On controllers that use resistance detection, you will of course need to fit a resistor in place of the pot, to override the pot fault detection circuit.
The reverse line on all controllers are high impedance and can easily be operated digitally. Easiest method is a pullup resistor to positive line (10K will suffice) with a transistor to 0v. The method is shown on Joystick Interfaces.
The ignition line is easiest operated by a PNP transistor pulling it to the positive line, since extra impedance here will increase the battery under-voltage protection level (assuming the feature is present!). The ignition current is small, about 600µamps, so a 100K base resistor is fine. Or you could use a relay with the contacts as the ignition switch.
In the Pro and VTX it is possible (via the optional expansion connector) to input direct to the modulator. This is still decoupled and will still accept a PWM signal.
See also Machine - good and bad practises for additional hints.
Dual Ramp Reversing and Regen Braking
Controllers such as the 4QD, Pro and VTX have 'Dual Ramp' reversing. If you operate the reverse switch when the motor is moving, the controller detects this and automatically slows down the motor at a rate defined by the deceleration ramp. When the motor has essentially stopped the controller reverses it and ramps it up tp speed in the new direction.
To know when to reverse, the controller has to be able to either measure the armature voltage (which will fall to near zero when the motor is slow enough) or it has to make a good prediction of when the motor is stopped.
These controllers all have regen braking. So the controller can safely assume that the motor is doing what it is told and that, when the controller's output voltage has fallen to near zero, then the motor is essentially stopped. But the controller's output is controlled by the demand voltage, after the ramp control. See our circuits archive for a description of how the pwm operation works.
Because of the regen braking, the controller can safely assume motor voltage. It does not need to measure the armature. This is exactly what is done on the Pro and VTX series. The 4QD however does measure the armature voltage.
So there is no provision on VTX and Pro to disable the regen braking. If this were done, the controller could not assume that the motor was doing as told and would have no way of knowing when it was safe to reverse. The 4QD series however, with armature voltage sensing, can have regen braking disabled.
In fact, if there was demand, the VTX series circuitry could be altered to allow regen braking to be disabled. The controller would then have to be used with pre-set reversing. See our guided tour of features. However in practise there are very few occasions where regenerative braking is at all disadvantageous.
Efficiency
A lot of users are rightly concerned about battery life. 4QD's controllers are carefully designed to minimise losses and therefore to maximise battery life. Older controllers used low (audible) frequency for switching: this increases losses in the motor, so the motor runs hotter. 4QD's controllers use 20kHz to reduce motor losses to a minimum. Also, 4QD controllers draw very little current for the control circuitry. Lastly, we employ regenerative braking which actually uses the motor as a generator during braking and recovers some of the energy which is returned to the battery. There is also some advantage in using a higher current controller and the thickest wiring practical - the higher the current handling ability of these the less power will be wasted as heat.
Other causes of loss are inefficient motors: a permanent magnet motor will always be a better choice than a field wound motor since the field draws current, the permanent magnet doesn't. We can do nothing about this in our controllers.
The other major cause of loss in most machines is the gearbox. Since this usually comes with the motor economics rule here, but a good pinion gear train is more efficient that a worm and wheel, especially with regeneration where worm drives are usually very poor indeed.
Do not use belt drive (toothed or plain) with a system employing regenerative braking. See Toothed Belts.
Fork-lift trucks
Historically most industrial vehicles, including fork-lift trucks, have used series wound motors. Modern trucks still tend to use these motors, because they are cheap and a tried and tested technology. There is also a large service industry supplying spare parts (contactors, resistors etc) for them.
See section on 'Series wound motors' for more information.
You can get electronic controllers specifically for series wound motors: the controllers also have outputs for the direction change contactors, to switch these at the correct time to control reversing. Switching the contactors at zero current stops arcing and extends the contactor life.
Fuses and circuit breakers
This is a difficult subject!
- Difficult because motor currents in most applications are very peaky.
- Difficult because the controllers battery current if not the same as the motor current.
- Difficult because users haven't really got a good idea of what they are tying to protect with the fuse.
- Difficult because fuses themselves are quite unpredictable in their carrying and blowing performance
The two things you could protect with a fuse are the motor and the controller. But you must first decide what sort of fault condition you are likely to encounter and what effect this will have on motor and battery current. Clearly the fault conditions will depend on the application!
Motor
Let's say you want to protect the motor against a stall current. What is the motor's stall current? Is this higher than the controller can give? It should be - part of the controller's job is to protect the motor against stall currents. If it doesn't then you have an over-size controller!
So it's not the stall current you need to protect - so exactly what current/time rating fuse or trip will give you protection where the controller won't? There's no simple answer - you need to understand the problem first. But selecting a motor fuse rating is not as easy as you first thought!
For relevant information, see:
- Current Limit
- Current requirements
- Motor ratings
- PWM controllers - how they work.
Controller
So how about a battery fuse, to protect both motor and controller? Again - what do you want to protect against?
Remember - motor and controller currents are not the same. Since the controller is very efficient, the motor power if very nearly equal to the power supplied by the battery. So motor volts times motor amps equals battery volts times battery amps.
So to define the maximum likely controller current, you need to guess the maximum motor voltage which will occur under the likely fault conditions. That is not easy. Probably your best choice is to measure the battery current under different conditions and chose a fuse which won't blow under normal conditions. But even that's not easy - you probably do not have suitable measuring equipment.
As an empirical method, try a fuse or breaker in the battery rated at the same current as the motor's nameplate (continuous rating) current. If it blows under real conditions, increase it until nuisance tripping stops. But we cannot guarantee that it will necessarily catch all faults!
Battery reversal
This is best protected against as explained in the section Reverse polarity protection. However, a fuse can, in most setups, give adequate and simple protection.
If the battery is reversed, current will be limited by the battery impedance, the battery wiring and the voltage drop in the MOSFETs of the controller and the voltage drop in the fuse or breaker you are fitting. In lots of set-ups, the battery wiring is sufficiently thin to restrict current so that a fuse will probably blow before the MOSFETs, and this is probably the most significant reason for fitting a supply fuse. It needs to be as small as possible but large enough so that nuisance tripping does not occur, but its exact value is impossible to state as it depends on your wiring and on your battery.
As a rule of thumb, a 25A fuse is about right for VTX-35, Uni-4 and similar controllers. For Uni-8, VTX-60 and VTX-70 a 40 amp fuse is indicated (mainly because higher value blade fuses are difficult to get).
Golf-carts
Unfortunately there is some confusion in nomenclature: golf cart and golf trolley are both indiscriminately used to describe both passenger carrying golf buggies and also motorised golf bags, or caddies.
Golf buggies
For passenger carrying golf buggies the current required is determined by the maximum speed, vehicle weight and payload (weight of rider) and the required hill climbing ability (usually 1 in 3 or 1 in 4). Given all these parameters we can work out the required current but, for one man machines, we would generally recommend our 4QD-150D-24 or Pro-120-24 for 6 mph vehicles and our 4QD-200D-24 for 8 mph vehicles, assuming you want to climb the 1 in 4 hills some courses have.
Alternatively our VTX-70 is recommended. This is designed so two controllers can be used in tandem, one driving each motor. This gives the required current and has independent current limits for each motor, so if one motor gets damaged, the other sill still work and should get you home with little chance of a burn out. Both 4QD, Pro and VTX series have reversing and full regenerative braking so are ideal controllers.
Commercial carts (outside of America) almost all use EMD motors, the smaller ones use two EMD 180w motors (PM50-63) with a 110 amp controller and a top speed of 6 mph. However these can be under-powered. A better choice, as used in the better buggies, are two EMD 300 watt motors (PM63-50) with a 4QD-150 amp controller. This combination is generally OK up to 8mph on even the most hilly courses, depending on all-up weight.
Two Sinclair C5 motors are a similar power to two EMD PM63s, but, although they can be used safely on 24v, their top speed is then 6000 rpm, so they need appropriate gearing. Otherwise use them in series, when they each will get 12v from a 24v controller.
For a cheaper, non-reversing system, our 'Scoota' controller is ideal.
Golf caddies
For motorised golf bags 12v at about 30 amps is adequate: less than 30 amps may mean that the trolley needs help up hills. Our Porter controller was designed for this type of use and it suits most caddies.
The Uni-4 or Uni-8. Like the Porter, these have reverse polarity protection. Also, like the Porter these potentialy have regenerative braking. However most commercial caddies have a free wheel device which renders regen inoperative, though it is not a hindrance. If there is no freewheel, them the regenerative braking is a definite advantage.
Heating.
All components or wires which are carrying any current get hot. How hot it eventually gets depends on the rate that heat is generated in the component and on how quickly heat can get out of the item into the surroundings. How quickly the item reaches this steady temperature also depends on how much heat the item can absorb or how 'large' it is thermally, i.e. on its 'thermal mass'. The situation is more complicated in most motor control application because heat generation is very intermittent.
Heat generation.
Electrical heat is generated by electrical current (amps) flowing through a wire or component which has electrical 'resistance' (ohms). All electrical components and wires have resistance but the larger the wire, the lower its resistance. The heat generated is the square of the current multiplied by the resistance. Electricians refer to current as 'I' and resistance as 'R' so the heat generated is I²R. Note that the voltage does not come into this: heat generation has nothing directly to do with the voltage. However the power the device is handling is directly concerned with the voltage since power is volts times amps. This is why a 24v system is better than a 12v one: it can give double the power at the same heat or, if you halve the current, the same power for ¾ the heat.
If you are a beginner to electricity and all these amps and volts confuse you, try reading our page Understanding Electricity - an analogy with water.
Heat dissipation
A hot object dissipates heat, i.e. it looses heat into the surroundings. The rate of heat loss depends on the area over which the heat is dissipated, how good the 'contact' with the surroundings, hot much hotter that the surroundings the object is, the speed of airflow over the object and even the colour of the object. In practise the dissipation is eventually to the air (unless you are on a boat and can use water) so the 'contact' is down to the surface roughness (whether the object is finned) and the airflow: blowing over your coffee cools it much quicker than letting it stand in still air.
Heat conduction
Usually the object which is generating the heat is not the same object that is losing the heat to the air. The heat has to flow from one object to another, usually through other objects on the way.
The heat conduction path does not let the heat flow unless there is a temperature drop all the way along it. All the objects through which the heat must flow 'resist' the flow of heat. An object which has a high resistance (be it to heat flow or to electricity) has a poor 'conductivity'. A good conductor does not resist the flow of heat so much as a poor conductor (this is the same for heat as for electricity and usually the two are linked).
If there are several objects through which the heat has to flow the contact between each item and the next must be good: a very small gap, even in the form of surface roughness, can dramatically reduce the heat flow (i.e. it can add resistance to the thermal path) by trapping a thin film or air, which is an exceedingly good insulator. Obviously this is bad so it is common practise to use 'heatsink compound' to fill the gaps caused by roughness when joining two metal parts through which heat must flow.
Thermal mass
If an object has a lot of 'thermal mass' it takes a lot of heat to increase its temperature. If we do not want an object to heat quickly, then we must have a large thermal mass. Thermal mass will even out the heating, so a few short, high bursts of heat don't cause overheating. However thermal mass alone won't cool an object. The mass only stores the heat - which must still get dissipated safely. A high thermal mass increase the time the object takes to get hot, but its final temperature is still determined by the balance between heating and cooling rates.
Materials
How do various materials perform? The table below shows relative values for several metals.
| Conductivity | |||
|---|---|---|---|
| Material | Thermal | Electrical | Specific heat |
| Aluminium | 50 | 339 | 20 |
| Copper | 92 | 562 | 9 |
| Gold | 70 | 413 | 3 |
| Lead | 8 | 48 | 3 |
| Silver | 100 | 602 | 6 |
| Steel | 11 | 50 | 10 |
| Water | 100 | ||
Specific heat is the ratio of 'thermal mass' to actual mass, note how very good water is. If you want a high thermal mass, use a bucket of water! Not very practical: electronics does not like water! The second best is aluminium. A kilogram of aluminium can absorb twice as much as a kilogram of copper or steel. However, steel is about 3 times more dense than aluminium, so a 10cm cube of steel is actually about 1.5 times better than the same size block of aluminium.
Thermal conductivity is related to the volume, so a rod of silver will conduct heat twice as well as a rod of aluminium the same size. Steel is 5 times worse than aluminium. Anyone who has tried welding these materials will appreciate the difference! Copper is pretty good too but, because of its low density, aluminium is, weight-for-weight, the best. Steel, weight-for-weight, is very bad indeed and, where heat conduction is important (as it always is in a heatsink) should not be used, except where there can be a very good contact area into a large mass or where there is sufficient thermal mass to even out the heat flow. If you need to mount on steel, then a heat spreader made from aluminium (or copper) will get the heat out of the source, spread it over a wide area and pass it into the steel. Make sure to use heatsink compound in the joints!
Industrial use
4QD controllers are ideal for industrial use as a 'building block' for designing into a system. We have many years experience in the design and manufacture of industrial machinery control systems and can usually help. We can also usually supply suitable circuitry.
Closed loop control
The Pro, Scoota, VTX, 2QD and 1QD controllers do not incorporate internal armature voltage sensing. For most applications this is an unnecessary complication since the operator compensates for speed changes automatically. The 4QD series does incorporate armature voltage sensing. However this alone gives a limited improvement since the motor's speed still changes because of the voltage loss in its internal resistance (IR loss). It is possible to compensate for motor IR loss, but the controller then must be set up for each different motor: this causes much confusion as setting up requires much skill, or a tachometer and a method of loading the motor whilst measuring the speed. The other problem is that the existing internal current sensing (which could be used for IR compensation, is sensitive to the MOSFETs internal temperature, so the IR compensation would drift.
The best method of closed loop control is via a tacho-generator. It is relatively simple to add on to the 4QD series as the internal circuitry used for armature voltage sensing can accept a tacho generator circuitry with minimal change.
It is also possible to add a tacho generator to the Pro 120 or the VTX with our optional tacho feedback board. Other controllers can also use this same board or, if you wish to make a circuit yourself, see Tacho generator motor speed feedback for the circuit of a suitable error amplifier to compare tacho feedback voltage with the demand speed.
It is also possible to use a pulse generator to measure motor speeds. If using analogue control this needs to be processed via a frequency to voltage converter to be used as feedback. The slowest motor speed will determine the maximum time between pulses and this will determine the minimum response speed of the F to V converter: if it responds too quickly the motor will accelerate between pulses! But if you chose a long time you will need slow response which can cause problems with motor speed stability. Pulse circuitry is not direction sensitive (unlike a tacho generator) so the error amplifier need only be uni-directional.
Because of these timing constraints, pulse generators are usually used with computer control systems: in this case the computer can give the controller a demand speed via a digital to analogue converter: the software can control response times and difference amplification (if the demand speed is input digitally), or the D-A can output to a conventional analogue error amplifier, working direct from the demand speed control.
The 4QD, Pro and VTX series have, at their inputs, sophisticated analogue ramps: ideally the error amplifier should be inserted after the ramping circuit and before the modulator: contact 4QD for more assistance.
See also Speed Stability
Constant torque mode
Torque mode is used on, for instance, a machine used to wind a continuous material onto a spool. Instead of the input controlling the speed, it ideally should control the motor current, hence the torque of the motor. This is not a very common application so it has not been designed into 4QD's standard controllers - such features inevitably increase the cost so that all customers end up paying for features they do not actually need!
However all controllers made by 4QD incorporate a current limit and this current limit can be altered by adjusting a few components, so a controller could be used to supply a constant current, which in some applications is adequate.
Remember that the standard current limit is there solely to protect the MOSFETs from transient over-currents which could be destructive. If you run on the standard current limit for too long, then the controller will simply overheat and melt - quite literally, it is indeed possible to melt the solder on the MOSFET leads before the controller fails! So you will need to reduce the current limit quite significantly, to less than, say, 25-35% of the nominal current to prevent overheating. The exact allowable current will of course depend on your heatsinking, the application and the controller you use. Contact the factory for advice.
Jockey wheels, motorised
Motorised jockey wheels (used for shifting caravans or RVs on site) normally use a 12v system: an EMD PM63-100 motor and a motor speed controller. The motor may take 60 amps or more for short periods. An VTX-70-12v will be ideal or you can use a 2QD if you don't want reversing.
Several commercial units exist: mostly these use the EMD motor. In fact since these are generally pretty slow moving, many of them don't bother to use a motor speed controller at all, simply putting up with the slight jolt at start up and accepting the necessity of replacing a relay occasionally because of the motor current surge causing contact arcing.
Joystick operation.
A 'Joystick' is a control operated so that the centre of stick travel is zero speed. Move the control forward to run forward and move the control back to run backwards. Some of 4QD's reversing controllers (i.e. the 4QD series) can be used in this way, others not, as standard. However for the simpler reversing controllers, we supply optional Joystick interface boards.
The 'Joystick' itself can be a linear control (slider pot) or, more usually, a rotary pot with a lever, used only over the centre of its rotation. This is a 'Single axis Joystick' which will control speed and reversing but not steering. To control steering two controllers are required, each controlling a separate motor, with a 'Dual axis Joystick' system.
Dual axis joysticks also allow transverse movement (in addition to front-to-back). In some applications (e.g. a remote controlled pan and tilt head for a camera) the two motors can be controlled one by each axis, i.e. one controlled by front to back movement and the other by transverse movement. This is two separate 'Single-axis' controls rather than a true 'Dual-axis' control.
In other applications two motors are used (with separate controllers) one to drive each wheel of a vehicle. In this case transverse movement of the stick must speed up one motor whilst slowing the other so the vehicle steers by the difference in speed. This is a true 'sum and difference', 'dual-axis' system, offered by our Dual-Axis JSI.
4QD-150 series. This has circuitry for joystick control: all that is required is a small modification, to be done at the factory.
VTX series. Suitable for interfacing to a joystick via our interfaces.
Pro-120. Suitable for use with our interfaces.
Kiddie-cars
This was one of the target applications we considered in designing our VTX range of controllers. The VTX-35 should prove a good choice, unless the car is particularly heavy, very fast or to run on very rough ground, when the VTX-60, VTX-70 or even the Pro-120 would be a better choice.
The current you actually require will depend on the weight of the child, the weight of the vehicle, the performance you expect (i.e. the top speed and the hill climbing/rough ground ability) and the voltage. For an average car operating at 12v with a top speed of 4 mph, a 35 amp controller should generally give adequate performance. If the top speed is more than 4 mph then you need proportionally more current. For 24v operation, halve the required current - or double the top speed, depending on your gearing. Because of the reduced current requirements, 24v is preferable - if you don't mind the extra battery.
For the reversing switch you can use any ordinary single pole switch. Or you could mount a reed switch with the magnet on the gear lever to give a proper floor mounted reversing switch. More information on the reed switch.
Alternatively, if you don't want reversing, our 2QD controllers are ideal. They include regenerative braking which is very useful. For heavier use, the Scoota may be indicated.
For the speed pot (for either range of controller) you can use a simple rotary panel mounted potentiometer, our 'Bicycle Bell' throttle or, if you want a foot pedal, the Plunger operated pot is easy to fit, but expensive. Otherwise you can use a rotary pot with a lever attached, since the controllers have adjustable gain to suit a rotary pot with only a few degrees of movement, but you will need to be ingenious with the operating linkage. The naked mechanism from our 'Bicycle bell' throttle lends itself well to this application.
Lighting
Low voltage lighting circuits may be run off a Uni (or 1QD or 2QD) controller: this enables a vehicle with 24v, 36v or 48v battery to use a 12v circuit for lighting, horn etc. The Uni can of course also be varied, so you can have a variable voltage lighting circuit if you so desire.
The Uni is designed to chop the negative feed to the motor, so the battery positive is the common connection: you cannot then normally use a negative earth system. However our controllers (those that include regen braking) do have a fast high side current limit so it is possible to use these controllers 'upside-down', i.e. with a common negative. The problem here is that the circuit won't switch on properly as it is designed to start up smoothly with a common negative.
The output from a controller used thus is not pure 12v, but chopped 48v (or whatever the battery is) so it is quite unsuitable for radios and similar units. It does however work well with bulbs, horns and small motors. Because of this restriction on its application we do not consider such a controller to be sellable as a general purpose dc-dc converter. However the Uni is hugely cheaper than any commercial dc-dc converter and (within the limitations above) it works very well indeed!
Also - because it is not d.c. but chopped battery voltage (i.e. a squarewave) you should be wary of trying to set the voltage up by a meter! Meters are not calibrated to read the heating effect of squarewaves and are more likely to read an average voltage. This can be very misleading. A bulb is essentially a resistive device (whose resistance depends on its operating voltage). Now the heating effect of a voltage is proportional to the voltage squared - so the bulb on 24v might get four times the heating as on 12v. We want the same heating - so a 25% duty cycle is called for: i.e. four times the heating effect for 1/4 the time! This is an average voltage of only 6v across the bulb.
Probably the best way of setting up such a system is by the use of a 'Grease-spot photometer' to set the brightness of the bulb to be the same as another operating from a 12v battery.
Locomotives & Trams
This was a target application for which the VTX series controllers were originally designed. Many are in regular use all over the world. The VTX has evolved into the VTX.
The current a locomotive will take depends on the loco's mass (including passengers), the top speed of the loco and the size of any gradients it will have to climb. Given all these figures, you can actually calculate the current required, and we have a calculator on site for this purpose. However a 5 loco takes typically about 30 amps at 12v and about 15 amps at 24v while an average 7¼ gauge loco may draw 50 amps at 12v or 25 amps at 24v. It is probably best to chose a controller slightly larger than you anticipate, but a typical garden loco could be built by choosing the controller from the chart below.
| Gauge | 12v | 24v |
|---|---|---|
| 5¯ | VTX 75 12 | VTX 40 24 |
| 7¼¨ | VTX-70-12 | VTX-75-24 |
For more heavy duty, especially for 7¾ locos, consider either a Pro-120 or a 4QD series controller - these are also available for 36v and 48v operation.
Some people prefer the regen braking that these controllers give - others dismiss it as 'unrealistic'. This depends on who is driving the machine. If it is a garden railway and the kids will be using it, we strongly recommend regen braking and a reversing controller that will handle this safely.
See also our accessories page for model locos.
|
|
Page Information
Document URI: www.4qd.co.uk/faq/bmnc2.html
Last modified: Friday, 27-Jun-2008 08:27:01 BST
Page's Author: Richard Torrens
© 1996-2006 4QD