[?]

The EMD SD45 was manufactured starting in 1965 so that it represents one of the newest prototypes available in Large Scale. SD45's are still running all over the country, but they are getting fewer over time due to very high miles in the last 30 to 35 years.

The SD-45 was the first locomotive that used an engine with more than 16 cylinders. The 645 model engine in the SD45 carried 20 cylinders. The "645" in the model number indicated the displacement in inches of the engine PER cylinder. This 2 stroke engine developed 3600 horsepower, the largest on rails at the time it was manufactured. The heat generated by the engine was so great that extra radiator area was needed. The wider radiator resulted in the flares on the rear of the long hood. Cooling air was taken in through the side vents and exhausted out the top. This technique was borrowed by other builders later to accommodate larger radiators.

There are two major variants of the SD45, the SD45-2 and the SD45T-2. The variants had all new electricals and controls. Both rode on an extended frame and lacked the sloped air intakes. The radiator was extended in a longer hood. The SD45-2 had its air intakes along the top of the hood. The SD45T-2 (tunnel motor) took in air through vents on the sides set down near the walkways. There was nothing inside the hoods in this location and you could look right through the loco and out the other side. This arrangement was developed by the Southern Pacific so that the engines could draw in somewhat cooler air in tunnels. The last unit of a MU consist would often suck in superheated air in a long tunnel because of the blast of engine exhaust from the lead units. It could then overheat and shutdown with disturbing results, such as a stalled train in a tunnel. The "tunnel motors" with the relocated intake grills had a better chance of drawing in cooler air from nearer the track and tended to hold up better in long tunnels and snowsheds. The SD40 and SD40T-2 had the same frame and cooling arrangement as their more powerful brethren, but carried a 16 cylinder model 645 engine which developed 3000 hp. Eventually, the higher maintenance costs of the SD45 lead the railroads to purchase far more SD40's than SD45's.

• The Aristo SD45
• Derailment Issues
• Pulling Power
• Mounting Kadee Couplers
• Lighting
• Smoke
• Scale and Clearance
• Disassembly of the SD45
• The SD45 Power Brick
• Wiring in the SD45
• DCC Installation
• Sierra Sound Installation9 Jun 02

The Aristo SD45 is a 1/29 scale model of a standard issue SD-45. These locos were found on railroads all over the country.

The sloping air intake grills are clearly visible in this photo. This is a major spotting feature of the SD45.

The engine is heavily weighted, stock it comes in at 18 lbs, and has the pulling power to match. See the section on pulling power below for the details of my initial round of testing. There is a long steel channel that runs the length of the frame and 3 lead weights inside the fuel tank. These lead weights can be removed if the weight is not desired by removing the tank (four clip tabs accessible from inside) and then removing 3 nuts on three bolts that come up through the frame. Each nut is secured with a locking compound that must be cleaned off the threads before the nuts can be removed. Once the weights are out, there is ample room in the tank for a large battery pack.

The engine is shipped without the handrails installed, they must be installed by the user. The handrails are all metal. Once the handrails are installed, the loco won't fit back in its shipping box.

The engine is equipped with a new smoke unit that does a good job at very low engine speeds. The lighting is constant intensity as well. There are battery/MU connectors at each end of the loco, but to connect two locos together, an adaptor connector (not supplied) is required. Internally, the unit has a socket designed to accept a DCC decoder. However, nobody yet has marketed a decoder or R/C RX that plugs right in.

Right out of the box, the engine made some pretty loud gear noise. It did quite down quite a bit during break in. All twelve wheels pick up power and none have traction tires. With the very heavy weight of this loco, it clearly doesn't need traction tires. Power pickup was quite good even on dirty track. This may be partially due to the weight of the loco which can allow the wheels to bear through minor track contamination.

The locomotive tracks very gracefully. I had seem some comments about how the unit "glides" through turnouts but I didn't understand it until I saw it in action. It tracks through turnouts without any sign of being on a turnout, no jerks, no dips, and no derailments. Each axle has considerable sideways slop to allow the long wheelbase trucks negotiate curves without binding. The method seems to work very well.

The slow speed performance of the unit is outstanding. It is very steady at very low speed and it is easy to control. When it gets DCC installed, it will be exceptionally smooth. The only other loco that runs this well at low speed is the Bachmann Shay and it became an incredible performer when it got DCC (and better power pickups). The Shay might make a better switch engine than the SD45 only because the truck mounted couplers allow it to couple better on curves. Other than that, the SD45 is an outstanding switcher as well as an exceptional road engine.

There are four switches to control the electrical functions of the loco. These are located under the dynamic brake pods/fan assembly on the top at the center of the long hood. Be sure that you don't lift the loco by the pods, they'll come off and you'll drop the loco. Pull the pod assembly straight off to reveal the switches. If the pod assembly doesn't come off straight, the clips will bend. Then they have to be straightened to get the pod assembly back on. I removed the spring clips from mine as I get into that area often to turn the sound system on and off and I got tired of straightening them. The pod assembly seems quite secure without the clips.

The POWER switch controls the connections to the track. If you run battery power of any kind, TURN IT OFF AND LEAVE IT OFF. If you don't, you'll get smoke from somewhere. If you install internal batteries or never intend to run the engine from track power ever again, drop a glob of hot glue on the switch so that it doesn't get flipped accidentally.

The MOTOR switch disconnects the motors in all modes of operation, even if a DCC decoder or R/C RX is installed provided that the installer actually used the provided DCC decoder connections. If you have a custom DCC or R/C installation done, your installer may elect to bypass Aristo's motor wiring so that the switch may or may not continue to function.

The SMOKE switch is pretty obvious. It turns the smoke unit on or off. The smoke unit itself is designed to protect itself in the event that it runs dry so it is not really necessary to turn it off, but if you don't use smoke, there is no need to power it. The smoke unit runs directly from the internal power bus which is permanently connected to the battery connectors at each end of the loco. When the Power switch is on, this bus is connected to the track.

The LIGHT switch turns on or off all of the lighting. It disconnects the lighting loads from the internal lighting power supply, but it leaves the power supply itself connected. This can be a problem for DCC installations, but a remedy is described below in the DCC section. If you have a custom R/C or DCC installation done, the installer may elect to wire some of the lighting from another source so that this switch may or may not affect those modified functions.

[ Top ]

I noticed a derailment problem with this engine as it came out of the box. The engine itself tracks very well and didn't derail at all during my short period of testing. However, when hauling a long train, the first car would often derail on a curve. This is a common problem that occurs with cars with truck mounted couplers when being pulled by a loco with body mounted couplers. The sideways movement of the engine's coupler while the engine is in a curve pushes sideways on the car's coupler and twists the car's truck clean off the rails. The coupler spring on this SD45 is pretty stiff and this may be contributing to this problem. Aristo has softened up the spring on current production units so this may or may not continue to be a problem. In any event, a softer spring can be bent from a smaller gauge of music wire.

[ Top ]

I ran my standard tractive effort tests on the SD45, and as expected it did quite well. As a matter of fact, it almost did too well. I kept bringing out more and more cars and it kept going and going. I nearly ran out of grade before I decided that the engine had had about enough. The train was about 56 feet long. See my Tractive Effort Tests page for more details.

The SD45 testing revealed that it was able to pull about as many cars as the very best previous puller that I have tested, the Bachmann Shay. I brought the DCC equipped Shay out for a retest and found that it was still doing well. Both the Shay and the SD45 will accept quite a load and even if they slip trying to start a train on the grade, they'll eventually find their feet and start gripping the rails. Since I wasn't able to tell which was actually pulling better on the grade (because I had literally run out of grade) I elected to stage a tug-o-war.

I put the SD45 and the Shay on opposite sides of a track block and put a car with plastic wheels between them as the "rope." The SD45 was running from a Train Engineer in PWC mode and the Shay from DCC and I didn't want a wheelset crossing the insulated block and shorting the two power supplies together. The only car that I had left with plastic wheels was Clarabel from a Thomas the Tank Engine set. Poor Clarabel was hard put to deal with the buffing forces it saw during this test. It literally picked itself up from the rails.

Neither the Shay or the SD45 could do much to pull the other engine when the other engine was stopped. There was some movement in either direction, but not much. When the SD-45 and the Shay were both pulling away, sometimes the SD45 would win and sometimes the Shay would win. Over a series of tests, the SD45 did slightly better and it was pulling uphill. Several times, I did pull one or the other engine across the rail gap and I did short the power supplies together anyway. I heard a lot of DCC noise coming from the SD45, but neither engine nor either power supply was damaged or even blew a fuse.

After the SD45 had been converted to DCC and the bricks modified to get all the wheels resting on the track (see the section below on the SD45 bricks), I ran the tug-o-war again. However, since both locos now had DCC, I didn't have to mess with a "rope." I could couple them together directly and run them from the same booster. To my surprise, this time the Shay did a little better. It could drag the SD45 just a little and the SD45 couldn't drag the Shay. With both pulling away, the Shay could drag the SD45 with all wheels on both locos slipping. Perhaps the "improved" weight distribution actually results in slightly less traction. The Shay weighs about the same as the SD45 but carries its weight on 8 wheels instead of 12.

Since the SD45 has four motors, I wanted to see how much average current that it would draw under no load at all. I set the engine up on the bench and connected it with a Kadee contact brush and measured the IV characteristic. At voltages less than 3.5 volts, the unit is stalled so that the motors look like a resistor. As soon as the wheels start to turn, the current drops due the generation of some back emf by the motors. At higher voltages, the current tends to flatten out as it typical of small DC motors. With no load at all, the unit draws 1.8 amps. During my pulling power tests, the average current ran near 3.5 amps. It'll take some good batteries to keep this loco running for any period of time.

The loco was measured to as close to true stall as I could get at 16 VDC. I had to REALLY press the loco into the test track to get it near a true stall, but the wheels never did quite stop. The steady state stall current was 8.8 amps. This is pretty respectable for a four motor loco, most with two motors draw nearly as much. This could be due to either low stall current motors, or due to higher internal wiring resistance.

[ Top ]

Kadee couplers can be mounted to the SD45 in a similar fashion to the method that I also used on the RS-3 and the Center Cab. I prefer to use #831 boxes and couplers if possible because I've got a whole bunch of them. In this case the large offset coupler that comes with the Kadee #831 kit is not suitable. I used the medium offset coupler that comes in the #836 kit and the #831 boxes.

The original coupler is a little difficult to remove, but if the pilot is gently flexed, it'll come out. It will come out more easily if the 4 screws that secure the pilot are removed. There are two that screw into the frame and two that hold the front and back parts of the pilot assembly together. With these screws taken out, the pilot will still not come off, but the parts can be flexed enough to get the old coupler out and a new one in. Note that the steps may partially or fully fall out. Make sure that you've got them back in correctly before reinstalling the screws. The spring can be removed by rotating the truck so that the spring can come out between a wheel and the brick body.

All of the cutting and grinding is done on the coupler body itself. Except for trimming and bending the centering spring, no permanent modifications are made to the loco itself. For a completely reversible installation, a new centering spring can be bent from music wire and the old spring and couplers can be kept on the side.

First the pivot hole in the coupler body needs to be drilled out to 0.218". Then a 0.080" styrene shim is glued in between the side rails behind the hole. 30 additional mils of shim material is glued along the side rails of the coupler box and on top of the 80 mils already added. I also cut a shim from some 0.218" OD styrene tube to use a spacer under the mounting screw to allow it to snug down without binding the coupler. The coupler then came out at the right height in a static test. The shims and the side rails will ride on the ridges of the coupler mounting post and this sets the coupler height. A #70 hole is drilled in the back of the coupler body to accept the centering spring. Finally, the back corners of the body are ground off at a 45 degree angle so that the coupler body will have sufficient swing to avoid dragging the following cars off the track.

The coupler box may still interfere with the brake hose assembly on one side. This did cause me some problems as I still got some derailments in right turns. I twisted the brake hose slightly so that the coupler box hits the pilot assembly before it hits the brake hose. If following car derailments are still a problem, then the opening in the pilot will need widening.

With the shims glued into the coupler body, it is difficult to get it back on the post. However, if the mounting screws for the pilot assembly are removed, the pilot will flex just enough to allow the coupler body to slip by and settle on the post.

The mounting post itself is sturdier than the posts on earlier Aristo locos but it will still flex under a heavy load. When it does, the coupler will tend to rise. There doesn't seem to be much available room at the base of the post to reinforce it. The stock Aristo coupler will not suffer a problem due to this difficulty because it has a lift lock feature that prevents a coupler pair from dismating due to mismatched height.

Under load, the Kadee coupler still tended to rise just a little. A spacer was cut from 0.125" x 0.250" styrene strip and glued underneath the end of the pilot (white block in the photo) so that the spacer rides on the top of the coupler box. This locks the coupler box at the right height and it cannot rise. After the photo, the spacer got a coat of Santa Fe blue and it became nearly invisible.

Tests with a heavy train with a very light car in front indicate that this coupler mount works better than the stock one did. The test was so sensitive that I found a turnout that needed a little filing on the points and one other spot that was out of level. When these things were fixed, I could run the train either way without derailments. The coupler body would rotate all the way to the stops so that it is pretty clear that a mount with less rotation capability would cause some kind of trouble especially in "S" curves of which I have several. With the stock coupler, the a test train (with a heavy car in the front instead of a light car) would derail in several spots around the layout.

The centering spring is a little difficult to remove initially, but when cut short and formed, it goes back in much easier. You may have to play with the spring forming a little to get it right. I trimmed the stock spring. The Kadee coupler has softer springs that center the coupler in the box than the Aristo coupler so that the slightly stiff box centering spring is not a major problem with the Kadee coupler.

The coupler body sticks out from the pilot a little more than I would like, but mounting it deeper would require that the opening in the pilot be widened to allow the coupler body to rotate enough. We'll see how well this mount works out in the long run. I may remount them later to pull them back in a little more at some future date.

[ Top ]

The SD45 has an unusual lighting arrangement. The lighting power runs from an internal PWM converter so that full lighting power can be obtained before the loco starts to move. The engine starts to move at about 4 volts, and this is about where the lights have reached full intensity. This photo was taken with the motors on, but the track voltage was not quite high enough to get the loco to move. There is no perceptible difference in lighting intensity when switching between PWC and Linear power.

As can be seen from the photo, the cab and number boards are also illuminated. There are safety lights for the crew at the base of the short and long hoods and at the step at the center of the long hood on the fireman's side. There are also two LEDs at either end that show red to the rear. They are not illuminated to the front.

If you look down into the rearmost fan, a green light is visible. This is just the number board bulb showing through the green tinted circuit board.

The internal PWM power converter generates a pulse stream at about a 200 uS period with varying peak voltage (depending on the track voltage) and pulse width. The duty factor is about 10% at full track voltage and increases at lower voltage. The internal bus voltage then appears to be about 2 volts RMS. The headlights are incandescent bulbs that appear to operate at nominal intensity at 6 volts DC. However, since they are being pulsed, they tend to draw more current than they would with straight DC and operate brightly at this low average voltage.

This is a schematic of the wiring of the SD45 headlight circuits. All the lights run from the internal PWM power source. The front headlights and the rear markers are connected together as are the rear headlights and the front markers. The number board lamps and porch lamps run in both directions.

A white LED cannot be operated at its best efficiency under these conditions. LEDs are current driven devices, but they get inefficient at high current drive. One just cannot ram 10 times the current in them for 10% of the time and expect anything like the same intensity. For an LED to work properly, it'll have to be run from a current limited DC source.

Aristo applies a filter capacitor and current limiting resistor to the red LEDs used for the markers. These parts provide a constant DC bias to the LED's, but the LEDs get only about 4 mA on the average. I tried to run a white LED with a 100 ohm limiting resistor and an appropriate filter capacitor to get about 20 mA of average current but it didn't work too well, the LED pulsated rather badly.

In my DCC installation, I just wired the power side of the two LEDs in series with a 1K resistor back to the blue wire on the decoder. If you don't have decoder, then run them from the +20 volt internal bus in the loco. The switched end of the LEDs can go back to the appropriate pad on the headlight board (the one connected to the diode, a cylindrical part). Since the stock headlight wire is blue too, don't get it confused with a decoder blue wire. A decoder will supply about 20 volts on its blue accessory wire.

The actual headlight assemblies at both ends are held in with great globs of hot glue and are fairly difficult to remove. The actual headlight bulbs are encased in more hot glue and are virtually impossible to remove without remelting the glue into a sticky mess with a hot air gun or literally destroying the bulbs if the glue globs are just pried out. This is not really a problem because if you are pulling the bulbs, you are going to replace them anyway.

The front headlight circuit board can be freed by popping off the hot glue that holds it in place and then by pulling the number board assembly straight off. The wires to the bulbs will break, but no matter. Then the board can be removed. If you are really careful, and lucky, you might be able to work the front PWB out the front without breaking the wires, but you'll be cutting them anyway.

The rear board can be worked out by removing the glue globs. Be aware that the glue may adhere to some of the chip parts on the rear board and if you are not careful, you may rip some of them off the board. Where glue covers some parts, it may be better to heat it with some hot air so that it will soften. It'll make a bigger mess, but it won't rip up the boards. The rear lighting board is retained by small ridges in the sidewalls of the body. The board should be worked straight forward so that the marker LEDs will clear the holes in the body before the board is moved upward or downward. It is much easier to work with this board if the speaker assembly is removed first. Note that the marker LEDs are actually installed in little sockets, one of the LEDs in the photo is laying loose on the board. They can be plugged in either way but will only work in one direction. The smaller electrode inside the plastic lens goes toward the nearest edge of the board.

The boards can be reinstalled without any hot glue and they will stay in place well enough. If you really feel the need to secure them, use a VERY SMALL dab of hot glue. The monster globs installed at the factory simply aren't necessary.

A 5 mm white LED will fit back in the headlight holes. Just enough sticks out to simulate a headlight lens. Then a small dab of hot glue can be dropped back in to secure the LEDs. The blue wire in the photo is the new wire that goes back to the DCC decoder blue wire or pin 6 of the DCC socket. I trimmed off an unused corner of the board to allow the wire to get past the board without the wire being pinched. The green wire solders back to the board to use the existing wiring to connect the headlight back to the DCC socket.

When I first tested this circuit, the headlights wouldn't go off, they got dimmer, but nowhere near off. It turns out that the red LEDs in the markers have a fairly low reverse breakdown voltage and they were clamping the control voltage to a maximum of 12 volts. I had to install another diode in series with the markers on both ends to get the headlights to go off when they were supposed to. Cut the red wire going to the front markers (middle wire on the connector) and the orange wire going to the rear lighting board and insert a diode with the arrow pointing back to the main board. These diodes will become reversed bias when the lights are supposed to be off and block the red LEDs from drawing any current and holding the headlights on. There are similar diodes already on the headlight boards to block reverse currents in the headlights.

The lights behind the number boards also illuminate the cab. However, they are pretty bright and actually show through the number board housing. While you have it off, you should paint the inside with a couple of coats of Engine Black paint.

Both the Soundtraxx Sierra sound system and the NCE D408SR DCC decoder that I have installed support both Type 1 and Type 2 ditch lights. I tested both and I find that the effect from the Sierra is better because the ditch lights stay flashing longer after the horn trigger is removed. The Sierra provides about 3 seconds of delay following the completion of the horn sequence, the D408SR provides about 5 seconds from the release of the F2 key which is the beginning of the horn sequence. Since I will be running the Sierra in the triggered mode, this results in longer ditch light operation. They will also operate if the horn autotriggers. For type 2 ditch lights (lights off if they are not flashing), the actual lighting effect is about the same. The type 1 ditch light (lights on when not flashing) doesn't work quite right on the D408SR. If the ditch lights are programmed to be off in reverse, only the flashing effect goes off, the actual lights stay on. This is probably a bug in the decoder firmware.

After looking at every photo of an SD45 I could find, I saw no example of any lighting that could be a ditch light EXCEPT the two lights on the low hood that Aristo uses as markers. I have elected to remove them and use them as ditch lights instead by replacing the red LEDs with white LEDs and wiring them back to the Sierra. This takes 3 additional wires between the body (where all the other electronics are) and the frame (where the low hood is). The Sierra's lighting output is 6 volts so a 180 ohm resistor is needed in series with each LED to set the current to about 20 mA.

The only original lighting left in the low hood is the porch light. Since the mounting points for the LEDs on the board were immersed in hot glue and I didn't feel like cleaning up the board, I elected to abandon the board and wire the lights direct. Note that these LEDs are in sockets as well.

I initially used the front hood lights as ditch lights, but I came down with a case of RCD (Rivet Counter's Disease) and made some more legitimate ditch lights. I scoured the internet and found several photos of SD45-2's with more conventional ditch lights (Golden State Rail Fan Site) so I figured that if ATSF backfitted SD45-2's with this kind of light, they might have done some SD45's as well. These ditch lights were made from a piece of 0.080" styrene sheet with an LED mounted in a 0.200" hole. A piece of 0.25" OD tubing makes the cowl. The wires were covered with black shrink tubing and the whole works was painted engine black. The LED lens was then cleaned off with a Q-tip soaked in lacquer thinner and mounted on the frame of the SD45.

There is a page on the internet that describes all manner locomotive lighting. Ditch and crossing lights are described there as well.

The ditch light mounts are simply glued to the deck and aimed "crosseyed" and down so that the lights illuminate the track in front of the loco well. Real ditch lights are crossed at an angle of about 1 degree so the lights will cross several hundred feed down the track. This maximizes the visual impact of the lighting so that the loco can be most easily seen from a great distance. Since even our while LED headlights don't make that bright a beam, these lights are crossed at a much larger angle to light up a swath in from of the loco. They are also aimed down slightly. I did my aiming by eye for the best effect, I haven't tried to measure the resulting angle, but it's probably about 5 degrees.

[ Top ]

The SD45 is equipped with a new kind of a smoke unit. This unit generates a fair volume of smoke and shuts itself off when it runs out of smoke fluid so that it won't ever burn up. This feature does seem to work but it behaves differently depending on what kind of power is applied.

With Aristo PWC on the track, the smoke unit will run until it runs nearly dry, then it will cycle off and on about a 2 second interval. With DC on the track, when it runs down, but is not completely dry, it will simply shut down and stay shut down until the power is interrupted for a few seconds and restored.

The smoke unit itself is built into a small plastic box which contains entire unit including its fan. On the end of the box, there is a two pin connector. The air intake is on the bottom.

The unit contains a fair amount of electronics to control the smoke unit and regulate its smoke output. The smoke element itself is a typical Aristo wire wound heater element wrapped in a fiberglass wick. The wick extends into a fluid sump that can hold a large quantity of fluid. Fluid is drawn by the wick up to the element. When the unit finally does shut down, the sump and the wick are virtually dry.

The small compartment is the fluid sump, the large hole is the air intake for the fan. This sump can be filled half full and it won't slow the unit down.

Even though the performance of the unit is not impacted much by the quantity of smoke fluid, the unit seems to like some kinds of fluid much more than others. The unit ripped through 60 drops of LGB fluid in just 8 minutes and didn't even make a very strong plume. 10 drops of Bachmann fluid lasted 3 minutes and was much denser. See the table below for the relative performance of various fluids in this unit.

The unit makes full smoke at about 8 volts on the track. The smoke element itself runs 5.8 volts (on my unit anyway) and this stays virtually constant at higher track voltages.

The power dissipation of the unit increases rapidly with increasing input voltage until the unit begins to regulate at about 8 volts. Then as the voltage increases, the input current decreases so that the unit dissipation does not increase nearly as fast with increasing power input. This characteristic is unlike linearly regulated constant intensity smoke systems which keep the smoke intensity constant, but draw significantly increasing power at higher track voltages.

LGB smoke fluid works exceptionally well in LGB smoke units, but it doesn't do too well in the Aristo units, the SD45 unit being no exception. The smoke density and duration are a strong function of the type of fluid used. The table below indicates the duration and relative density of smoke obtained from various fluids. Aristo's own smoke fluid is not in the list because none of the retailers near my home carry it so I don't have any.

It is interesting to note that the fluids that make the most smoke density also last the longest. I believe that the LGB and the Lionel fluid are similar, both being formulated to operate in Seuthe smoke generators.

[ Top ]

I've measured the SD45 as best I can with a caliper and measuring tape and it comes out pretty close to 1/29 scale for the body dimensions that I evaluated. The "Corrected" column is the actual loco dimension referenced to the rail head as the height dimensions were actually measured from the shelf that the track was sitting on.

My reference for scale is the drawing shown in the Model Railroader Cyclopedia - Volume 2. There is a 10% error in the wheel gauge, but this is by design, 1/29 scale doesn't fit on 45 mm track (1/32 scale) without some error.

The SD45 does have some significant overhang, but it isn't any worse than the other very long rolling stock available. On an ~4' radius (8' diameter AKA R3 curve) the front and rear steps overhang to the outside by 1 7/8" measured from the inside of the outside rail. The center of the loco hangs inward by 2" from the inside of the inside rail.

While the loco won't reasonably run on 2' radius curves (R1), the bricks will and without binding. I have heard from one user that has a few 2.5' radius (R2) curves and the loco makes it around those curves, but not gracefully. The coupler swing is too severe and the loco will derail the following cars.

[ Top ]

The SD45 has been, unlike many other locos, designed to be disassembled by the user for installation of DCC, sound or R/C. There are 8 screws on the bottom, clearly marked with arrows, that need to be removed to remove the long hood. Pull these screws and the handrails that attach to the cab and lift up the long hood. The top part of the cab will come off with the long hood. The frame and the hood are connected by two keyed plugs that cannot be interchanged or plugged in backwards. At this point, virtually all of the wiring is accessible to allow accessory installation.

The short hood comes off with 4 additional screws located near the corners of the assembly.

Don't remove any screws along the center line of the loco, these hold on the weights and do not have be removed for interior access.

The entire power bricks can be removed by pulling 4 screws located between each wheelset. Just the bottom covers of the trucks can be removed by removing 8 small screws around the perimeter of the cover.

The speaker assembly can be removed once the long hood is off by removing 4 screws.

The lighting boards are glued in with excess quantities of hot glue. this glue should be softened with a hot air gun before removal is attempted.

[ Top ]

The Aristo model is powered by FOUR can motors, two in each truck. All six axles are powered through individual gearboxes. All six axles ride in ball bearings and each gearbox has two more ball bearings for a total of 24 ball bearing assemblies in the loco.

The entire motor and gear train assembly can be easily pulled in one piece. If you remove the 8 screws that hold on the brick lower cover and turn the engine over, the whole assembly will literally fall out provided that the motor contacts have not been soldered to their contacts as has been done on some units.

Before the assembly is taken out, it is recommended that each motor be marked for its position. I numbered the bottoms starting at 1 from the front and drew an arrow pointing front. It is important to get the motors back in correctly. If a motor is rotated a half turn, it will run backwards. If you do lose track of the motor rotation, there is a red dot on the end of each motor. This dot goes nearest the bottom of the brick.

Its reasonable to mark on the gearboxes as well so that they can reinstalled in the same positions, but this is not entirely necessary. They can go in in either direction and they will still run right in principle. Due to tolerances, some gearboxes might run better in one position vs. another.

All the truck power pickup contacts are via sliding contacts. This photo shows the power contact for the center gearbox. It is the metal tab going up the inside of the brick housing. The motor power tabs slide into the fork contacts. A soft plastic spacer is visible at the top of the brick. There is another one on the other side of the brick. On some units, there were TWO pads incorrectly stacked which materially messes up the wheel loading of the brick.

The purpose of the compliant plastic pad is to limit the rotation of the center gearbox. If the pads were not there, all the axles could rock and the engine body would flop from side to side.

On some units, Aristo has soldered the motor tabs to the fork contacts. I my opinion, this should not be done because is materially complicates disassembly of the brick. There was apparently some concern that the heating of the fork contact was contributing to brick heating and lubrication failure. I suspect that soldering the tabs is overkill and not really necessary. If your tabs are soldered and you need to get the brick apart, the only practical way to do it is to remove the brick from the loco and then cut the motor contact straps (the ones closer to the centerline of the brick) and bend them so that the half straps can be pulled out with the brick assembly. Of course, the straps are then history and must be replaced upon reassembly. If you have to do this, get some replacement straps BEFORE you reassemble the brick.

The motor used in the SD45 appears to be a standard issue Aristo motor. It has a DC resistance of 6.7 ohms which puts it right in the class of the motors used in other Aristo locos (except the Pacific which uses a larger one). The shaft is treated with a plastic ball-hex fitting that mates in the hex socket of the worm in the gearbox. The red dot which marks the motor tab polarity can be seen in this photo. When the motors are reinstalled in the brick, all the tabs should be nearest the bottom of the brick.

While the motors were out, I measured the free run current along with the stall resistance. These two motors were clearly in the same class as these other two motors from a typical Aristo diesel. The stall resistance of the older motors is a little lower (5 ohms or so) than the SD45 motors. SD45 motor #2 draws a little more current at high voltage because it is not perfectly balanced and it buzzes a little. This increases the load and therefore the current.

The power contact on the gearbox is screwed to the side of the gearbox and contacts the ball bearing race. The half round bump slides along the tab inside the brick housing. The housing is held together with 4 screws, but the wheels have to be removed to access the screws.

The wheels are secured by a screw and lockwasher in the end of the axle. However, since the wheel sits on a tapered shaft, it will usually be stuck in place even if the screw is removed. A gear puller can be used to remove the wheels but it is not really required. The wheels aren't on that tight. A straight blade screwdriver can be placed next to the axle and rotated against the back of the wheel. It doesn't take much pressure to pop the wheel off. If more than light pressure is required to release the wheel, the a gear puller should be used. Others have reported that when they put too much pressure on the wheel to release it, the half axle actually pulled out of the worm gear. A gear puller presses down on the end of the axle and pulls the gear without placing any force on the axle attachment.

The gearbox has plastic worm that engages a plastic worm gear. The worm rides in two ball bearings. The half axles are screwed into the sides of the worm gear and each rides in a ball bearing. This gearbox was adequately lubricated. Since it is a relatively time consuming task to get all six gearboxes apart, this would seem to be a good place for Aristo to install a lube plug on some future version of this gearbox.

On the first production run of the SD45, the factory apparently used a less than ideal grease that breaks down at too low of a temperature and runs out of the gearbox. This results in a dry worm and the worm will eventually grind itself up. You can check to see if you have adequate grease by running the loco with the lower truck covers removed. If the grease in the boxes is adequate, you will see globs of it moving around through the translucent gearbox material. If you suspect that your boxes need grease, its a straightforward but time consuming task to remove each one in turn, disassemble it and regrease it. Use a good plastic compatible grease such as LGB gear grease.

There is a single ball bearing contact in each gearbox half. This contact rides on the axle and provides an electrical path around the ball bearing race. The race is electrically connected to the pickup contact so some current will flow in the bearing races. If enough electrical current flows in a ball bearing, it can pit the balls and races and cause the bearing to wear much too quickly.

The ball is sitting on top of a spring that is in turn retained by the metal clip on the outside of the gearbox half. Nothing retains the ball except the grease. If the ball falls out, scoop up a little grease on a small screwdriver and pick up the ball with the grease and then slop the whole works back in the hole. Apply more grease to the ball if it doesn't stay in place. When the axle is inserted, the ball is depressed, the spring is compressed and the ball is then captured in the hole that the spring sits in.

I did notice that the bricks seem to be "high centered" on the center axle by maybe 20 mils. Even with the weight of the loco on the central axle, the pads didn't seem to compress enough to cause all the wheels to ride on the track with even loading. I took out the pads and I found that it didn't get any better. On this brick, the central gearbox didn't ride at the same height as the other two axles, pads or not. It is riding on the circular ridges on the gearbox that make up the motor mount which are hard against the stops in the brick housing. I laid a piece of rail across the wheels and found that it would rock on central axle.

I couldn't tell by measurement that the support ridge was higher than the others, but it was clearly too high. Maybe the brick housing is slightly warped or there is some other difficulty. I don't have a microflat table and the dial indicators that it would take to characterize it the problem exactly. Instead, I just ground about 10 mils of plastic off the ridges that support the central gearbox. Then the gearboxes all rode at the same height and the test rail didn't rock. Without the pads in place, the central axle might have been set in a few mils more than the other axles causing the end axles to take up all the load.

I put the pads back in and then the center gearbox was actually resting on the pads. It was then a little high with no load (the engine is on its back) but when the track is pressed down, the pads compress and the gearbox settles down into the pads and the track doesn't rock. The end axles then provide the hard support with the center axle being compliant and conforming to the other two.

The other brick had exactly the same difficulty and required exactly the same modification. Now all the wheels ride on the track and take up load. The center axle is still loaded somewhat heavier than the other two, but it is much better than it was. Without a truly equalized suspension, getting the axle load fully balanced on a 3 axle rigid truck will be very difficult. After the modification though, it is close enough.

This issue may or may not afflict other locos. Before you do any grinding, you should carefully evaluate your trucks. I initially detected the problem by setting the engine on a flat piece of track and then used a small screwdriver the lever up each wheel in turn. Before the mod, it was obvious that one wheelset wasn't even resting on the track. After the mod, it takes some force to lift each wheel and when the force is released, the wheel drops back to the track with a very audible click.

Aristo has addressed this problem when they rework bricks by moving the pads to an end axle. This also tends to line up the axles without any grinding on the brick housing.

The SD45 power brick is very easy to remove entirely. Just remove 4 screws (one between each wheel) and the brick will literally drop out. There are no wires to mess with. The contact straps on the brick contact pads on a board on the loco truck bolster.

The long outer strips are the power pickups. The shorter inner strips are the motor contacts. Each strip has two spring fingers formed in it to provide a contact to the circuit board on the truck frame.

The contact board on the truck bolster has large patches on it to contact the strips along the top of the brick. This board is then wired through the bolster and the frame to two small boards mounted on the frame topside. These boards act as distribution points for track power and motor power. There are transient suppressors installed on the other side of this board that seem to be effective in suppressing radio noise.

After the bricks were broken in for awhile, the slow speed performance of the loco degraded instead of improving. The loco wouldn't run as slowly and it tended to jerk to a start. The problem was in the front truck, it didn't want to run smoothly anymore. After much fiddling around, I determined that there was a concentricity problem between the some of the motor shafts and the worms. This effect is indicated by a wobble when a motor and gearbox are run together outside of the truck. Why this began to be a problem after break in, I don't have a clue. In any event, Aristo had apparently discovered the problem as well and asked for the bricks back as they had a fix already in hand. When the bricks were returned, the motor shafts were equipped with brass hex fittings instead of the former plastic hex ball fittings. The old fittings fit quite tightly into the worms, the brass ones slip fit. I don't know when or if the brass fittings were incorporated into production but Aristo indicated that all new production (at least after Oct 2001) would be built this way.

These brass hex fittings are not user installable, it takes a gear press to get them on.

The slow speed performance of BOTH trucks was better than before. It used to take 4 volts to get the motors to run and the slowest no load wheel speed was about 6 rpm for the "good" truck and 10-12 rpm for the "bad" truck. When the bricks were returned, both of them would run at about 1 volt at about 4 rpm or less with a DC power source. When running from an MRC Tech II power pack with back EMF control, the trucks would run steadily at about 0.7 rpm (90 seconds/turn).

The reworked bricks also had the pads moved to one of the end gearboxes of the brick. Because I had modified my housings, this move slightly upset the wheel loading that I had adjusted before so I moved them back to the center gearbox.

[ Top ]

The SD45 has the most complicated internal wiring of any loco that I've come across to date. This is the first large scale loco that I have come across that comes prewired to accept DCC or R/C control without having to tear into any existing wiring. There is one complication for DCC installation though that needs to be fixed in some future version of the main circuit board. I got around it with some cuts and jumps.

The unit is designed to accept power either from the connectors available at both ends or from the track. The connectors can be used to bus multiple locos together to share power pickups or a battery trail car, however a female to female adaptor connector (not supplied) will be necessary to connect two locos with the supplied connectors. The power switch on the loco either connects or disconnects the power pickups to this bus. If you leave the power switch on while the loco is connected to batteries in a trail car, the loco will backfeed power back to the track which is undesirable.

I've converted the power connectors on this loco to my personal standard DB-25 type Power Connector Pins.

All the accessories in the loco (except the smoke) run from an internal bus derived from a PWM power converter. This converter reaches full output out with about 4 volts in so that the accessories are nearly constant intensity. The headlights run from this internal bus are returned directly to a rail through a diode. This is a fairly common method of directional headlight control. If a DCC decoder is used, then the lights will return through a DCC function output instead. The blue wire output of a DCC decoder would not be used. The Lamp On/Off switch disconnects all loads from this converter. The headlight circuits draw about 40 mA average each.

The smoke unit runs at full rail voltage of either polarity or on DCC track power and is connected through a power switch to the rail inputs through a jumper on the jumper plug. One lead goes directly to a rail so that the smoke unit should not controlled by a DCC decoder function output unless it is rewired.

There is a bridge rectifier on the board that supplies DC power to the PWM power converter. Both outputs of this bridge are available on both the DCC and the accessory connectors. The DC voltage on these pins will roughly equal the peak rail voltage.

There is a problem with this approach during integration with DCC. The PWM power converter is on all the time so that the current that it draws will confuse a DCC command station during decoder programming. The command station assumes that ALL loads are disconnected or wired through the decoder during programming. During programming, the command station communicates with the decoder by the decoder drawing a pulse of current for a affirmative answer to a query posed by the command station. If there are parasitic loads, the command station cannot tell when the decoder is answering. This will be a major problem when programming on the programming track. When doing programming on the main in OPS mode, all decoder functions except the 2 digit address can be programmed but without read back of the actual CV values. The lighting switch should be wired to disconnect the INPUT to the PWM power converter instead of disconnecting the loads from its output as designed. This requires some cuts and jumps to move the lighting switch to its proper electrical location.

The motors are connected through a motor switch and the jumper plug which comes installed in the "DCC" socket. A DCC decoder or R/C RX could be substituted for this jumper plug so that no cutting or hacking on the wiring is necessary to get the basic functionality of command control. No one currently (Nov 2001) supplies anything that plugs right in. The jumper plug is symmetrical, it can be plugged in either way. The plug jumpers pins 1 through 4, pins 5 and 8, and pins 9 through 12. Pin 1 of the DCC socket is furthest from the J1 designation.

A limited number of accessories are available for control at the DCC connector. There are pins for power, the motors, the front headlights, the rear headlights and the smoke system. There is no cross connect between the DCC connector and the sound connector so sound function control will have to be wired around the main circuit board.

There is a second connector on main board which is configured for connection to an accessory system. It has a combination of sound and power functions. The tables below lists the pinouts and functions of each pin for both connectors. Again, pin 1 is furthest from the J2 designation.

The following three diagrams are the complete schematic of the stock SD45. These roughly follow Aristo's drawings dated March 9, 2000, but some errors on those drawings have been corrected. The drawings are given in three parts, the first one is the wiring in the shell, the second is the wiring on the frame, and the third is the wiring of the sound and accessory connector.

[ Top ]

Except for the complication involving decoder programming, DCC installation was straightforward. I elected to install an NCE D408SR decoder because it is the only one that can handle the average current of the SD45 at full lug. The DC resistance of the motor used in the SD45 is 6.7 ohms minimum. This implies that the stall current at 16 volts is 2.4 amps per motor. A DCC decoder will see all 4 motors in parallel with a little series resistance in the wiring from the DCC socket. The total stall current should be a little under 10 amps which jives with the 8.8 amps that was actually measured.

The decoder itself was just stuck to the inside wall of the shell with double back tape. One of the blue wires (accessory power) was run to the rear to power the rear LED headlights and the other goes forward for the same purpose. The two sound trigger wires go back to a set of optoisolators (see the section on sound installation below) to trigger the horn and bell. All the other connections (track power, motor and headlight controls) are wired through the DCC connector.

The connections to the DCC socket are listed in the table below. A 12 pin strip of single row 0.1" pin grid connector is an exact substitute for the Aristo jumper plug.

The problem with the lighting circuit can be fixed with some cuts and jumps. The fix is to disconnect the traces from the lighting switch and permanently jump the connection that the switch used to make and break. Then the same switch is rewired to break the connection between the DCC connector and the rectifier diodes on the board. I elected to break BOTH connections so I needed to cut the short traces between the pads on the lighting switch too. Since I had already removed the switch to be sure that I knew what I was doing, I was able to do this. If only one side of the bridge connection is broken, then these two short traces don't need to be cut and only two wires need to be added. Getting the switch off is a difficult task and consumes quite a bit of solder wick.

On the top side of the board, the traces between the DCC connector and the diodes need to be cut too. Then wires from the switch are wired back to the cuts to switch one or both on or off. This fixes the decoder programming function and leaves the lighting switch functioning essentially the same.

This schematic shows the changes that need to be made assuming that only one side of the bridge rectifier is switched out. This method does not require that the swtich be removed from the board to cut the traces underneath. Three cuts and jumps are necessary.

[ Top ]

I have elected to install a Soundtraxx Sierra sound system in the SD45. There is an internal connector intended for installation of an Aristo Digital sound system, but I have been pleased with the Sierra and I will hack it in instead.

The SD45 comes with a speaker preinstalled and prewired to the sound connector but the acoustic mount of the speaker is less than adequate. A fairly high power 8 ohm speaker is mounted directly under the radiator fans projecting upward. This is good. However, the front of the speaker and the back of the speaker are both exposed to the inside of the shell. This is bad. The internal acoustic coupling will allow some of the waves from the front of the speaker to interfere with the waves from the back. This has the bad effect of reducing the volume emitted by the speaker, especially at the low end of the audio spectrum. The low frequency sounds provides the "guts" of the diesel engine sound and are important.

The speaker is mounted on 4 posts on the radiator fan assembly. The assembly itself is easily removed from the loco by removing four screws. The sides of the speaker mount are partially blocked by plastic ridges, but these don't extend high enough to reach the speaker seal. The ends are entirely open.

There are lots of ways to fix this problem, I choose this one. A set of walls made from 0.125" x 0.250" styrene were assembled to enclose all three radiator fans. A 0.030" strip was glued to the top of the existing sidewall to seal against speaker. All these parts were attached with Zap-a-Gap CA. All the joints between the pieces were filled with a little more Zap-a-Gap CA. This method isolates the front of the speaker from the back and allows the entire long hood to act as an enclosure for the speaker. The sound is directed out all 3 radiator fans.

Then two covers were made from 0.080" styrene sheet and glued over the top of the walls. The covers fit tightly against the edges of the speaker. The result of this modification was quite successful. The overall volume of the sound system was at least doubled and the low end was significantly enhanced.

For those that aren't using the Aristo sound system, the speaker terminals are available as two solder pads on the edge of the main circuit board. These pads are the two closest to the speaker connector. Check it by using a ohmmeter set to the X1 range. When the pads are probed, the speaker should click.

The sound system connector is also a standard 0.1" spacing pin grid header. These can be purchased in lengths of 40 to 80 pins and simply cut to the desired length. This same pin grid header is used for the DCC connector. This particular version does not have captive pins as it is designed for the short ends to be soldered into a circuit board to hold them in position. The plastic header just provides support until the strip is soldered. When wires are attached, the pins may move a little. A small drop of superglue placed at the joint between the pin and the header after soldering will hold it in place.

The Sierra sound system installation to go along with DCC is a little complicated by the peculiar characteristics of the Sierra. It takes quite a bit of interface to get the Sierra to work properly with a DCC decoder or the Aristo CRE-55490 on board R/C receivers. Other R/C systems may be afflicted by the same difficulty. Regular track powered installations won't need this interface. The schematic diagram of the interface circuit and an operational description can be found on my Sierra Tips page.

Due to the addition of the interface circuit, the installation looks complicated but it isn't really all that bad. The interface board is on the lower right, the Sierra is at the lower left. Both boards are fastened to the shell wall with double back foam tape. The Sierra battery is mounted with double back tape to the speaker enclosure.

Note that the frame weight protrudes upward from the shell edge by about 3/4" so that all the boards and parts have be mounted high enough to clear it.

The sound power and volume/external programming switches are mounted under the dynamic brake housing with the other switches. There isn't a lot of vertical clearance so the switch handles had be trimmed short with a pair of diagonal cutters so that they would clear the cover.

There were times that the Sierra sounded distorted with an irritating buzzing sound. I had assumed that the sound was coming from the fans but it was not. It turned out that the inner grills were buzzing. It wouldn't happen all the time and sometimes one or more of the six grills were causing the problem. Anyway, after poking around with a toothpick at full whistle I found the sensitive spots. A single drop of Zap-A-Gap CA applied with a toothpick at the bottom center of each grill would prevent the grills from buzzing. The glue is shiney in this photo because it hasn't dried yet. This was the last grill that I glued after determining that the patch worked on the more serious offenders.