which itself has an impedance, so it will insert an unwanted signal into the earth line, which had not been designed to cope with it. In bad cases, earth tracks can easily burn out.
So, as well as being aware of signal returns, you have to be aware of possible alternative signal paths. Even though the signal earth may appear separate from the power earth, in practise the two may be more or less coupled.
Spider earthing and chassis earth point
A good earthing scheme in most cases is a ‘Spider’ earth. Chose one point in the machine and run each and every signal and power earth back separately to this one point. It is often convenient to have this ‘spider’ point as the chassis earth. It should (as we have seen) be close to the battery negative terminal of the main power control device as is possible (i.e. to the motor speed controller), without being silly.
Clearly this requirement can get a bit tricky when the machine in question uses many power controllers, but the spider earthing system is then really the only way to minimise problems.
Radiating Wires (Transmitters) and Sensitive wires (Receivers)
As we have seen, any wire carrying a varying signal can transmit that signal and induce unwanted signals in any other wires. So be aware of which wires carry high currents and so can transmit interference and of which wires are carrying low level control signals, so may be able to be disturbed by unwanted received signals.
Keep these two sets of wires well away from each other.
Sensitive wires are also subject to the same rules as transmitters: a loop with a large area is as good at receiving interference as it is at transmitting it, so keep signal wires and their returns together, at least running parallel. Twisted, or even screened if possible.
Solenoid switch off noise
Solenoids (e.g. relay, contactors and similar) give a very high voltage transient when they are switched off. This has two effects. Firstly, is causes burning of the contacts used to break the solenoid current, because it causes arcing at these contacts. But also the transient spike is a very good source of noise, which can be picked up in sensitive places. Although such switching transients are usually not a huge problem, as a matter of good practise the coils should always be suppressed. In a battery operated machine, this is best done by connecting a normally reversed diode across the coil. The switch off transient then causes current to re-circulate harmlessly through the diode. This ‘coil catching diode’ is discussed in our circuits section.
Transients
Not only contactors and relays can cause transients: wiring of large machines is unpredictable in terms of inductance and capacitance. Machines are also generally used in noisy environments.
Large voltage spikes will destroy MOSFETs. It is very difficult to predict what will cause such transients, so 100% suppression is difficult! It is therefore very advisable to fit transient suppressors across all sensitive components.
It is therefore suggested that a suitable suppressor be fitted across the battery + and Battery – connections at each controller, and also across each motor in addition to the motor capacitor.
Tranzorbs do not have a sharply defined clamping voltage, so you will normally need to chose one with a clamping voltage well below the maximum voltage of the MOSFETs you are trying to protect. You may then prefer to use a semiconductor transient suppressor with a more accurately defined clamping voltage – but of course what you need to use rather depends on the transients you are trying to suppress.
Fighting robots
This can really test the reliability of a complete system. Motors can be stopped [jammed] instantly, can be switched from forward to reverse rapidly, or can be pushed backwards against their chosen direction. In these cases you need to think about what happens to the kinetic, magnetic, and electrical energy that is flowing in the system. It is beyond the scope of this article to cover all these scenarios in detail, it’s probably best just to say that there will be a lot of large voltage spikes so fit as much protection as possible to as many places as possible.