High-Performance Throttle

by Rich Weyand, with Bill Pistello & Bill Reid

This model railroad electronics article originally appeared in Mainline Modeler magazine in the February 1995 issue. Excerpts of it will also be published in N-Scale magazine sometime in 1995 or 1996. We would like to thank Bob Hundman of Hundman Publishing, publishers of Mainline Modeler and N-Scale, for being so supportive of our projects and encouraging us to make these articles available on-line.

Information on ordering a commercial kit or assembled and tested unit of this circuit is available from the TracTronics Price List.

A Modular Electronic Control System

This is the twelfth and last article in our series describing a set of electronic building blocks we have designed to control our model railroad layouts: Rich Weyand's N&W Pocahontas Division, Bill Pistello's Union Pacific, and the Reid brothers' Cumberland Valley System. This article will finish the series with a description of a high performance throttle which delivers excellent slow speed control without heating or damage of the motors of our locomotives.

As a side note, the circuits AutoBlock, AutoSearch, and WhichWay, from our last three articles on ABS, APB, and CTC signaling, are now also available as commercial products from TracTronics, 1212 S. Naper Blvd., Suite 119, Naperville, IL 60540.

The Throttle Problem

Throttles are the ultimate model railroad control problem. The train is the center of our model railroad world, and the locomotive gives it life. The control of a model locomotive electric motor at the slow speeds required to simulate real locomotive operation and maintain the illusion of a real railroad is a serious challenge. One way to meet this challenge is through the technique of pulse-width modulation, or PWM.

Pulse-width modulation solves two very real problems in throttle design. When one chops the voltage to the motor with the PWM technique, pulsing the motor with the full voltage of the throttle, slow speed performance is excellent, as the pulses continuously bump the motor over the detents of the magnetic field in which the armature resides. In addition, the output transistor of the throttle runs much cooler than it would supplying a steady DC voltage. The transistor is either on or off: the voltage across the transistor is zero when it is on, and the current through the transistor is zero when it is off. The power dissipation in the transistor, the voltage times the current, is therefore zero except during the actual switching transition, and the transistor runs very cool.

However, motors run hotter when driven with the PWM technique. Some motors run much hotter, hot enough to damage the motor when run at slow speeds for extended periods. These effects are very dependent on the way the PWM technique is implemented, the peak pulse voltage, the actual wave shapes of the track voltage and current, the type of motor, the way the locomotive is used, and a myriad of other parameters. But the very real prospect of destroying the mechanism of a prized switcher, which with its continuous slow speeds is in the most danger from a PWM throttle, made us want something better.

What we wanted was a throttle which would give us the slow speed control of a PWM throttle, but not cause heating of any kind, even in a switching locomotive which might run almost continuously at 5 to 10 scale miles per hour with 10 to 20 cars in tow during an operating session which could last five to six hours. And we did not want to give up walkaround control, memory, momentum, push-button or rotary knob operation, and four-wire connection from our hand units, or the ability to use radio control hand units as described in the August 1994 Mainline Modeler.

Origins Of The Circuit

The circuit we started with appeared in Model Railroader magazine in January 1986, in an article by Kirk Wishowski. Kirk and his friends had come up with a throttle for their HO scale club which met most of our requirements. When we built the circuit and used it for our N scale layouts, however, we had some problems. The memory did not last very long, and locomotives would noticeably slow while we moved from one throttle plug-in to the next around the layout. Also, the performance of the throttle seemed to be dependent on the particular parts we used. In particular, swapping the two output transistors for identical parts caused different throttle characteristics. This is a taboo in electronics design, as a circuit should always work the same when identical parts are used.

Worst of all, the output of the throttle did not always go to zero when the input was set to zero. For our little N scale engines running with Faulhaber or other high-efficiency motors, this was a big problem. It is very disconcerting to watch your locomotive happily steaming down the rails with the throttle emergency brake held down.

What was most disappointing about the problems was that this throttle had the finest slow speed control we had ever seen in a throttle that did not use PWM. Even notoriously poor performing locomotives ran beautifully with this throttle, and fine locomotives were a joy to behold. And they always ran cool. After four hours of 5 mph running with twenty cars in tow, a brass switcher was stone cold to the touch.

We decided that we would try to improve on the design, solving the problems we had found using it for our N scale layouts. After nine years, many experiments, and many destroyed electronic components, we came up with a modified circuit which greatly improved the memory, performed in a reproducible way, and brought the output voltage all the way to zero when the input was set to zero. In addition, we improved the slow speed performance over the impressive performance of the original circuit.

It's important to note that during this nine years of experimentation we never smoked a locomotive. Variations on this circuit have been used for as long as nine years on dozens of model railroads, including those of Bill and Wayne Reid, Bill Denton, Mike Hurlburt, and Bill Pistello, all of whose layouts have appeared in the hobby press recently.

The Base Unit Circuit

The current throttle base unit circuit, which we call CoolerCrawlerTM, is shown in Figure 1. This circuit is very difficult to explain; it is in fact the most difficult of the series, but for those who want to know:

Capacitor C1 is the speed control capacitor; the voltage on this capacitor determines the output of the throttle. This capacitor provides the memory and momentum for the throttle circuit. Transistors Q1 and Q2 are wired as a Darlington pair, and provide amplification and wave shaping. Q3 is the pre-driver transistor and Q4 is the driver transistor. Note that Q3 is internally a Darlington pair also. Q4 is the output transistor, and is mounted on a sizable heat sink to dissipate the substantial heat generated due to not running Q4 in a PWM switching mode. Capacitor C2 is a wave shaping capacitor, and ensures that the output waveform is smooth.

Relay 1 is the reversing relay, with D3 as the anti-surge diode and R5 as the latching resistor. D4 is a zener diode which limits the output current by spilling base current off of Q3 when the voltage drop across Q3, Q4, and the output series resistance R3 is high enough. R4 is the output parallel resistance, which keeps the output voltage from floating when there is no locomotive connected. Note that diode bridge DB1 and diodes D1 and D2 generate two separate positive voltages, one for the bulk of the circuitry, and one for the output transistor.

Building the Base Unit Circuit

The circuit etch patterns and component placement diagram for the CoolerCrawler circuit are given in Figure 2 for those who want to etch boards, rather than perfboard the circuit. The component placement diagram includes the hole locations to aid in drilling your board. The component listing is given in Table 1. Note that the etch pattern is always printed as seen from the component side of the board, per electronic industry standards. The solder side image must be reversed on the board you build, so that the text and the image are correct.

A large area has been left on the board for the heat sink for transistor Q4. In general, the larger heat sink you use on this transistor the better, as long as it fits in the space. The heat sinks listed in Table 1 are the largest we have found that will fit in the space. Note that two sets of mounting holes for transistor Q4 are included on the board, one parallel to the board edge, and the other at 45 degrees to the board edge, to accommodate different heat sinks. Mount the transistor with #6-32 x 3/8" machine screws and #6-32 nuts. Use heat sink paste (Radio Shack 276-1372) between the transistor and the heat sink to ensure good heat transfer to the heat sink.

Please be very careful to install C1 and C2 to match the polarity indication in the component placement diagram; electrolytic capacitors will explode when power is applied if they are wired backwards. Also be sure to install DB1 to match the polarity indication in the component placement diagram; installing this component incorrectly will result in frying both capacitors and all four transistors!

Wiring the Base Unit Circuit

The pinouts for the base unit are given in Table 2. The leads from the transformer are connected to XFRMR1 and XFRMR2, pads 2 and 3. We normally use Radio Shack 273-1511 as the 12.6 volt 3 amp transformer for this throttle. Do not use the center tap transformer lead, but wrap the end in electrician's tape or heat shrink tube and leave it unconnected. Be sure to fuse the 110 volt side of the transformer with Radio Shack 270-1238 or similar.

The leads to the track are pads 6 and 7. Pad 6 is the common rail lead to one rail of the entire layout. When multiple throttles are used, these leads from all of the throttles are connected together, and connected to earth ground, such as a water pipe. Pad 7 is the throttle lead to the track, and should be connected to the cab selector controls for the track blocks.

Pads 5 and 9 are the reverse loop leads. These leads will not reverse direction when the throttle direction is reversed, and should be used to power the track of any reverse loops on the layout. Note that the cab selector controls for reverse loop blocks must switch both of these leads from each throttle. These leads cannot be wired in a common rail configuration for the reverse loops.

Pads 9 and 10 are the ammeter leads. If an ammeter is desired, the trace between pads 9 and 10 must be cut on the board, and pads 9 and 10 then connected to the ammeter.

Pads 1, 4, 5, and 8 are the SPD, DIR, GND and POS leads to the hand unit. These leads are the four throttle control leads we described in our August 1994 Mainline Modeler article on walkaround throttle upgrades. Figure 3, reprinted from the August article, should be familiar; compare this with Figure 1 for the throttle base unit.

Hand Unit Options

There are several options for walkaround hand units to control the throttle base unit. The three main options are: rotary potentiometer control, push-button control, and radio control. The point to keep in mind when building the hand unit is that it only does two things: control the speed of the locomotive by controlling the voltage on capacitor C1 of the base unit, and control the direction of the locomotive by momentarily connecting the coil of Relay1 of the base unit to either positive voltage or ground.

The schematic and components list for a hand unit with a rotary potentiometer is given in Figure 4. This hand unit contains a top speed limiting trim pot, as well as a momentum adjust trim pot. We prefer this type of control for our yard cabs, as the rotary knob makes it easy to set switching speeds, and we can then use the service brake to control locomotive speed. The emergency brake is a good safety extra, as it bypasses the momentum of the rotary knob. You may wish to experiment with the values of resistors R1, R2, R3, and R6 to get the throttle and brake response you want. Note that a larger resistor value will slow the action of the corresponding control, and a smaller resistor value will speed it up.

The schematic and components list for a hand unit with push-button controls is given in Figure 5. We prefer this type of control for our road cabs, nudging the speed up and down with the push-buttons. The emergency brake on this unit also bypasses the momentum feature. Once again, you may wish to experiment with the values of resistors R1 and R2 to get the throttle and brake response you want.

Radio Shack project box 270-220 makes a nice-sized hand unit enclosure for the rotary potentiometer and push-button hand units. We do not use any PC boards for this, we just wire the components directly to the pots and push-buttons. Various push-buttons available from Radio Shack can be used for the direction, throttle, and emergency brake buttons. You should experiment with different control layouts on the box to get the feel you want for the hand unit. Any four-strand wire and connector can be used for the hand unit tether and the fascia plug-ins. Be sure to strain relief the cord where it enters the hand unit!

The radio control hand unit was described in the August 1994 Mainline Modeler, and we refer you to this article for information on using radio control to control the throttle base unit. We like the radio control hand unit for road cabs, but the extra expense of a good RC transmitter and receiver pair means it's not for everybody.

Throttle Operation

The output voltage waveforms of the throttle for various throttle settings are shown in Figure 6. Note that the 60 hz pulses of the throttle are rounded, not square-edged. While these rounded pulses result in a bit of heat dissipation in the output transistor Q4, they also result in locomotive motors running very cool, while yielding excellent slow speed operation. Some motors will make some noise due to the rounded pulses of the output signal, but no motor we have tried will heat up when run with this throttle, even for slow switching speeds with heavy trains for hours on end.

Maximum throttle current with the design shown here is three amps, although we are also working on a six amp unit. The throttle design will limit the output current to three amps even into a shorted load, regardless of the current capacity of the transformer used. When running at high currents for extended periods, Q4 and its heat sink will heat up quite a bit. When the output is shorted for extended periods, R3 will also get quite hot. These two components should be located away from little fingers and flammable objects, and where free air circulation will aid cooling. To aid air circulation and cooling, do not mount these throttle boards in an enclosure.

A Note On Commercial Availability

Bill Pistello has been selling complete assembled versions of this throttle, including hand units, for over five years through his company, Modular Model Railroad Electronics. The newest version of this throttle with all recent enhancements, including the SafeTrain reverse blocking circuit and the RadioTrain RC interface from the August 1994 Mainline Modeler, should be available by the time you read this. Complete assembled throttles, including your choice of rotary, push-button, or RC hand units, are available from MMRE, 625 South Princeton, Villa Park, IL 60181.

For those who prefer to assemble this throttle from a kit, or to purchase the base unit assembled and tested but build their own hand units, both kits and assembled & tested throttle base units are available from TracTronics, 1212 S. Naper Blvd., Suite 119, Naperville, IL 60540.


The throttle we describe here, improved and enhanced over the last nine years, gives us excellent low speed performance without the disadvantages associated with pulse-width modulation techniques. The performance of all our locomotives is dramatically improved with this throttle, and visitors to our layouts never fail to notice the fine operation of our engines. Most guess that our more modest performers must have been re-motored to achieve this performance, but re-motoring has been made unnecessary by this throttle.

With this twelfth article we conclude the series, as we now have all of the electronics modules which we need to control the switches, signals, and locomotives on our railroads. The article series has been very productive for us in meeting our electronics needs, and we hope you have enjoyed reading these articles as much as we have enjoyed bringing them to you. Happy railroading!

Index to the Electronics Articles

For those of you who may have come in late, the electronics articles of this series appeared in these issues of Mainline Modeler. Back issues of these magazines are available from Hundman Publishing.

Copyright 1992 - 2012 by TracTronics, Incorporated.
Last updated: December 28, 2011
These pages maintained by: Marc Haas