There are two sorts of analogue electronic controller: closed-loop and pulse-width modulation. In closed-loop controllers the control knob sets the output voltage, a proportion of which is fed back to the input as a control loop. They have the advantage of quiet operation, but the output transistor can run hot and train starting and slow running, although good (far better than the old rheostat controllers) leave room for improvement. Pulse-width-modulation controllers deliver pulses of full power: the speed control sets the duration of the pulses. At low settings the pulses are brief and the spaces between them correspondingly long. As you advance the speed control, the pulses get longer and the spaces shorter until (on some controllers) at full speed the spaces disappear and continuous full power is delivered. Trains run noisily - sometimes they buzz like bumble bees (which can be quite realistic with diesel-outline models!), the output transistor runs cool, starting and slow running are unbelievable, but beware of prolonged slow running as it causes motor heating.
The PWAyMan combines the best of both types. It is a Pulse Width and Amplitude Modulation (PWAM) controller. (I expanded the initials PWAM to PWAyMan to make it sound more railway-oriented!) The output consists of square-wave pulses. As on a closed-loop controller, however, the control knob sets the pulse height, the voltage, directly. During the spaces between the pulses a servo monitors the motor voltage: as the controller output itself is temporarily zero, any voltage present must come from the train motor acting as a generator (this is the so-called 'back EMF'). This voltage is directly proportional to speed and can be displayed on an optional speedometer. The servo compares the motor voltage with the output voltage set on the control knob and the result is used to adjust the pulse width. The result is very accurate speed regulation: as the train negotiates an uphill gradient or a tight bend, the pulses are lengthened to provide the extra power needed to maintain the set speed; as the train encounters a downhill gradient, the pulses are shortened as less power is needed to maintain the set speed.
Although it may appear complex, the circuit consists of a number of interconnected sections each of which is fairly simple. It does make use of some industry-standard i.c.s (chips). The circuit is shown in the following diagram:
It is important that the controller is powered from a supply delivering 16 V (r.m.s.) full-wave rectified a.c. D1 and C1 derive from this a 23 V smoothed supply for the part of the unit that delivers the power to the trains. IC1 and C3 provide a regulated 15 V supply for the rest of the circuit.
Important! You may use several PWAyMan controllers on the same layout, but each must have its own separate power supply. Short circuits will result if you attempt to operate two or more controllers from the same power supply. Similarly the controller power supply cannot normally also be used for other equipment such as track circuits, train detectors or signal drivers.
Also important! It is essential with this controller, as with any controller, to provide some form of overload protection. Its purpose is to protect both the train motor and the controller output device if a short-circuit occurs on the track or if the train encounters some obstacle while running at high speed. Some power supplies are fitted with a thermal device, which is perfectly adequate: in the event of an overload, a button pops up and the power supply output is disconnected. When the cause of the overload has been removed, you press the button and can resume your operating session. If your power supply is not fitted with one of these, you could add the electronic cut-out shown later on this page.
The circuitry around around VR1, T3 and IC3 forms a simple closed-loop controller. You could build and test this before proceeding with the rest of the circuit. Ignore T2 and T4 for the moment. The output device must be mounted on a heat sink as it is called upon to dissipate appreciable power. If you cannot obtain a TIP147, any PNP power Darlington will do; alternatively you could use a discrete general-purpose PNP transistor and a PNP power transistor connected in Darlington configuration.
The circuitry around IC2 generates the pulses. IC2 is a 555 timer set up as an interval timer adjustable between 0 and 10 ms. A new timed period is triggered whenever the output from the power supply (on pin 2) drops to zero; with a 50 Hz mains supply and full-wave rectification this will happen every 10 ms. If your mains supply is 60 Hz, it will happen every 8 ms and you may need to adjust the timing chain components (R2, C2). This description assumes a 50 Hz mains supply. During the timed period the output of IC2 (on pin 3) is high, supplying bias via R5 to T2 which clamps to ground the control voltage being fed to the closed-loop controller. The timed period, then, corresponds to the spaces between the power pulses that the controller delivers; as the timed periods get longer, the power pulses get shorter. During the triggering of the 555 timer its output is high for about 1.5 ms. This is useful as it ensures that there will always be a brief period between power pulses when the motor EMF can be monitored.
This is the circuitry around T4, T5, T6 and T7. Quasi-Darlington pair T6 and T7 sample the motor EMF via R8 and charge C4 up to this voltage. T4 and T5 conduct during controller power pulses and clamp the input of T6/T7 to ground during these so that only motor EMF is sampled. The values of C4 and VR3 are critical for accurate holding of the monitored value. Speedometer M1 is optional but adds interest to operation of the controller. If it is a 100-microAmp type, it is easy to pretend that the units of calibration are scale miles per hour, the maximum of 100 being suitable for a steam-era or early diesel-era layout.
This is the circuitry around IC4, the 741 operational amplifier. Set up as a non-inverting amplifier with a voltage gain (determined by R10/R11) of about 5, this compares the sampled motor EMF via R12 with a proportion (set by VR2) of the closed-loop control voltage. Its output is fed via R13 to T1 in the 555's timing chain. So, as train speed rises relative to the setting of the speed control, so does the 555's timed period, reducing the pulse length and therefore the power supplied to the train. The combination of R12 and C6 introduces a slight delay in the response which is amusing to watch: as the train enters a down gradient, for a moment it accelerates and then on go the brakes! Reduce the value of C6 if you prefer a more rapid response.
The pair of LEDS (D2/D3) is optional but, if they are a well-matched pair, useful in set-up. They flash alternately, D2 during power pulses and D3 during spaces, although both will appear to be on together. With a train cruising at moderate speed (no gradients or other features to affect its progress) adjust VR2 until the two LEDs are equal in brightness or until a voltmeter on pin 3 of the 555 indicates 7.5 V; this corresponds to 50% duty cycle, that is, power pulses and spaces are of equal length. You should now find that the speed compensation works well, speed being reasonably constant on gradients and tight curves. If you grab hold of the train and give it a helping pull on its way, you may find that it stops altogether and then gradually resumes its original speed! With the train running at an estimated 30 scale miles per hour, adjust VR3 until the speedometer indicates 30. You should now find that its reading is proportional to speed and that it goes to zero if the train is lifted off the track. You may need to adjust VR3 for different types of train, although on the prototype one setting suited a remarkable variety of types. If you have track circuits (see the page on TEKTOR) you may get a spurious speed reading, caused by the track circuit voltage, in the absence of a train; when a train is present, the track circuit voltage is short-circuited away.
Performance is generally excellent with a wide range of locomotive types. There is no buzz or evidence of motor heating because at low speeds the pulse amplitude is also low. The only disadvantage is that the maximum output is 22 V at 50% duty cycle, equivalent to 11 V continuous; this makes the top speed a little lower than on some controllers, but only those who run their trains at full speed are likely to notice this. Closing the optional switch S1 disables the pulses, turning the unit into a pure d.c. closed-loop controller. This gives a higher top speed, but at the expense of the speed compensation and the gentle starts and slow running associated with pulsed power. When used in this way the controller is not compatible with live-rail track circuits (see the page on TEKTOR); in its normal mode it is compatible with them.
The circuit shown above is an add-on for PWAyMan to provide overload protection. (It can also be adapted for use with some other types of analogue electronic controller.) It shares the PWAyMan's 15V regulated power supply. Two small modifications to the PWAyMan controller are needed. Firstly R1 in the circuit above is inserted into the return side of the PWAyMan's output circuit (before the reversing switch) so that all the current flows through it. Secondly the collector of T3 in the circuit above is connected to the base of T3 in PWAyMan.
The value of R1 depends on the current level at which you wish the cutout to trip. For a typical trip current level of 1A it needs to be 0R7 (that is, 0.7 Ohm). If this component proves difficult to find you could connect three 2R2 resistors in parallel, giving 0R73 which is near enough. On one occasion I retrieved a length of element from a failed toaster and cut a short length of this to give a suitable resistance - the only difficulty was soldering to it! By adjusting VR1 you may select a higher trip current, e.g. 2A.
When an overload occurs, current in R1 raises a voltage across it high enough to bias T1 into conduction, lighting the overload indicator, LED D1. T1 also supplies bias via R4 for T2 which conducts, providing an alternative bias source via R5 for T1 so that the overload condition is self-sustaining. T2 also supplies bias via R6 for T3 which short-circuits the control voltage in the PWAyMan, so that its output is reduced to zero, protecting the train motor and the output device. When you have removed the cause of the overload, switch the power supply off, wait a few seconds, and switch it on again to reset the cutout circuitry.
© Roger Amos 2012