One Mans idea of a battery-electric loco.

Juggernaught started off as a simple 0-4-0 shunter, a quick and easy job to improve my engineering skills and get something on the track while I thought about building a steamer. However, in typical Ken fashion, while thinking through the design it slowly grew into a 1.7 metre long, 200Kg Co-Co .

I spent some time trying to decide how much power was “enough” to cart the loco and driver around the track at a reasonable speed.

Then of course, the question had to be answered, “what was a reasonable speed?” After some enquiry’s around the place, it seems that speed limits at tracks vary quite a lot, from a low of 7kph to a high of 12kph. I decided that a maximum speed of 15kph would cover all eventualities (and satisfy my “need for speed” when no-one was looking…). I had already decided on 4 inch diameter wheels, so given the maximum rpm of the motors, a gearing of 3.5:1 would give me 15kph.

So now back to the question “how many motors, and what power?” Being a great believer in “you can’t have too much overkill”, I eventually decided on four, 300watt motors, two in each 6-wheel bogie.

The next question was the batteries. A search through manufacturers data, along with looking at as many pictures of full-size co-co’s as I could find showed the Class 55 “Deltic” as having the longest fuel tanks in relation to overall wheelbase that I could find. That gave the proportions to allow me to use a pair of 75 amp-hour deep cycle batteries sitting low in the dummy fuel tanks, right in the middle of the loco. A low center of gravity was important to me, because, as those who knew me in my earlier (motorcycling) days know, I don’t like slowing down much for the bendy bits..

An early decision was made to make the loco out of aluminium for ease of manufacture and so I could use good sections of material to give rigidity without excessive weight. This decision came back to bite me later when the loco needed extra weight to give enough traction to suit its power, but it has resulted in a low centre of gravity that allows for some interesting cornering speeds…on Juggernaught you don’t need to slow down too much for the bendy bits!

I didn’t like any of the designs I saw for the bogie pivots so I spent a long time thinking about it before coming up with the ones I made for Juggernaught. Those of you familiar with the “fifth wheel” coupling on articulated trucks (semi-trailers) will see where my design came from. The result is a very stable locomotive with very little “lean” on corners. Another aspect of the couplings is the use of teflon bushes for both axis, they need no lubrication and allow the bogies to pivot with very low friction.

I had a look at various sources to find out what materials go best with aluminium as a bearing surface, and cast iron seemed to be the go, so the axle boxes were machined out of some 3 inch diameter, continuos-cast, cast iron. As this was my first real go at machining things to any sort of tolerances, I didn’t trust myself to make 12 bearing pockets to the fit needed for pressing in ball races, so the steel axles run direct in the cast iron. I machined oil galleries in the axle boxes, arranging the bottom one to just break in to the axle hole at the bottom. I then filled this gallery up with pipe-cleaner to act as a wick for the oil. The top of the axle box was hollowed out to form an oil reservoir and two 1mm holes were drilled in the vertical sides of the slots that the frames slide in, so some oil leaks out to lubricate the sliding surfaces.

Of course, me being me, I forgot to oil one axle one day, this coincided with a day when we had guests at the club, so were running round all day with lots of passengers. By the time we noticed it, the axle boxes were chewed out to about half the axle diameter- yet Juggernaught didn’t even notice the problem, it just kept running as if nothing had happened! So now two axle boxes are running in ball races, and 10 in the original plain bearings.

Because the motors are so close to their respective axles, the difference in chain tension between the extremes of suspension movement meant the chains slipped under load frequently so I had to devise some chain tensioners. These were complicated by having to operate on both runs of the chain, depending on whether the loco was going forward or reversing

In the end I used a “scissor” arrangement, with small sprockets running on the outside of both runs of the chain, with a tension spring pulling both sprockets towards each other. No matter which run of the chain is slack, chain tension is maintained at all times.

To make chain lubrication easier, I drilled a hole in the bogie frame and pressed in a small brass tube, which points downwards and finishes just above the bottom run of the chain. Now all I have to do is squirt some oil in to the tube which then drips on to the chain and keeps it well lubricated.

Of course, all this oil floating around the underside of the bogies makes for a bit of a mess (it won’t rust easily though..) so I fitted oil drip trays to the bogies, which also double as protection for the motors in the event of a derailment.

I originally wanted to make the bodywork to resemble the class 55 Deltics on which Juggernaughts dimensions are based. After many abortive attempts to find the dimensions, or to scale it from pictures, I eventually just sketched out my own design and had a local sheet metal place fold it up from 1.6mm steel sheet. A long time after, when it was far too late to do anything about it, I discovered that not one single dimension or angle was as I had drawn it- that’s what comes of trusting these so-called tradesmen. If there’s ever a next time, I will be checking the job with a ruler and protractor before accepting it.

Before I launch in to the description of the internals, a little nomenclature- there are three types of “controllers” in this loco. The first are the ones that directly control the current through the motors, there are four of these and I will refer to them as the motor controllers. Next is the PIC microcontroller-based controller that sits in the loco and tells the motor controllers what to do, as well as turning on (and off) the lights, horn, engine sound and brakes. I will call this one the loco controller. Lastly is the hand controller, the bit you hold in your hand to tell the loco what to do, this I will call the hand controller.

The motors and motor controllers came from Oatley Electronics, the motor controllers are “switch-mode”, hence are very efficient and seldom get even warm, let alone hot. (300w motors) (motor controllers).

At first I made a simple hand controller just using a potentiometer and a couple of resistors but quickly became dissatisfied with the response, there being a very pronounced dead band at the beginning of the pot’s travel. It also required too many wires in the cable to the loco, with power and ground, plus one wire for each function that I wanted to control. As I spend a lot of time playing with PIC microcontrollers, this was an ideal opportunity to mix two of my hobbies. Since then the hand controller and its software has been through innumerable iterations, so I will settle for describing the current “wired” setup. (I was nearly going to call it the “final” setup- I don’t think that I will ever truly finish the software, there’s always something to change or improve on!)

I used an old model aircraft radio-control handset, throwing away all the electronics and substituting my own. One joystick is spring-loaded to return to centre, this is the forward/reverse speed control and the other joystick is modified to get rid of the spring-loaded return, this is the dynamic braking control. The idea is that when “hands off” the speed control springs to off and the brake control stays where you last put it.

The motor controllers drive the motors when the control voltage is between 1.4 and 3.5 volts, hence with a simple potentiometer control there is a lot of “dead” travel. The pic microcontroller is programmed to translate the joystick movement into this voltage range, as well as sensing when the reversing relays need operating. When the braking joystick is operated the pic disables the motor drive and shunts the motors via a relay to the dynamic braking circuits in the loco controller. The pic also provides the pulse-width modulated signal to these circuits. It sounds complicated (and it is) but it is completely transparent to the user- which is how it should be, the object of the game is to enjoy playing trains, not handling complicated control boxes with multiple, non-intuitive controls.

Another important part of my philosophy is to keep high current leads as short as possible, and have only low current leads going to the hand-controller. In it’s “wired” configuration, the hand controller only uses 5 light-gauge wires between itself and the loco, using a 9600 baud serial data link between it and the loco

Charging the batteries took a bit of thought, as I wanted to use a good deep-cycle battery charger (Altronics kit # K1675) to look after the expensive batteries, and this meant charging at 12 volts while the loco’s system voltage is 24 volts. I also didn’t want to have to keep taking the bodywork off and connecting and disconnecting the batteries whenever I needed to charge them.

The solution I came up with uses some 50-amp “Anderson” unisex connectors (Jaycar cat. # PT4420), 3 in the loco, 3 in the operating plug and 2 on the charger plug. Each connector has 2 poles, the loco wiring goes to one pair of poles, and the 4 leads from the batteries go to the other 2 pairs. It’s important to make sure one of the three connectors is fitted in the opposite sense to the other two, so you cannot accidentally plug the charger in incorrectly.

The removable plug is wired to put the batteries in series and connect them to the loco, and when the plug is removed the loco is completely dead, a useful safety feature. The connector on the charger cable connects the two batteries in parallel for 12 volt charging but does not connect to the loco wiring at all. Thus, charging is only a matter of pulling out the operating plug and plugging the charger plug in in its place. Very simple, and only requiring a 3 inch by 2 inch hole in the bodywork which is easily covered by an embellishment of some sort such as exhaust stacks or air intake, etc.

One important point when fitting the connector blocks together in is to not clamp them tightly to each other, I found out the hard way that it makes them extremely difficult to pull apart. It’s much better to leave a millimeter or so of slack so they can self-align when being connected. This tends to leave much more skin on your knuckles when pulling the plugs out.

As mentioned earlier, I use dynamic braking on the loco. Originally I was going to fit “normal” air brakes, but there just isn’t room under the bogies for the equipment.

The dynamic braking is simple in principle, short the motor leads together and the motor stops in an instant- or at least, it tries to!

In practice I use pulse-width modulation to apply a brief short circuit to the motors, 800 times a second. The length of time the short is applied is varied to vary the effective braking effort. The end result is braking of a high order with no mechanical complications in the already-crowded under-bogie space.

The serial data link from the hand controller to the loco controller means, in principle, I can have almost as many switches, buttons and knobs as I like on the hand controller to operate things in the loco. In practice, there is a limit but it is far, far more functions than any practical loco would ever require.

The controls so far are forward/reverse, braking, engine sound, horn, and headlights.

There is one extra control on the hand controller that uses it’s own pair of wire to operate, and that is the dead-mans-handle. There is a button that has to be pressed at all times the loco is operating. This operates an 80 amp relay in the loco that carries all the current for the loco. If the button is released, the relay drops out and the loco coasts to a halt. I originally did not have this in, but a small software glitch while I was developing the software caused a runaway and convinced me instantly of the need!!! You know the old chinese curse “may you live in interesting times”—well that was a VERY interesting time!

Instrumentation consists of an expanded-scale voltmeter, speedo and ammeter, all mounted where the rear cabs window would be. As you may expect from me by now, none of them are quite as simple as they appear.

The expanded-scale voltmeter reads from 21 volts to 27 volts, making it much easier to read. There is no point in reading below 21 volts as the motor controllers automatically cut out at 21.5 volts to preserve the batteries.

The speedo is driven by a pulse generator consisting of a slotted disk and photo-interrupter, running off the center axle on the rear bogie. This was a little tricky to make as there is not a lot of room under there. The pulses go to another PIC microcontroller mounted on the back of the speedo. This PIC counts the number of pulses per quarter second and converts this number to a current to drive the meter. Using this method, I could calibrate the speedo on the bench and know when I fitted it to the loco, it would be accurate. When Juggernaught says you are doing a given speed, you are doing that speed +/- ½ kph

The ammeter was a little tricky, as I wanted to use matching meters for the voltmeter and ammeter. The ones used are rather insensitive, needing 40 milliamps for full-scale deflection. The shunt resistor I am using in Juggernaught has a 1 millivolt drop per amp passing through it, so I needed to amplify a very small voltage to drive the meter. This requires some very careful work on the circuit and I am still not too happy with it. In real terms, it is probably plenty accurate enough, but I know it is not quite right and that bugs me..

Juggernaughts sound system is based on the diesel sound simulator published in AME as part of the 422 series- except of course (as you might expect by now) it is somewhat modified. I spent some time playing with the output to get the tone right, and fitted an adjustable schmitt trigger circuit to the throttle input of the sound generator. Thus, as the throttle is cracked open, the “engine” runs up to operating speed and stays there regardless of loco speed, until the throttle is released, when it slowly drops back to idle. This more-or-less follows full size practice- the alternator on a full size loco of this size runs at a constant speed when operating normally, and cannot instantly drop to idle, the large rotating mass takes a finite time to spin down. A further development when I have time, is to have the tone of the sound “deepen” in response to the load on the motors, again, to mimic the full size locos.

The horn is currently a standard car air horn, but an electronic version is on the bench, this uses a home-made mp3 player driven by yet another pic microcontroller. I am trying to develop this to play a real diesel loco air horn sound, mixed in to the same audio amplifier and speaker that the engine sound uses. However, this is still very much a work in progress, having been put on the back burner while I was getting Juggernaught ready for our annual play-fest in Adelaide.

Juggernaught has lived up to my expectations as far as performance is concerned- battery life is extremely good thanks to switch-mode controllers and short, heavy-gauge, low-resistance wiring in the high current circuits, it corners rather well, and goes like a cut cat when required.

Further developments already planned are a wireless camera in the cab, 10 new, ball-bearing, axle boxes, and four new H-bridge PWM motor controllers to get rid of the reversing and braking relays and their attendant wiring. Software changes planned are to make the throttle a constant-speed control, so you can set a speed and have the software control engine power or braking to keep the speed constant regardless of load and track gradient.

Number of visitors to this page so far:

Visitor counter
Copyright 2007-07-15 Ken, VK7KRJ.