Initial studies:

The process, stumbling blocks, example runs and initial results are shown in the  movie.

Subsequent work will look into other factors that will influence these characteristics. 

observations

a thread asked what engines exceeded expectations.   So what are expectations

the tractive effort for a prototypical locomotive is 25% of its weight.   I've found this to be roughly the same for models (at least 20%).

then the question is how many cars can it pull.  what is the expected force required to pull a car.   That depends on how well the trucks roll.   I went thru my trucks when i saw some cars rolling on level ground.   I was able to make improvements but for some it was a challenge and I accepted that a truck would roll on it's own on a 2% grade.

we measured a pull force of 2.4oz for 30 high quality cars estimated to weigh ~4 oz.   this comes out to ~2% (2.4 / 120 oz).   this is equivalent to the additional force needed to pull a train up a 2% grade.    this doesn't consider curves which have an equivalent grade of 32/R.   

from above, a loco weighing 12oz can produce a drawbar for of at least 2.4 oz (20%) and could just pull this train.

it could pull half this train (15 cars) up a 2% grade.

if you're having problems pulling a train up a grade there are two things to consider: 1) increasing the weight of the locomotive, 2) improving how well trucks roll with a truck tuner and lubrication

if you're having problems

Or, do as real railroads do, add another engine!

When I built my first "large" layout in may garage in Indiana, it had a 2.5% grade.  I tried pulling a 25 car train with 2 Atlas GP7's and found that was about the limit of what they could pull up that grade.  I'd day the cars were all fairly free rolling cars (a mix of Walthers, MDC, Athearn etc).  If I stopped the train on the grade and then applied power, the wheels would slip on the loco's but would start moving and stop slipping when moving.

On my last layout, I had a 2.9% grade and two Athearn RTR SD45's (six axle) could manage 25 cars up the grade but were near their pulling limit as a pair.

measure

are they?   part of my point is to verify the performance of both the locomotives and cars to see if there's room for improvement.   

if the average rolling resistance of the cars is 2%, locos weighing 10oz are required to pull 25 cars on level ground

locos weighing 24 oz is required to pull the same train up a 2.9% grade.

Yes, the kit built cars rolled fairly well as stated.

Yes, they rolled fairly well as I stated, which I could tell as part of the process of building my kits..

Of course if you feel the need to go through a measurement/verification process thats up to the individual.  I only have so much hobby time and there are so many hobbies within the hobby; I don't feel the need to actually apply metrics to rolling resistance.  How many people actually do that?

A number of people have

A number of people have complained about the video transitions among other things, I will take note of that in the future.  

It seems that several are wanting to use a simple derivative approach to this problem. Motor performance is not linear and thus the linear approach will not always yield the result your looking for.  

At this point, I am not necessarily looking for the best approach to yield the longest train length and or the highest grade. At this juncture, I am looking for what do the performance characteristics for a random set of engines look like.

The following chart is shown in the video with additional animation and explanation:

The y-axis is the total car weight divided by the quantity of the engine weight on the drive wheels times the ratio of the drive wheels divided by 8.
This normalizes the results. Because the track segments are the same and the cars are the same, the the differences are from the engines. Primarily from the motors.

There are a number of things that come mind. The motor operating condition to the design relationship is a primary factor, this is when the operating point has put the motor torque output near its maximum capability. Others include the strength of the magnet flux, the condition of the electromagnetic poles and the internal motor bearing drags.

If these are the case, adding weight will not add more pulling capability.  

Clearly the data indicates that one rule of thumb is not adequate.

The number maximum number of cars on the video charts is the largest number that can be pulled up the grade three times. For this case, all the drive wheels were slipping to some some degree. When the wheels start to slip the current draw shape changes.  

The purpose of this set of experiments was to define an initial set of data base characteristics. With this insight comes many questions. Where possible & of value, I will work to answer some of these.

.

you don't need to measure every truck.   you can simply put a car on a 2% incline and see if it rolls freely.   those cars that don't could probably be improved.   i have cars that seem to roll uphill.   others that were worse than 5% before tuning.

on another thread, someone said their 4.7 oz 0-6-0 starts slipping when it pulled 8 cars weighing 2.5 oz each.   i estimated that the resistance of those 7 cars was the equivalent of a 4.7% grade.  based on that,  the person said he felt the loco "has pretty good pulling power".    but i feel with tuned trucks that loco could pull twice as many cars.

Just check rollability

Per my Rolling Stock book, you just want to set performance standards for your cars and then test each car to make sure it meets those standards. Here's my "rolling stock run like a dream standards" checklist for my Siskiyou Line 2.

Note item #15 -- car freely rolls down 24" test track when one end elevated with 1/2" piece of wood (~2% grade).

(Note, this is after the wheelsets have been changed out with InterMountain all-metal wheelsets. Metal axles against plastic sideframes have better rollability. Kadee's plastic wheelset axles in plastic sideframes exhibit a poorer rollability.)

If the car doesn't roll freely down this grade (if it needs *any coaxing* then it first gets the conical axle holes burnished with a 4B graphite pencil to see if that solves it. If not, then each of the four or six conical axle holes get a light ream* with the truck reamer tool. Most of the time, that fixes it. If it's still a little sticky, I give the conical holes another burnish with the 4B graphite pencil.

I've found significant issues on the SL1 with cars having poor rollability -- in fact that has been documented in a TrainMasters TV video. With SL2, I will not allow that to happen!

*Do not ream the truck sideframe axle tip holes excessively. If the wheelsets are just too tight in the sideframe, replace the trucks with something better. Out-of-alignment side-to-side slop can develop if you ream the sideframes excessively, making for extremely derailment-prone trucks that dog-leg down the track at a slight angle.

Joe, has a good point

Joe, has a good point regarding rolling of cars. I will provide an actual test on a grade a scale 3/4 of a mile long in HO scale. The grade is 2% to about 2.5% The train in question is longer than the grade and traveled around some curves at the base and the crest. Locomotives used were 2 Atlas gold series H24-66 locomotives and one Atlas H16-44 classic with dcc and no sound. Locomotives were cleaned and lubricated and reassembled. All cars were rebuilt and tuned to my standards which are nearly identical to those Joe mentioned in his book. The exception is my cars will roll on a very slight grade and I use both code 88 and 110 wheel sets. My cars are all heavier than the nmra standard by about 1 to 2 oz. With the three engines listed 160 cars were pulled up the grade and around the layout for hours on end. Using the same three engines a later run with 187 cars was made. The other 27 cars were just what was available on the layout and some were poor rolling cars. They were just handy. At no time did the locomotives slip on the grade or exhibit any other issues with running on the layout. In some cases the train was traveling around 540 degrees of curves and with grades both up and down with out incident.

I suspect the data you are collecting in your test is flawed by the use of unrealistic test cars in your test scenario.

An open question

Have been asked what my G&TL studies mean to a persons layout.  That is an open question.  It depends on a number of factors.

The track resistance was intentionally set to be minimal.  

1- Even at that I was surprised at the pulling capability at grade.  Nine of the twelve pulled 2X the drive wheels at 4% grade.  If you assume a worst case of 1.5% grade of track resistance in a helix or complex nolex, then 2.5% grade is doable in a layout design.

2- The ability to pull a few cars at 11% grade was a surprise.  This has ramifications for certain types of trains and layouts.  

3- the change in scale velocity with load was less than expected prior to wheel slip.  

After wheel slip there is a limited number of additional cars that can be pulled up the grade.  

The above can be drives from the data taken using the current techniques.  

Additional insight can be achieved by examining a number of one, two or more off variations.  This would be done on selected engines.  These might include:

1- increasing the engine weight, by 3 steps for a total of 1 pound.  This may be academic because it may be impossible to add that much weight.  It will get at the motor capacity and the grade impact of the weight change.

2- motor variation, Kato, Athearn Genesis, Mashima, etc

3- general tune up

4- electronics, DCC, Rail Pro, etc

This list is by no means complete.  I would be interested in hearing other ideas from you all.

Rob, was the grade more than the length of the train?

The approximately 100 foot train would climb 24 to 30 inches depending on the actual grade.  That would mean each of you engines would pull neatly 100 cars on a no grade layout.  Are those stock engines?  All you did was clean & lubricate?  That adds credence to that option.

document

can you post you data in a document that can be downloaded and reviewed?   

I can post the three other results charts

Like the one posted earlier.  Bacause of the video those are in the public domain.  Anything else I prefer to keep to my self for now.

Larry, 2 of the engines each

Larry, 2 of the engines each will pull approx 120 cars on a level section of the layout but curves are involved, I do not recall the exact number right now. That is the H24-66 Atlas Trainmasters. I believe I was able to pull approx 80 up the 2% grade with each of the engines. I did not record the data at the time of the tests, it was only to see if my efforts in cleaning and lubing new engines were worth the efforts. I concluded they were. Out of the box my Trainmasters did not perform as well as the test data from model railroader. My conclusion was my efforts were worth the time it took to do them. The only thing I did was disassemble and clean them, paying particular attention to reassembly to make sure there was no binding or out of alignment and lubricate them. My freight cars are extremely free rolling even though they are 1 to 2 oz heavier than the NMRA recommended practice. The H16-44 was cleaned and lubed as well and I suspect on level track it would pull about 60 cars, again I checked this some time ago.

The general changes from published test data from out of the box units from Model Railroader to my actual testing shows a great increase between their test results and my actual tests with my rolling stock and my tuned locomotives.

I attribute this to both the very free rolling nature of my cars and my elimination of any binding in the mechanisms of the locomotives and a general reduction of friction in the drive train.

The things I do are things that Joe covers in his excellent text on rolling stock. He says he wishes he had that when he was starting out in the hobby and I agree. It seems that I was doing 90% of the things he listed in his book after discovering them the hard way, it is a great reference on how to do it right. It also lists several alternatives to the weight issue and discussion that seems to polarize many modelers. I have found that on curves below 36 inch radius it is very helpful to have heavy cars if one is going to pull 100 car or more trains.

If train length will be less than 50 cars I have found the nmra standards work fine with curves that range from 36 inches to approx 26 inches in HO scale. It seems the builders of my clubs layout had standards everywhere except where they did not, which baffles me.

I hope this response was helpful.

Grade & train length investigation published on 21Feb2020

In my 11 years of reading MRH, this is my first blog entry.  This investigation is outstanding, to say the least.  I have been asking questions of MRH and other model railroading resources for information like this for years.  As my layout is built [I am just restarting in N Scale], I will hope to contribute to this excellent investigation with additional information [concerning N Scale], including possible amperage draws of the locomotives, my ultimate goal.  KEEP UP THE GREAT WORK!

Here are the three charts of results from the movie

The next two are engine only results:

All of these are at a power supply voltage of 12 volts DC.  The Athearn SD45 does not show in the Engine only data because It was early in the series & I did not take a complete set of EO data.  That engine has subsequently been sold so the missing data can not be defined.

Thanks for your kind words and encouragement.  I have a plan to get more answers, will post the highlights here when they are available.

Larry

i'd also appreciate the text in the video

if you did this in powerpoint, could you post it.   if not, could you just create a blog entry with the information in the videos, the various bullet lists viewgraphs: goal of experiments, potential solutions, needed cars, intial test plans,  summaries, conclusions, tables, ...

Yes need explanations of what the charts are telling us

Power = VI

Because the results are all done at 12 volts, the the current change represents the power change.  This chart is for the engine only, so the power change is only for the grade impact.  I quote as supply power because the current measure is on the supply.  The only load is the engine so it is most likely the engine power change as well.
the current draw for these engines vary.  It made sense to normalize each curve by the zero grade value for each engine.  This aligns them at zero grade and equal change in the resulting curve is an equal percent change in power.  
until the wheels slip, I anticipated that these curves would follow a very similar slope.  They do not.  
When the wheels start to slip, the power drops, the motor is initially drawing less current.  This corresponds to a reduction in the velocity.  As the grade increases further, the current increases again, even with the wheels slipping.  
As indicated in the video, there is no commonality to the way these engines load up to handle the grade.  This tells me that operating to design points on the motors are very different.  I am saying that as a  person who spent a career working with motor performance.

I personally find this detail intriguing, but most people will not.  For this reason I limited the video to the highlights.  Making the obvious conclusions.  I have some theory’s about what is happening, but at this time that is all they are.  
 

As indicated in the video, reason for doing the work is many fold.  Primarily it is to understand what was happening when I was exercising the RS-18 with grade.  The problem with these activities is the “relative to what” question that naturally follows.  For this reason I included 12 engines in the examination.  It is a start of a data base for comparisons.  I am not sure where it will lead.  I have a number of ideas to follow and fortunately it is on my nickel.

boggling nomeclature

as an engineer, i wasn't shy to ask if that was a technical term when i didn't understand someone's explanation and to explain it.   I worked with some really smart guys who could always explain things in mister rogers terms

what do you mean by equal percent change in power  (presumably this data is at full voltage, so by power do you mean current)?

why would you expect them to "follow a similar slope"?   what do the slope of the curves imply? 

what do you mean by "start to slip",  doesn't the locomotive speed decrease when slipping?  and isn't current expected to decrease as slip occurs and motor rpm (and bemf) increase?

isn't current expected to increase as grade increases?    not sure what your trying to say, whether your describing expected or unexpected behavior

where in the video? what do you mean by "load up"?   what do you mean by "operating design points"?

perhaps a short tutorial on motor characteristics would help understand the measurements and charts in the video 

Loco efficiency

I'd have thought that as long as you can make the loco slip its wheels, anything you do to make it more efficient (motor tune-up, gear lube etc) is irrelevant. If enough torque gets through to the wheels to haul the train and spin the wheels if the train is too heavy, you can't make use of any extra power gained by a better mechanism.

Maybe the ideal would be to have a locomotive so heavy that when it starts to spin its wheels, the mechanism is just about to stall. And by the time that happened, the loco would be dragged down to some very low speed even at full power, which seems like a prototypical situation! Then you'd need to think about how rapidly the motor is heating up, and whether you can climb to the top of the grade before something gets damaged.

Sand. Can we use sand? (Better not, I think. And it might stop the loco from getting electric power.)

Tractive effort

Why is this necessary?  Tractive effort is based on weight on drivers, period.  Number of drive axles does not change that.

Scott Chatfield

It is an attempt

To put them on an equal basis. This tells you which system is the most effective puller. Adding elements, like weight, traction tires or something else to that engine should yield more cars.
 

The 4-6-6-4 is an examples an engine that looks good in the tests  It pulls near the most actual cars, but it is the heaviest engine I have tested of the 400 plus that I have run tests on.  Plus it has six drive axles and traction tires.  when normalized this way, it does not look as good as others.  It may benefit from Robs clean, tune & adjust technique or its mechanism is just less effective.  

equal basis to what?

why doesn't total car weight / engine weight put them on an equal basis?   or total weight / number of drive wheels?    can you explain why that approach puts them on an equal basis?

it would be useful to have the raw data posted