Fullscreen HTML5Med Res (??MB)HTML4

REMOTE CONTROL BY BOB YOUNG Designing a speed controller This month, we will discuss some of the factors influencing the design of a simple low-cost speed controller which will be presented in this magazine. The overriding consideration is that of "milliohms" - the fewer, the better. To begin, a brief description of how our speed controller functions is in order. Speaking in very broad terms, a speed controller is an electronically controlled variable resistor inserted in series with the motor (Fig.1). The control elements are usually FETs (field effect transistors) or bipolar transistors. The control voltage is derived from the receiver battery which brings us to the first problem. FETs in particular require drive voltages of around 12V to reduce their ON resistance to the lowest possible value. As the re- ing from the receiver battery to provide the drive voltage. The choice between bipolar or FET transistors is an easy one. There are problems associated with paralleling bipolar transistors and since we are after a very high current with a very low ON resistance, parallel semiconductors are more or less forced upon us and the clear choice is FETs. There is only one practical way to use the FETs in a continuously variable speed control and that is to rapidly pulse them on and off. A linear control system (ie, virtually a variable "It must be stressed that in this project we are moving into the world of high current devices and therefore we are primarily concerned with making improvements which will be measured in milliohms." ceiver battery is usually only 4.8V, some voltage doubling or tripling is therefore required in the drive circuitry. We could use the motor drive battery which is often 7. 2V or as much as 30V in some cases. But in some applications, the motor battery is only 4.8V, particularly in the case of very small motors. To allow maximum versatility of the final unit, I have therefore decided to use a voltage doubler work- resistor) could not work because at the half throttle setting, the amount of power dissipated in the FETs would be the same as in the motor. Clearly, this is very wasteful of the batteries but also means a very large amount of heat that the controller must dissipate. So modern speed controllers are all switching designs. This switching frequency usually ranges from 50Hz to 2.5kHz and I have decided upon 2kHz as an achiev- able figure. The problem here is that each gate has an input capacitance of about 1500pF and this will begin to round off the gating pulse. The 2kHz signal to the gates of the FETs is pulse width modulated (PWM) and the output power of the speed controller is proportional to the width of the switching pulse. Thus, a 50% duty cycle delivers half throttle. At full throttle, the pulse signal becomes a constant DC level. High currents The foregoing factors are the major design considerations. However, from the very outset it must be stressed that in this project we are moving into the world of high current devices and therefore we are primarily concerned with making improvements which will be measured in milliohms. The idea of dealing in milliohms is a little strange to most people, particularly to electronics buffs who are more used to dealing in ohms, kilohms and megohms. Most of us are very familiar with the milliwatt (mW), milliamp (mA) and millivolt (mV), but how many of you have ever encountered the symbol for milliohm (mQ)? Yet lmQ is a very significant amount of resistance in circuits drawing 1000 amps or more. A resistance of lmQ at 1000A will result in a voltage drop of 1V and a power dissipation of 1000 watts. By the time we have finished the next two articles, you will all be very aware of the effect of just a few milliohms on even reasonably low current circuits. To fully illustrate the point, let us look at a real life situation. Perhaps one of the best examples of high current devices we all encounter in an everyday situation is the automotive starter motor. A typical starter motor draws about FEBRUARY1992 69 400 amps when cranking the engine. I have not seen the instantaneous startup current quoted for a starter but it must be in the order of 1000 amps or more. (Editor's note: typical "locked rotor" currents are around 1200 amps or more). If we take 1000 amps as an average figure, then parasitic resistances in series with the armature of just 12 milliohms would result in most of the energy available from the car battery being dissipated as heat before it ever reached the starter motor. Such a situation does arise from time to time as the starter components age. Batteries sulphate up and the internal resistance begins to rise. The solenoid contacts burn and the corrosion that forms on the battery terminals gradually builds up until one cold morning the system finally falls over. Out comes the spanner and emery paper for the battery terminals and warm water for the battery. There is little we can do for the solenoid contacts, as they will have to wait until we can get to a garage. In the meantime, if we can at least shed a few milliohms then we can get the car started and be on our way. You can now afford a sate IIite TV system For many years you have probably looked at satellite TV systems and thought "one day". You can now purchase the following K-band system for only: $995.00 Here's what you get: * A 1.8-metre prime focus dish antenna, complete with all the mounting hardware. * *magnetic One Ku-band feedhorn and a signal polariser. * 30 metres of low-loss coaxial cable with a single pair control line. * A 99-channel infrared control satellite receiver with adjustable IF One super low-noise LNB (1.4d8 or better). and audio bandwidth, polarity and dual digital readout. Extraneous resistances Your receiver is pre-programmed to the popular AUSSAT transponders via the internal EEPROM memory. This unit is also suitable for C-band applications. Call, fax or write to: AV-COMM PTY LTD PO BOX 386, NORTHBRIDGE NSW 2063. Phone (02) 949 7417 Fax (02) 949 7095 All items are available separately. Ask about our C-band LNBs, NTSCto-PAL converters, video time date generators, FM 2 and EPAL. I I I I I I I I I I ----------YES GARRY, please send me more information on K-band satellite systems. Name ............ .... .................... .. ........ Address ........................................... ........................... P/code ................ Phone .............................................. ACN 002 174 478 01/92 I I I I I I I I I I ·----------- ■1 70 SILICON CHIP Fig.2 shows the basic problem in simple terms. A motor fitted with an llQ armature is supplied from a 12V battery. The motor installation has a number of miscellaneous parasitic resistances in series with this armature. These are typically made up of the switch, connectors, wire, battery internal impedance, and brush and commutator resistances. These we will lump together and define as RLP (lumped parasitic). RLP does not include the armature winding resistance as this is defined as the load resistance and is a separate item. RLP is the actual parasitic resistance introduced by the practical requirements for using the device in the real world. They derive from the insertion of wire looms, contacts for ease of charging and servicing, fuse holders, switches for turning the system on and off and ageing of the brushes and commutators. Now the problem is that all of these lumped parasitic resistances can add up to a considerable resistance if care is not exercised in the design and maintenance of the motor installation, and this problem compounds with the amount of motor current required. + T I I ...,_ PULSE GENERATOR Fig.1: the basic scheme for an electronic motor speed controller. The pulse generator controls the on-time of the FET which in turn acts as a variable resistor & thus controls the speed of the motor. l +T 1 ...,_' 12V Fig.2: a parasitic resistance of just Hl in series with a motor with an 11n armature will rob the motor of 1W of power. This problem gets worse as currents get higher. Thus, ifwe give this lumped parasitic resistance a value of lQ, the total cir" cuit resistance becomes 12Q. This will result in an instantaneous start-up current of one amp. With a start-up current of lA, one full volt is dropped across the parasitic resistances, thereby robbing the motor of one watt of badly needed power. This is serious enough but consider the situation where we require much greater motor performance such as in our car starter motor. Consider what happens if we replace the 1 lQ armature with a 12mQ winding. With no parasitic resistance in series, this" armature will require 1000 amps instantaneously at start up. With just 12 milliohms of parasitic resistance in series with the armature, we halve the instantaneous current.to 500 amps and worse, half of the initial power is wasted. This means that the available power at the instant of startup is down to a quarter. Thus, it is TABLE 1 Model WHD Max. Voltage Min. Voltage Max. Current Cont. Current KoCx-111 412544 8.4V 6V 2100A 510A 0.05V 0.004 ohms Novak 410-M1c 3619 40 12V 4.8V N/A 250A 0.06V 0.005 ohms Novak 410-MXc 45 19 43 12V 4.8V N/A 500A 0.04V 0.003 ohms Tekin 411 P 41 19 36 13.75V 5V 1050A NIA 0.06V 0.005 ohms Tekin 420F . 38 19 51 20V 4V N/A 400A 0.04V 0.003 ohms easy to see that in order to deliver the maximum power required for starting the car engine, we must hold the parasitic resistances down to less than lmil Now lmQ is not a lot of resistance. Car designers do not specify 10mm thick cables for the starter leads for nothing. Even car designers get it wrong occasionally, though. The starter solenoid is a high current switch which has two windings: the "pull-in'' winding and the "hold-in" winding. At switch on, both coils are activated but once the solenoid is fully engaged, the "pull-in" coil is released, leaving only the "hold-in" coil energized. The total current consumption of the full solenoid is typically 35A, with 20A a typical figure for the "hold-in" coil only. . The problem is that the designers originally passed this current (35A) through the keyswitch on the steering column with the result that the keyswitch usually melted at some stage in its career. Modern practice is to mount a solenoid boost relay very close to the solenoid and run only the coil current of this relay through the keyswitch. The solenoid coil current is run through the relay contacts with very short leads. "ON" resistance Similar problems arise in the design of speed controllers for model aircraft, with one additional problem. As already noted, we are deliberately introducing an electronically variable resistor in series with the motor. In the ideal world, this would not be too bad, for there would be no voltage drop across the control semiconduc- tor in the full "ON" condition. Unfortunately, in the real world, there is a voltage drop and it can be quite considerable and extremely detrimental to the performance of the electric motor we are controlling. The typical "ON" resistance for even an exotic FET such as the IRFZ44 is given at .028Q or 28mQ Now we can see immediately that here is a potential source of power loss of very significant proportions. Remember, too, that these figures were derived under laboratory conditions. We are faced with the task of converting this device into a practical working model so great care must be exercised if undue losses are to be avoided in our finished design. Commercial benchmarks At this point it would be interesting to establish what modern commercial speed controllers are achieving in the way of performance to establish the benchmark for our own efforts. Table 1, based on data from the November 1991 issue of the American magazine "Radio Control Racer", shows comparative figures for several modern speed controllers. The figures quoted are, at first glance, quite stunning, with instantaneous currents of up to 2100 amps and sustained currents of 250-510 amps being quoted. The resistance of the systems was measured at the 5cm point, being in the range 3-5mil The voltage drop at this point was typically 0.04-0.06V. But hang on, 2100 amps at 4mQ gives a voltage drop of8.4Vacross the 5cm of wire alone. Where do we find enough electrical headroom for a mo- Voltage at 5cm Resistance at 5cm tor armature in series with the 4mQ of parasitic resistance? And what is the 0.05V drop at the 5cm point all about? Here we encounter a real problem in that figures quoted at random without their companion figures are completely useless. Thus, we need to know what was the test voltage for the 2100 amp result. In all probability it was a short circuit test at 8.4V but we can only guess, for nowhere in the text accompanying that table does the magazine give any hints. Likewise, the 0.05V drop at the 5cm point; at what current? Here at least we have a clue. The resistance at this point is given as 0.004Q (4mQ). Thus Ohm's Law gives us a figure of 12.5A, a reasonable enough figure. These are still very good figures and one wonders just how accurate they are? To prove the point I obtained access to a brand new car fitted with a Novak 410-MXc, a 6-cell pack and a Kyosho Super Stock 20 motor. At 7.51V, the motor drew 3.65A driving the back wheels with no load. A voltage drop of0.02V was measured across the negative battery terminal and the FET side of the motor terminals. Loaded to BA, the voltage drop was 0.04V which gives a calculated 0.06V at 12A. At 8A, the tyres nearly took the skin off my assistant's hands, so we did·not try for 12 amps. You will notice that I was too smart to try to hold the tyres. I did the hard stuff on the transmitter throttle. Let me tell you right here and now that these units may be expensive but they are good and comparable performance represents a difficult goal to achieve. But we think that our proposed design will stack up pretty well. SC FEBRUARY1992 71