[?]

Tractive effort is the force that a locomotive can apply to its coupler to pull a train. In the US, it is usually expressed in pounds. Tractive effort is not often directly measured. Most locomotive builders calculate tractive effort based on a percentage of the weight placed on the driving wheels. For rod type steam engines, this factor is usually 25% for clean dry rail. It was found through tests that this factor worked quite well because most steam engines have all their drivers locked together so that slippage of an individual wheel is not possible. Articulated engines could have half of the engine slip, as would often happen, but the 25% factor was still applied. Application of sand could increase this factor considerably.

Diesel engines can slip axle by axle. Over the years, steadily improving motor control systems have been developed to control wheel slip such that the latest batch of large diesels can apply 1000 hp per axle to the rail without serious slip. First generation diesels had trouble applying 400 hp per axle without slippage. The old 25% constant doesn't hold well with diesels, tractive effort as measured by dynamometer cars have shown that the factor can often exceed 35% with the newest AC powered high power diesels.

In most cases, tractive effort limits the pulling power of locomotives only at very slow speeds. Above a few miles per hour, prime mover horsepower limits usually prevail.

What does this have to do with large scale models? Our locomotives can slip as well. Most large scale locomotives have enough motor power to slip their drivers under load at any speed so that wheel slippage, not motor power, is usually the limiting factor in determining how long a train our engines can pull.

My layout has a ruling grade of 1.6% for about 60 feet. There are two quarter circle 4' radius curves in the middle of the grade as well. There is also a parallel track with 5' turns in it, but the track with the tighter turns is the ruling grade as it limits train length more severely. The performance of my engines on this grade was of interest to me so I ran some marginally scientific tests. It occurred to me that others may be interested in my data as well. Your mileage may differ....

I ran every engine that I had available. Maybe some interested SoCal large scale railroaders could bring by some other kinds to be run as well. I'd be happy to spend an afternoon messing with trains.

• Test Plan
• Test Methods
• Test Results
• A Second Round of Tests 26 May 02
• A Third Round of Tests
• Testing a Battery Power Conversion
• Conclusion

Since the object of the tests was to determine how many cars an engine could pull, it seemed reasonable to quantify the results in number of cars that could be pulled under two conditions, slipping and not slipping. I also desired to record the power consumed by the engines, so I metered the total average current delivered to the rails.

The plan was to load each engine up in four conditions:

• Loaded to a stall at full track voltage with the wheels slipping
• Loaded to almost a stall with the wheels slipping with enough power to keep the train moving on the grade
• Loaded just below wheel slippage
• Loaded to an actual stall

Under each condition, the number of cars pulled and the current was recorded. In the stalled and slipping case, the number of cars was irrelevant so only the current was recorded.

A test train was created from a set of fairly heavy cars. Most of the cars were Aristo boxcars, reefers or stock cars. When I ran out of those, I added some hoppers and tank cars. The lighter cars were placed near the end of the train so that when the train was cut to lighten the load, the lighter cars were cut. All the cars were equipped with San-Val metal wheels and all have their wheel bearings lubricated with LGB 50019 oil. Before the test, each car was roll tested down the grade to make sure that none of them had their "brakes" set.

A 10 amp digital meter was inserted in series with the track feeder. The power system is an Aristo 5460 power supply feeding an Aristo 5471 TE receiver set to PWC mode.

The engine was positioned just upgrade from the second turn, which is historically the worst spot on the grade, with all the cars in the train actually on the grade. Four tests were done as follows.

Stall Test. The engine was prevented from moving by physically restraining it as the power system was set to full power (18 volts on the track) and the current was measured to the nearest 0.1 amp.

Slipping Test. The train was selectively cut until the engine could maintain speed on the grade with wheels slipping. If the engine couldn't actually start the train on the grade, the train was given a push. The power was set to the minimum setting that would allow the engine to maintain speed. When the wheels are slipping it is usually not possible to accelerate so the engine was allowed to run at whatever speed it could maintain. The number of cars and the current was recorded.

Starting Test. The train was cut further to determine the number of cars that could be started and accelerated on the grade without wheel slippage. The number of cars was recorded. The current was recorded after the train managed to get moving and the current had stabilized. The actual starting current was usually a little higher than the running current.

Actual Stall Current Test. In the actual stall case, the track voltage was set at 16 VDC and the engine was pressed down onto the track until it actually stalled. This was done several times. If the stall current was stable enough, the reading was recorded to the nearest 0.1 amp. This case is important for proper sizing of a DCC decoder or radio control receiver.

The test was repeated for each engine in the GIRR's fleet. Each test was run with lights running but no smoke. Where sound systems were installed, they were left running as they could not be shut off.

In the case of steam engines, the tender is not counted as a car.

The track is Aristo Code 332 brass rail that had been cleaned with a few passes with an Aristo track cleaning car. The track was dry and no oil had been used for track cleaning for several months. Dirty track tends to provide better traction. Oil materially degrades traction. All of the initial tests were run within a two hour period so I assume that the track condition was the unchanged during the span of the tests. The Rogers and following engines were run subsequently as those engines arrived, so the track might have been in a somewhat different condition for those tests.

The results of the tests are shown below. Aristo engines are coded by their box color to indicate roughly the year of manufacture. Gray box locos are the earliest, black box production started roughly in 1994, and yellow box units are produced in 1997 and later.

1. Could not actually be stalled under as much pressure as I dared to apply
2. Digitrax DG580L already installed, running in analog mode
3. Test power supply current limited at 10 amps
4. Motor only
5. Without traction tires

I just finished running a second round of testing, this time with a mostly different set of engines and under different conditions. This time I tested the roster of the GIRR, Mountain Division. Some of the engines that I had tested above have migrated to the Mountain Division so they show up in both places. Note that the Shay is a different unit that was purchased at the same time.

This time, the environment is different. I used all Bachmann box cars, reefers and stock cars all equipped with San-Val metal wheels. The grade is different too and much more severe. The ruling grade of the Mountain Division is a 5.2% 4' diameter spiral and this killer grade taxes the poor engines heavily forcing me to double or sometimes triple the grade.

I also recorded the current running light on level track. Light current is pretty variable so I recorded an eyeball average. The actual currents varied up to 0.2 amp above and below the average for most of the engines. The power source used was an MRC 9300 power pack with an Aristo ART-5471 Train Engineer trackside receiver running from the 20 volt DC output of the 9300. All testing was done with PWC on. The stall current testing was done with an Aristo 5460 power supply driving a 5471 trackside receiver set to 16 volts average on the test track.

Engine	Stall Current (amps)		Pulling, Slipping		Pulling, Not Slipping		Current Running Light	Notes
	Slipping	Hard Stall	Cars	Amps	Cars	Amps	Amps
Aristo Rogers	2.3	6	5	1.9	5	1.9	0.7	No added weight 2 traction tires
LGB 2017 and matching powered tender	1.4	5.51,3	5	1.9	4	1.3	0.7	approx 12 oz each added to engine and tender
Bachmann 2-4-2	2.1	4.5	2	2.1	1	1.8	1.1	approx 8 oz added to engine, power pickup on 2 tender axles
Lionel James	2	6.8	2	1.7	1	1.5	0.8	12 oz lead added, power pickups added to center drivers
Bachmann Big Hauler ATSF #49, first run unit	0.8	1.52	3	0.75	2	0.7	0.4	Approx 1.5 lbs lead added, power pickup on 2 tender axles
Bachmann Big Hauler Emmett Kelly Circus 2nd run unit	1.5	4.5	3	1.4	2	1.3	0.7	Approx 1.5 lbs lead added
Bachmann Shay	2.3	71	8	2.0	7	1.9	0.7	box stock
Lehmann Porter	1.5	4.51	2	1.2	2	1.2	0.9	Approx 1 lb lead added, 5v smoke unit running
Aristo C-16	1.4	3	4	1.35	3	1.3	0.7	no added weight
Aristo Center Cab 26May02	n/a	n/a	7	n/a	6	n/a	n/a	converted to battery power, weight of 15 NiMH AA cells added
Aristo Railbus 26May02	5	1.09	n/a4	n/a	1	0.83	0.7	sound added but not running

1. Could not actually be stalled under as much pressure as I dared to apply
2. Engine would not stall and gears skipping teeth
3. Tested as a set
4. Wouldn't pull 2 cars even slipping

I just finished running a third round of testing, this time on engines that had been tested in Round 1 but converted to DCC. I wanted to see if DCC had improved or degraded the pulling power of these locos

The environment is pretty much the same as the first round of tests, but since the data was taken something like year later, the track is probably in a different state. It is dry and lightly cleaned.

I recorded the AC supply current being drawn by the DCC booster. Since the metering is different and the control circuits are different, these currents cannot be directly compared to the currents in Round 1. They can be compared between themselves though. All the current readings include the overhead needed to run the DCC booster and command station (0.25 amp) and the power to the loco's sound system which was running in each case.

I did a conversion of the black box FA to battery power with an onboard Aristo 5490 receiver and 18 volts worth of gel cell batteries (3 Ah rated). I wanted to see what the conversion did to the pulling power of the locomotive. I expected the tractive effort to increase because of the added weight and it did pull better. The batteries weight about 4 pounds. The small steel weights were removed in the processes, but the engine is significantly heavier.

This locomotive was wired for "tri-modal" operation as described in my FA Tips page. It can run as a straight track powered loco, as a constant track powered loco and as a battery powered loco. This allowed me to compare the results of the three modes.

Note that when running from track power, the 5490 is active even if it isn't being used and the headlights (LED conversions) are on so that there is a 200 mA static load on the track.

The Aristo Center Cab and LGB 2060 were converted later. The Center Cab is also a tri-modal job, the 2060 is battery only.

Loco	Power Mode	Stall Current (amps)	Pulling, Slipping		Pulling, Not Slipping		Current Running Light	Notes
		Slipping	Cars	Amps	Cars	Amps	Amps
Aristo FA-1 black box	Straight Track Power	2.8	20	2.5	16	2.4	1.1	used to pull 11 cars when stock
Aristo FA-1 black box	Constant Track Power	2.6	20	2	16	2.1	0.9	PWC on the track set to max speed
Aristo FA-1 black box	Internal Battery Power	n/a	20	n/a	16	n/a	n/a	no track current due to onboard batteries
Aristo Center Cab 1Jun02	Internal Battery Power	n/a	8	n/a	8	n/a	n/a	15 NiMH AA cells and CRE-55491 75 MHz RX
LGB 2060 1Jun02	Internal Battery Power	n/a	3	n/a	3	n/a	n/a	15 NiMH AA cells and CRE-55490 27 MHz RX

The initial winner was really no surprise. The best puller in all tests was the Bachmann Shay until it was matched by the Aristo SD45. It has consistently outpulled any of the other engines in real world service. What REALLY surprised me was the pulling power after DCC had been installed. It's hard to believe that DCC did this, it may just be a traction improvement due to track conditions. However, the engine would pull this load from the first speed step when it was barely crawling along. The back-emf feature of this decoder is simply amazing.

The older RS-3 did surprise me however. When it was new, it could handle only 6 or 7 cars on a similar test. After the weight was added, it improved to be similar to the FA's. However, a year later after the wheels have become all scuffed up, it pulls stumps. The newer one isn't quite as heavy and its wheels are still shiny and fully plated. The difference is pronounced.

After DCC was installed in the RS-3's, they got even better, but again, this might be just a relative traction issue. The two RS-3's together didn't pull as well as the sum of their individual loads and the two together didn't do as well as the Shay did by itself.

The FA's and FB's are mostly similar. They weigh less than either RS-3 and even though their wheels are pretty scuffy, they don't pull nearly as well as the older RS-3.

After battery conversion, one FA pulled much better, probably due to the added weight of the batteries. The motor currents, where they could be measured, are essentially unchanged.

The Pacific either grabs or it slips badly. When it slips at all, it hops all around and chatters like mad. The 0-4-0 does the same, but doesn't pull nearly as well. All of the other engines slip fairly gracefully and pulled a little harder when slipping. The Pacific pulled about the same slipping or not. I believe that this is because once it starts to slip, traction levels off abruptly as the wheels hop on the rails. I didn't test the Pacific under DCC as I had removed its decoder to put in the Bachmann Climax.

Even though the Lionel Atlantic is heavily weighted, it still doesn't pull really well. Before it was weighted, it wouldn't pull its own shadow. It would growl and hop around a little, but not nearly as much as either of the Aristo steamers.

It is no surprise that the Bachmann 4-6-0 Big Hauler doesn't pull well, they never do. The one in the first run of tests is a 3rd generation unit, but it is similar to the 2nd generation unit. The ATSF unit is a first run unit. Even though it is considerably noisier, it draws the lowest current of any of them by a significant margin.

The little LGB industrial diesel does surprisingly well considering its weight. It does have a traction tire.

The Lehmann Porter is not a very good puller and it also seems to draw a lot of current. 250 mA or so of that is going to the smoke unit which cannot be turned off.

The LGB 2017 (a Lake George and Boulder 0-4-0) and its matching powered tender pull fairly well, about the same as the 2 motor Rogers.

I was surprised at the range of stall current in the various locos, especially the Aristo diesels which are of essentially the same design. The Aristo 0-4-0 drew the most current, and in only one motor at that. It would seem prudent to run this engine carefully so that it doesn't slip or it might burn up quickly.

Thomas the Tank Engine and his buddy James are clearly among the weakest of the lot. Even with weight added, it'll handle only two cars. With the engine running forward wheel slip is fairly graceful. While pushing, it hops and chatters like mad. I've never figured out why it does this.

The Aristo Rogers 2-4-2 did quite well considering it only has 4 drivers. This is probably due to the fact that it has two motors and two traction tires. The Rogers doesn't hop and chatter like the other Aristo steamers when stalled.

The new USA Trains GP-7/9 pulls very well also, nearly as well as the Bachmann Shay and the broken in RS-3. It has 4 traction tires and it is not nearly as heavy as either of the other two star pullers. With another pound or two of lead, it might be a new leader. Without traction tires and with NWSL stainless steel wheels, the tractive effort degraded by a few cars, but the engine still pulls well. It did respond well to extra weight. The engine will do one additional car per half pound of added weight at least up to two additional pounds.

The Aristo C-16 pulled surprisingly well for such a small engine. I had heard reports that it could only handle 10 cars or so on the flat, but this one worked fine with 12 cars on a 1.6% grade. Without wheel slip, it pulled better than some of the two motor diesels.

I expected the Climax to pull better than it actually did. It isn't as heavy as the Shay, but it pulled only half as many cars. When I ran this test I rechecked the Shay as a reference and it still handled its 23 cars. The Climax did a little better with DCC, again this may be a traction issue. It did pull smoothly at the slowest speed step due to the back-emf control of the DCC decoder.

The Bachmann Columbia 2-4-2 was true to form and pulled very poorly. I was surprised at the current that it drew, even running light.

The new Aristo Center Cab Industrial switcher pulled better than I expected it to considering its fairly light weight. It is right up in the range with the other Aristo diesels without additional weight.

The new Aristo Classic Rail Bus is a weak puller, but this is to be expected. There is only one powered truck and not much weight. However, it did as well as most other engines with only 4 driven wheels.

While I didn't recheck this observation during the Round 1 tests, I have observed in the past that a pair of nearly matched engines, such as a pair of FA's, will pull more than the sum of cars that each engine would pull alone by maybe a car to two. During the DCC test, the two RS-3's wouldn't pull the sum of their individual loads by several cars. I don't have a good explanation for this result.

The hard stall current tests were surprising in that so much current could be drawn by some of the locos. The hard stall current is mostly determined by the DC resistance of the motors and the parasitic resistance of the loco wiring. See my Small Motor Tips page for an explanation of what determines motor current. The smaller one motor locos often drew as much current as the two motor locos in a hard stall. The GP-9 really sucks the juice, it could draw more than 10 amps all by itself.

Stall current is important for sizing radio receivers or DCC decoders so that they won't burn up if the engine actually stalls. For most of our locos, this is unlikely as they will usually slip instead and the lower of the two currents will apply. However, I have been able to stall the Lionel Atlantic and the decoder complains quickly by shutting down.

The locos that already have a DCC decoder installed draw less stall current than non DCC equipped locos. This is probably due to the internal voltage drops in the decoder and perhaps some current limiting effects. For example, all the FA and FB's draw 6 to 7 amps. The RS-3's which have essentially identical motors draw about 4 amps.

When a loco is just starting out and the motor is hardly turning, the peak currents supplied by a decoder can nearly reach these hard stall levels during the short duration PWM pulses. For a large scale engine, I don't see any reason NOT to use the largest decoder available.