In search of smooth running: my lessons in loco construction

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

In search of smooth running: my lessons in loco construction

Postby Julian Roberts » Sat Jan 04, 2014 11:29 am

I was looking on the Forum for ideas for an LMS Compound. Not finding much, I went ahead anyway. Now that the model is a working proposition, I thought I'd post up what I'd done in case it helps anyone else. But as some of my ideas don't make sense without explaining them more fully, first I'm posting this. Hopefully it's not too banal for ALL the readers of this Forum, though no doubt it will be for many, for which I apologize in advance; and I hope it may be of interest to some others.

There are around 8 numbered topics - here is the first.

1.) Lessons in compensation.


My approach to making locomotives is that first and foremost they must work well, and as a close second they must look as good as I can make them. But - if there has to be a compromise between one and the other, running well takes priority.

My adult modelling life started in 1995, but only really got going when, after a considerable amount of frustration, light dawned as I read Iain Rice’s three loco construction books in the Wild Swan series.

My first compensated chassis was in 00 gauge - a Perseverance SR Terrier chassis under the Nucast kit. Although it basically worked well, I found that it tended to derail on sharper curves at speed. At around the same time I was operating Mike Walshaw's superb 00 gauge finescale "Westport" where derailments were almost unheard of, and noticed he had the same problem in the hidden fiddle yard with a compensated 4-6-2 Pacific 'spam can' (the bogie wheels were not part of the compensation). He said he suspected not enough weight was on the front wheels, and this was causing the flange on the outside curve to climb the rail. I surmised this might also be the problem with my Terrier and concluded that I would in future aim with compensated suspension to have the majority of the weight on the four corner, or leading, wheels.

So, a straightforward example is an 0-4-0 - the weight is equally on all four wheels whatever the beam arrangement. Any additional driving wheels would not bear an equal share of the load, but more likely only half as much as the corner or leading wheels. With beam suspension it is perfectly clear how to arrange such weight distribution, by adjusting the fulcrum point from half way between axles to a 1/3 - 2/3 proportion. The same would apply if I were to make an 0-6-0 twin beam arrangement instead of fixed axle.

I did not in fact make any more locos for 00 gauge - I moved to P4 where it seemed to me that this issue was far more pertinent given the very small flanges. Consistently I have arranged any compensation to be along these lines: so, in an 0-6-0, assuming a rear fixed axle, I place the fulcrum of the beam 2/3rds of the way towards the front axle from the centre one. Thus the front wheel has twice as much weight on it as the centre one. Where that weight is in the body is not particularly critical assuming the centre of gravity is roughly in the middle.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Sat Jan 04, 2014 2:15 pm, edited 1 time in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 11:31 am

In the 0-8-0T LNER Q1 that I made in about 2009 the suspension is as for a compensated 0-4-0 with a rocking front axle on centre beam. The springing of the top wiper pick ups is all that holds the two centre pairs of wheels down. Maybe the adhesion of the loco suffers compared with what it could achieve but that is of academic interest to me in that it does everything asked of it on the West of Scotland 4mm Group’s layout “Calderside”.

The middle wheels of a six, eight, or ten coupled loco can perfectly well, rather than compensated, be sprung with the Gibson hornblocks. However these don't have a strong enough internal spring for the corner wheels to my way of thinking, to ensure these wheels steer the vehicle round curves with 100% reliability.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Sat Jan 04, 2014 5:05 pm, edited 1 time in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 11:43 am

In fact my first loco in P4 was the DJH kit of an 0-4-4T Caledonian Railway Class 439.



This I arranged in an adaptation of the normal Flexichas chassis. The outer driving wheel is fixed as usual, but the bogie supports all the weight of the rear of the loco, the inner driving wheel being sprung by top wiper pick ups and some music wire acting on top of the axle. The other end of this wire is soldered down towards the bogie. The bogie wheels are each individually sprung, not very strongly but sufficiently to hold them down on the track. The strength (gauge) of the music wire determined whether the front driving wheels had sufficient weight on them to guide the loco. Too strong and that guidance was no longer assured - confirming for me the importance of the weight on the outside wheels of any model locomotive.
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 11:50 am

2) Lesson in 0-4-4 bogie guidance

I learnt something here by accident. The Class 439 being my first P4 chassis, I made up the Gibson milled frames with his P4 spacers. I arranged the inner driving wheels to have sideplay with just the thinnest Romford .005" washers, (the outer ones being washered up to running clearance only) but after I had made the bogie and come to running trials I found the loco would not go round my test track 4' reverse curves without derailing. After thinking about the foregoing weight distribution issue I realised that there simply was not enough sideplay - and that there was no way I could make the frames narrower without starting all over again.

(Of course a big lesson here was not to take the width of any spacer on trust, but to carefully work out how wide the chassis needs to be to allow for the necessary side play. The chassis on my models is of course rather narrower than the prototype to allow for unprototypical 4 foot curves, but I find that a bearable compromise.

0-6-0 locomotives frequently have an asymmetrical arrangement of the wheel spacing, so this can reduce the amount of sideplay needed. I still have to obtain a formula that I understand that allows for this, so far 1mm total sideplay has sufficed for me (= 1/2mm each way) but the less sideplay needed, the easier things like brake blocks become – and, especially, outside brake pull rods!)

For a while I thought the too-narrow frames a terminal problem and that the loco would have to be restricted to very shallow curves, but then it dawned on me that if I shifted the bogie pivot point to nearer the inner driving wheel the wheelbase would effectively be shorter and thus less sideplay would be needed. So the bogie (which had been pivoted at its midpoint) was fixed – not pivoted - to a 'steering arm' (for want of a better term) which is only free to pivot around a pin situated a little behind the bogie towards the driving wheels. A roller was arranged on top of the bogie to bear the loco weight, so that the bogie was free to swing.

This arrangement, I suggest, effectively replicates the sideways movement normally expected of the bogie in the prototype, but in a more positive and controlled way than the guidance spring that Mike Sharman suggests in his Flexichas book for this situation.

The bogie wheels springing arrangement is four rather weak pieces of nickel silver wire bearing down on each wheel bearing which is free in a slot. This was rather tricky to install and is not really satisfactory, the springing being quite flimsy because of limitations of space, though so far the bogie has never derailed. In future I decided I would have the much simpler compensated system as per Mike Sharman's drawing in the Flexichas book.
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 12:04 pm

3) Lesson in bogie pick up

The Ex-CR Class 439 loco ran well when eventually complete, with pick ups on the driving wheels. For experiment's sake I tried adding pick ups to the bogie. Try as I might, I couldn't arrange them to pick up without adding such drag that the loco's haulage power was dramatically reduced. It was a relief to take them off and find the loco could again haul a decent load - the minimally improved low speed control was definitely not worth it.

This lesson was replicated when I bought a second hand P4 Ex-CR Class 113 4-4-0 tender loco which had pick ups on the tender wheels. The locomotive could barely even pull the tender without slipping. I removed them, adjusted the Gibson springing, added weight and a coupling to transfer that weight to the loco, put pick ups on the driving wheels, and it now can haul anything required on "Calderside".


So my lesson here was that electrical pick up needs to be from the axle on unpowered wheels, bogie or tender.
You do not have the required permissions to view the files attached to this post.

User avatar
Andy W
Posts: 884
Joined: Thu May 21, 2009 8:11 am

Re: In search of smooth running: my lessons in loco construc

Postby Andy W » Sat Jan 04, 2014 12:12 pm

Very interesting Julian. Nice work. Rather than springing the bogie, have you ever tried having one frame of it pivoting so that it forms its own 3 point compensation unit? Much easier to set up and no maintainance needed.
Make Worcestershire great again.
Build a wall along the Herefordshire border and make them pay for it.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 12:14 pm

4) Lessons in smoothly revolving coupled wheels.

I knew fine well the importance of the coupling rods being exactly the same length as the distance between the wheels, from Iain Rice’s books and others. However there is more to it than that.

Between my Terrier and my Class 439 I made a 4mm scale 009 model of the Welshpool and Llanfair “The Earl”. Because the kit came with flycranks I used them without examining them. If you look carefully you may be able to see on the photos that the centre wheel crank is a different shape. In the kit this was catered for but I didn’t see that the distance from the axle centre to crank pin was different on the two types of crank.

The flycranks had to be soldered to the axle ends, so the usual adjustment of trial and error for quartering was thereby more complicated. Did I have fun….

What I learnt (mostly from David Franks – LMS Models - many thanks again David for your generously given advice!) was that for a locomotive to work the crank throw must be EXACTLY the same on all wheels on the left hand side of the loco, and EXACTLY the same on the right hand side of the loco. It doesn’t matter if it’s slightly different on the left side than the right. Additionally, the angle of the cranks, if it veers a little from the perpendicular, that won’t matter terribly much, so long as all the cranks on one side of the loco have the same inaccuracy: the same applying to the other side.

So I carefully made new cranks in two sets of 3, with a pillar drill to drill the holes, so each set was identical except in the eventually filed shape, that being of no importance mechanically on the model. Crankpins were soldered on carefully (with high temperature solder) as near as dammit to the perpendicular as I could manage. Again I tried to get them quartered without success, and realised that it was utterly hopeless without a jig to ensure that on all pairs of wheels the cranks were at exactly the same roughly 90 deg angle relative to each other. Making a jig was just a couple of hours work, a tiny proportion of the time I’d spent trying to get smooth running by trial and error. Immediately on using it to solder the cranks in place I had smooth running.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Fri Jan 31, 2014 9:30 pm, edited 1 time in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 12:35 pm

In fact my very first chassis for the Terrier had been a rigid unpowered one, using Romford wheels. Even for this I had carefully set the axle bearings with the rods using jig axles. When it came to fixing on the ready quartered wheels and rods, the very first few revolutions had a slight tight spot, but that soon went just by pushing the chassis along the rails. So I learnt that even on a perfect chassis there will be a tiny bit of a tight spot, and the best way to get rid of it is just to run it.

I have subsequently found on other chassis where I was dissatisfied, that further easing the holes of the rods beyond a free movement on the crankpin makes no difference. If you’ve done all the thinking in advance and used a jig to quarter it, there’s no improvement to be made fiddling around with it. The only way now to get smooth running is to just run it in. For this a rolling road or autoshuttle is needed – I have the latter: the loco reverses automatically at each end of a 6 foot length of track, thanks to light sensitive diodes from All Components.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Sat Jan 04, 2014 4:10 pm, edited 1 time in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 12:59 pm

5) Lessons in back to back and wheel fitting

One of the reasons I moved to P4 was that I didn’t think the effort involved was much greater than what I was putting into my 00 gauge attempts. For me the chief difference in difficulty of construction is the much finer tolerance of getting the correct back to back dimension when fitting the wheels. It is necessary to avoid damaging the wheels and causing wheel wobble by assembling them in one hit only. This is really a subject on its own, of techniques, suffice to say in this context that obviously the correct B2B is the other critical part of not getting derailments particularly on points and crossings given that the track is correctly gauged. At the same time the correct amount of washers to allow running clearance only have to be installed on the corner wheels.

It will be clear that I consider the use of a jig for quartering the wheels absolutely essential, and the GW Models jig quarters and gauges the wheels in one hit. It has fine adjustment possibilities, and I use them to set the B2B, in conjunction with my (to me) essential digital vernier, just using the actual gauge for checking. If one closes Gibson wheels in the jig using the B2B gauge as a stop I find they are in fact set too narrow – because of the stiffness of moving the wheels on the axle.

Each loco that I make I learn a little more about this critical process, and this I still find the most challenging part of the procedure.

6) Gearing.

To me the Achilles heel of model railway running is the low speed control. To get the impression of the weight, of the inertia, of the prototype is pretty difficult. My approach is to gear all my locos for a maximum speed of around 30mph – whatever the maximum speed of the prototype. Shunting locos I gear lower for around 20-25mph. This is in order that I can have the best possible low speed control. In most cases this has been High Level gears, with the 80:1 or 108:1 ratio.

If I was making locos for a layout where the speeds were higher, I would gear them higher. But most layouts are the short station type – certainly the ones suitable for exhibiting. Few have trains rushing through at 60mph. I don’t see the problem with running the motor to its maximum voltage, and I don’t see the point of making models able to run at a speed that they will seldom achieve. Only real engineers like Chris Pendlenton can apparently get both the proper top speed and low speed control, and for the moment I can’t get up to that level.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Sat Jan 04, 2014 4:45 pm, edited 1 time in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 1:18 pm

7) A recent lesson

My Ex-CR Class 782 chassis was made in an unusual way. The details aren’t important, but I made the rods (Gibson “universal” rods) fit the chassis, not the other way round. The normal advice is to treat each 4-coupled section independently. I have found I can have a loco where each group of 4 wheels seem to run sweetly, but not when they are coupled together to form the 6 or 8-coupled loco. I noticed making this loco that if I put all four rods on (sorry, I still joint mine over the crankpin) the jig axles (which were not the tapering type), that one did not go on so readily as it had done when just doing it as a pair with its opposite half. So with all four rods on the jig axles I held the iron to the one that didn't sit quite so happily, so that it adjusted its length – it was a minute amount: but this loco runs the best of any of my 6 coupled locos.

Perhaps that is helped by the fact that I used a pillar drill to drill all the crankpin holes (the wheels came with just dimples for crankpin holes), holding each wheel on a jig that ensured the crankpin hole was exactly the same distance away from the axle on every wheel, and ensuring its verticality.
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 1:37 pm

I have spent some long rail journeys just around Christmas 2013 skimming through quite a lot of the Forum for the first time. It seems that CSB suspension is seen as the way to go. I find it interesting that fairly early on in Will L's masterly exposition of the whole subject he says that there will tend to be more weight on the outside wheels of a chassis (or ‘corner’ wheels in my terminology). Just the same as I have concluded works best for compensation. However I haven't properly studied all Will's writings to see if that continues to be a feature of CSB suspension.

I see that the 4-4-0 is seen as a tricky candidate for CSB treatment, just as it always has been seen tricky for the traditional compensation. I have been thinking about how I would do the LMS Compound’s suspension since I bought the semi-made kit in 2011, and this June and November I had free those few days necessary to concentrate properly to get a working chassis. So if I try CSB it will be in a future project.

Just before describing what I did in a future post, an aside on the subject of the chief benefit claimed for the CSB - the smoothness of the vehicle on the rails. For me, in the heirarchy of problems in P4 running, there are other ones higher up. For me, the problems are:
1) derailments
2) jerky stops and starts
3) disappointing operating.

1) - how many of us have seen a real train derail? I have only once and that was on a narrow gauge steam line in a siding….Ahem, I was driving it…Appropriately enough in the context of my subject, the reason was that, unbeknownst to me and very unusually, one of the wheel spring hangers had come right off, so there was no weight at all on the front left wheel ….. we left the track on a right hand bend…
Yet derailments are all too common on some of our layouts.

2) - Seldom do trains start and stop in a way that shows their inertia and momentum like real ones do, and this is often because locos simply can’t run slowly smoothly - it's very difficult to achieve.

3) - Even on the most perfectly made layout and with the best made stock, some people still tend to operate trains like "electric mice" as model trains are caustically referred to in preserved steam circles - a (say) 5 coach train will be seen accelerating to 30 mph within a platform length or decelerating as rapidly.

Compared with these faults I find the slight bumps of the compensated loco very minor. Iain Rice in his chassis book talks about this and says that if this really bothers you, there is a way round it by using twin beams instead of the fixed axle – though whether that gives such bump free running as CSB is obviously a question, and a narrow gearbox is required.

I guess that there isn't that much more work to do in either approach, and I'm not declaring myself to be fixed in the compensation camp. So far, for me, the principles of compensation are more easily to understand and implement, in a kind of subliminal way, though perhaps that is because that is what I have done so far.
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 1:44 pm

8) Hornblocks

Because it came with inaccurately spaced holes, my Q1 chassis I originally made in a convention defying way. It didn’t work for long. I had to retrospectively fit hornblocks, which themselves are quite convention defying, but only in manner of construction. The essential thing is to keep the revolving axle wearing point separate from the sliding suspension wearing point. Mine here are made with square tubing containing a round tube of inside diameter 1/8th inch.

This rather unusual arrangement was only necessitated by my to desire to avoid dismantling the frames altogether. However, it does work!
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construc

Postby Julian Roberts » Sat Jan 04, 2014 2:01 pm

"Ealing": - thanks, certainly won't make a bogie like that again. LMS Compound bogie will follow sometime soonish...

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: In search of smooth running: my lessons in loco construction

Postby Julian Roberts » Wed Dec 27, 2017 8:12 pm

It suits some people seem to suggest I never actually do anything but obsess about hundredths of a millimetre. Just to put the record straight, here is the list of locos made since around 2007

1. Illustrated previously on this thread: CR Class 439 0-4-4T; LNER Q1 0-8-0T; CR Class 812 0-6-0; CR Class 113 4-4-0 (loco and tender chassis upgrade)

2. Picture attached CR Class 782 0-6-0T; LMS Compound 4-4-0 (unfinished); NCB 0-6-0T Barclay. Current build: LMS Crab
You do not have the required permissions to view the files attached to this post.


Return to “Chassis and Suspensions”

Who is online

Users browsing this forum: ClaudeBot and 0 guests