Experimental Gradient and Curve Predictor

[NIK]

Hi,

 

I've been working on a way of predicting how different trains will behave on different gradients and curves.

At present it is a spreadsheet that uses the values for gradient, loco pull, curve radius, loco and train weight and static friction of train being hauled.

 

I've just done some more tests using a club layout that has a 1:40 gradient combined with a 36" radius curve that starts part way up the gradient.

 

A Bachmann OO 2H DEMU Power car slipped to a halt with 9 mk1 vehicles in tow. It slipped with 8 vehicles and went up the gradient/curve with 7 vehicles.

 

I did a crude test of the hauled vehicles static friction with an adjustable incline plank. I measured the pull of the 2H power car when I got home and added the data to a spreadsheet that models the gradient combined with curve (using the experimental gradient and curve calculator results). The gradient track starts straight then curves, comes to a summit and continue to curve round. The model divides the track up into 1 foot sections - so a section can be straight, straight with gradient, gradient with curve and finally level with curve.

 

The model predicted the result to the nearest coach and the point of slipping to the correct section at which slipping occurred.

 

The gradient/curve calculator predicted the power car would have 58.5 grammes pull at 1 in 40 gradient. With two Mk1 vehicle removed the peak drag dipped under 58.5. The predicted drag was just over at 8 vehicles and 10 grammes over at 9 vehicles.

 

I had only brought along 5 Mk1 EMU trailing coaches as the pull figure I had written down was 32 grammes - the lowest pull of a loco/power car that had a DCC decoder. It turned out to be the best 4 wheel drive pull figure (without Powerbase magnets). I had to borrow Mk1 BGs and a GUV to add to the train - and I could only do crude static friction tests and the BGs were a lower weight (haven't weighed one of my GUVs yet).

 

I plan to do another test before the layout is put back into storage after our open day in Basingstoke on Saturday. First I will try and get a good record of the pull when I retest my weediest DCC fitted loco/power car and I will try and get better static friction records for each of the coaches I use for the test. Its probably better to have a weedy loco and a few coaches so the coach static friction has less influence.

 

I'd like to find out why the 2H power car is so grippy. It seems to have the same weight and weight distribution as the Bachmann 4CEP and 2EPB but 50%+ more pull. I will also repeat its pull test using DCC to see if that makes any difference.

 

Regards

 

Nick

[nswgr1855]

Hello Nick,

 

I have done the experiments in H0 a couple of years ago so my  layout grades would not be to steep for the trains I wanted to run. My results will be relevant to 00. I tackled the problem from a slightly different perspective. Firstly I made the assumption all wagon  weights were proportional to their length. I use the Australian Model Railway Association Mass standard http://amra.asn.au/standards/

I measured what I call the coefficient of rolling resistance of my wagons by using the inclined track method, that is inclining a piece of track with a wagon on it and measuring the incline when the wagon starts to roll. There is a large difference measured between different wagons. I only use pin point axles and bearings which are usually free rolling. The value I picked was the largest typical value of 0.02.

Like you I measured the coefficient of friction of my locomotives using an inclined plain. I wanted the worst case so for calculation purposes I used the lowest value, 0.2 which was on track with a small amount of oil, which I use for improving electrical contact. From these values you can calculate easily the maximum train mass that can be pulled on flat straight track by a given locomotive mass which works out at train mass =10 times locomotive mass. Experiments confirmed this. Measuring  curve friction is more difficult and to confirm by experimentation. Over may years I have tried to do the experiments on different layouts with Helices. The problem is the radius, grades, locomotive and wagon variables are hard to control. Any way I came up with a rough rule of thumb for layout design to compensate for the curves on grades, and that is multiply the prototype grade by 1.75,(1:40 becomes 1:70)  If you want to run prototype length trains up to 2700mm long on typical H0 model railways using curves down to 900mm radius. Conversely going the other way, a 1:40 grade on your layout is like having a 1:23 prototype grade.

DCC will make no difference to pulling power other than if it can control the motor better at low speeds. What will make a difference is wheel tire material. Generally harder materials used for model railway wheels have a lower coefficient of friction but are preferably used due to lower electrical contact resistance and less dirt pickup.  Also what makes a big difference is if all the loco weight is on the driving wheels. If you have non powered bogies with inside plain bearings, that is the formula for a poor pulling model.  These models tend to have traction tires in order to pull enough.

 

Cheers,

Terry Flynn.

Edited April 13, 2019 by nswgr1855
fix grammer

[NIK]

Hi Terry,

 

Many thanks for the info.

 

On the subject of oil I had a Bachmann 2H power car that measured 31 grammes pull after running in on my rolling road. I then tested on my club layouts gradient/curve and it pulled far more 4CEP trailer coaches than expected. When I got it back home it measured almost twice as much pull as before. I then put it on my rolling road for 1 minute and its pull then went back down to 41 grammes.

 

I then retested it on the club layout and it behaved as a loco with 41 grammes pull should behave according to my predicter.

 

Back  home I retested a Bachmann 4-CEP power car and it too was pulling a lot more than previously. So prompted by your mention of oil I put a bit of Peco power-Lube on the pull test track and the pulls of the two power cars dropped back to 31 grammes.

 

Cleaning the track with a cloth and with a track rubber had little effect - the pull was reduced and the oil presumably on the wheels of the power bogie.

 

I then got some talcum powder which is very fine and is thought to soak up oil and put a small amount on the track.

 

The pull increased immediately and kept climbing, the CEP power car reaching 143 grammes!. That's the same as a Heljan Class 128 unit which has twice as many driving wheels and 50% more weight.

 

So I'm guessing that oil is getting from my rolling road rollers onto the wheels of the locos under test but over time it is dispersing or being soaked up by dust on the rails of the club layout (I can't recall it ever having been hoovered).

 

I'm not suggesting putting talc on the tracks of exhibition layouts at this stage but I may repeat your technique of putting a small amount of oil on my pull test track.

 

Regards

 

Nick

[NIK]

Hi,

 

Due to not having access to the gradient and curve on the club layout I help with (due to COVID restrictions) I'm thinking about extending my gradient and curve predictor to include predicting the speed the train may traverse the up gradient and any curve.

 

I'm going to try measuring the speed of a train on a flat test track and adding vehicles that have had their friction measured and see how the speed drops off as vehicles are added.

 

Might then be possible translate from vehicle friction into gradient/curve drag and predict the speed for any feasible gradient and curve.

 

One problem is I don't have room for a test track with infinite radius curves so maybe I will have to make a long test track and measure the speed between two positions along the track. I've used JMRI in the past so I may use it to make the train do say ten runs in each direction, recording the speed each time, then I manually add a coach, JMRI repeats the runs and repeat until the test is complete.

 

Another problem is that the speed reduction may vary according to the power applied to the motor. Something to investigate.

 

Regards

 

Nick

[DCB]

You are getting into the detail of wheel and rail profiles and cleanliness of the rails here.  Generally I find heritage profile Romford wheels and Code 100 are a couple of wagons better than RTR wheels on code 100 all other factors remaining the same. Wheels grip on the inside corners of the rails not the tops, the gauge corners of East Coast infamy where 100 year life rails failed after 10 years or so

Long experience has shown me that haulage degrades as the track gets cleaned by passing trains.  Last week a 28XX on test hauled 27 wagons for over half an hour before slipping to a halt on a gradient. I reduced the load and ended up with first 24 and then seeing she was slipping I went back to 22 which is incidentally the capacity of my reception roads.   The 28XX is unusual in that she has a long flat can motor out of a computer with absolutely bags of torque and 40:1 gearing.  The wheels revolve at pretty much a constant speed depending on the voltage ( V not VA ) whether or not she is moving.  It is also quite normal for locos which have no issue with a 7 coach train up the 1 in 36 ish grade to the terminus on starting an operating session struggle with the same one an hour later.

Curves are a difficult one, locos grip better on curves while four wheels are taking the weight, but when they start rocking due to the twist in the track traction diminishes rapidly, Stock drags on curves so the loco will often slip to a halt on the straight following a curve. I found the place where my locos were slipping to a halt was actually downhill on one occasion while fiddling with riser spacers trying to improve haulage.  It was a hump a couple of feet before that causing the grief.

sprung locos are a lot better than unsprung locos weight for weight.

Coach weight and drag also varies.  Many recent RTR coaches are ridiculously heavy and draggy, We had to bin most of the weights in the Hornby Hawksworths to get a decent train up hill. 

Finally speed comes into it.  A fast ascent allows the momentum to carry the train over humps and dips where a slow ascent would cause it to slip to a stand.  We arrange a banker for trains stopping at the base of the climb where if given a run they would have no difficulty. Then again two locos neither of which will pull 4 coaches each will double head on 8 easily.

It is a science but there a so many variables it starts to look more like a black art.

Edited June 26, 2020 by DavidCBroad