It has been debated many times about the impact of car weight and track grade on train performance.
Thinking about this problem further, I realized that the result I’m looking for is the train length on a 2.5% grade. There are a number of examples where the conventional method did not allow that determination. The usual technique is to measure the drawbar force on level track and then convert to the train length based on some correlations found online. This technique is weak in many ways. The biggest issue is the ability to measure the draw bar force accurately and consistently. Clearly actually running a train at the required grade takes most of the uncertainty away.
Looking at the recent projections, the likely range of cars is between 15 & 25 4 ounce cars. This means a weight on wheels of 60 to 100 ounces. I say it this way, because my test track will not allow 15 to 25 cars. The train lengths would be roughly 15 to 25 feet. This does not work in my 7 foot test section.
An alternate option is to add weight to cars to allow the weight on wheels in fewer cars. One or two is likely to few, even if you could get the weight in the cars. Three to five cars is the number that could work on the test section and yield a reasonable result. Three cars requires the resulting car and added weights to be more than 2 pounds.
Experimenting with the washers I usually add to cars, a resulting weight of 16 oz (1 pound) can be easily added to an Athear BB 50′ boxcar, as shown below:
Using this type weight, another 4 oz can also easily be added. Thus each car would equal 5 traditional cars. This would require five of these weighted cars to equal 100 ounces on wheels. Still a problem.
An obvious way to add additional weight is to fill the holes in the washer stacks shown in the above figure. An option is to use a metal dowel. A dowel that is 1 inches long and 1/4 in diameter. With these dowels (actually flat head bolts) plus appropriate added washers allows the car weight to achieve 24 oz(6 cars). At this weight, four cars will equal 96 ounces(24 cars). with seven of these cars 42 cars can be simulated. This is likely adequate for most engines. This will be explored as part of this activity. Some recent attempted tests have shown the need to have as much capacity per car as can be accommodated. For this reason, the use of wheel weights are being explored. The hope is to get to a 7,8 or 9 car representation per actual car. The wheel weights are 1/4 ounce steel. Getting lead is very difficult today. The best that I could do with these wheel weights was 28 ounces in a 50’ car. Thus simulating seven cars in one is my limit.
To be able to pin down the train length to one car, the additional partially weighted cars will be needed. Weighted cars representing 1 through 6 4 ounce cars will need to be created to cover most possibles. So a total of 12 weighted cars are needed to do this testing. This way in a 5 car string you can represent 35 cars.
All will be 50’ boxcars except the the 4 ounce car. The base 50’ car exceeds that weight, so it will be a 40’ boxcar. The representative car numbers will be taped to the roof walk on the car.
This process has show to be quite successful. Following this technique the Rapido RS-18 has been examined over grade levels from zero to 17.2 percent. Using equivalent train lengths from engine only to sixteen cars.
These test have shown some interesting results.
1.) The RS-18 can climb a 17.2 percent grade with little impact on the engine speed voltage function, as shown below:
To run larger grades requires some additional stabilizing to the test track, for basically an academically interesting result. As time goes on this may be revisited, but for now higher grade testing will not be attempted.
2.) Even though, the engine can climb a 7% grade, when four cars are added, the train slides down the hill. Additional tests show that when the train length is near the maximum for the grade, when the power is turned off, the train will slide down the grade.
A practical interest for this type testing is to determine some useful train grade limits.
3.) the first of these is how many cars can be successfully pulled up a 2.5% grade. In this case a series of tests have developed a map of the solution space for this engine. These results are shown in the charts below:
The RS-18 data is shown in green and red. The green is the engine only configuration on a flat test track. The solid red is the same configuration on a 2.5% grade. Here we see the minimal impact that the engine shows moving from a flat to a 2.5% grade.
The red data moving down in speed represent the train lengths of 4,8,12 and 15 cars. This shows the dramatic speed Chang from 8 to 15 cars. This engine could not pull 16 cars on the 2.5% grade.
Thr red lines are from a train on a track with a 2.5% grade. The green symbols are the engine only results at zero grade. As you progress up in velocity, the subsequent curves are developed at constant power supply voltages of 4, 6, 9, 12, & 16 volts.
Clearly, this engine can pull 15 cars up this grade. The other results discussed with the charts are equally interesting.
4.) The next important finding is what is the maximum grade that a realistic train length. Using the real world rule of thumb of 2X cars to engine powered axles, a realistic train length is 8 cars. The question is what is the maximum grade where the engine can pull that train length. To pursue this answer, an additional series of tests were performed at a grade of 4%. These results are shown in the following charts:
Taking this further the grade was varied until the nine could not pull 8 cars up the grade. This was shown to be 6.2% grade. This brings into question about the 7% grade observation earlier. At 6.1% the train moved. At 6.2 it did not. Based on these results the following charts show the velocity and current functions with grade.
This activity has generated a number of questions.
1.) What is the grade limit that you should impose on your layout?
2.) How is this impacted by modern technology engines?
3.) These tests have been performed with a pure DC signal, how does a PWM DC signal change this result?
4.) Does a DCC derived PWM signal change this? What about Rail Pro?
These questions tee up other useful tests to bring more insight. Those will be included in other posts.
To help answer these and other questions that are stirred by these tests, additional engines will be tested in this manner.
The next engine getting this examination is an Athearn Genesis Challenger 4-6-6-4 steam model. (For more details see the post on the 4-6-6-4 activities.
The more samples, the clearer the answers will be. Twelve additional engines have been identified to supplement this information A result that is desired is the function of maximum cars to grade and/or the maximum fractional engine weight to grade. The latter is more appropriate of the engine capability. These charts will be developed here as the data comes available.
The first activity using the weighted car replications achieved the intended result. In addition a situation has arose that may have been missed before. Namely the impact of wheel slip on these results. Previously I thought wheel slip would only effect a high grade high load point. During these tests it became clear that the impact of this slip, namely the heat developed on the wheels & track seems to hang on.
What was observed is that a given condition experienced wheel slip near the end of the run. Moving the train back to the start a second run experienced wheel slip over the entire test length with a dramatically reduction in velocity and current. A third run repeated in close proximity stalled before reaching the end of the sets section.
Reducing the number of cars back to the car number that previously yielded three successful and repeatable results this time showed excessive wheel slip with largely reduced and inconsistent results. This has implications on both the the process & the relative results. The first two engines did not show much wheel slip and demonstrated very repeatable results. The more recent two engines both experienced very pronounced wheel slip, with questionable results. Wheel slip will occur, but time between runs is required when it occurs. The testing will not be quick. Waited a day and reran the above tests. the previously no slip condition was back and repeated the original tests. This verifies my suspicions about the impact of excessive wheel slip.
Completed a third engine on this series. Everything seems great. However, it is prudent to run some checks on the weighted cars. Appropriate to run some actual 3,4,5 &6 car trains against those weighted cars. Also should check the seven cars against each other & against two car combinations equaling seven cars.
spent the day doing several more tests. Specifically examining the validity of the car simulations. Looked at three actual cars comparing to the 3 car simulator. As well as 4,5,& 6 car simulators. These were tested behind an engine and the velocity and current were measured at 12 volts. In every case the data was within the expected variation. This indicates that the car simulators are representative of the cars they are simulated. They may be slightly optimistic, but not grossly. Likely one car in 10, if that.
Four additional engines were completed through this process.
I have decided to break this into a discussion of the car simulation cars, a summary of the engines tested & a few charts summarizing the results & conclusions.
1/20/20 spent the day working with engines seven & eight in the series of these tests. The two are both fairly recent releases, 2015 I believe. The results are showing that the focus on sound has led to a power reduction that shows up in train length reduction.
This is being posted in an incomplete mode. Charts and additional comments are forth coming.
