Friday, 30 November 2012

Magnetic Coupling Trials

I spent a couple of hours today, despite the freezing weather, preparing a length of garden track and running a 12-coach test train round the reverse curves on a 1 in 50 gradient. This was the first fully "live" trial of the "McBogle (or DOGRF) Magnetic Coupling" (see previous post). The couplings between coaches 1, 2, 3, 4, 5 and 6 were to the Mk.1 design using a single 3mm x 2mm magnet on each coupling. The couplings between the loco and the first coach, which were taking the greatest load, were Mk.2, with an opposite-polarity pair of 2mm x 2mm magnets. None of the couplings failed during the trials, which comprised several low-speed runs up and down the test track. (Higher speeds were not possible this time because of undiagnosed limitations in the control or  power distribution equipment, i.e. it couldn't supply enough current for a Hornby class 50 pulling 12 coaches up 1 in 50 at more than a scale 20mph.)

  1. Both the Mk.1 and the Mk.2 magnetic couplings have a more than adequate holding force for a 12-coach train of 160g coaches on a 1 in 50 gradient with reverse curves of radius 4ft and greater. Tests were carried out at scale speeds of 20mph and lower.
  2. The Mk.2 is easier to use, as identical couplings wll mate with each other. Before the start of the trial it was necessary to change the Mk.1 couplings between the loco and first coach for Mk.2 because one of the Mk.1s was the wrong polarity.
  3. Further tests are needed to assess the long-term reliability of the adhesive joint holding the magnet to the "pipe", with the stress of regular coupling and uncoupling.
  4. (Not related to couplings.) An unexplained gap between two lengths of track on a reverse curve has now grown to about 5mm (it was about 3mm in the summer). It is now causing regular derailment and needs investigating.
  5. (Not related to couplings.) A rail joiner near the top of Quarry Bank has no wire jumper and is not making good contact.

Thursday, 29 November 2012

Magnetic couplings.

An important decision that all railway modellers must make is which types of coupling to use between vehicles. There can be many factors affecting the decision; cost, time, realism, ease of coupling and uncoupling, .....

For my own set of priorities, I came up with the following guidelines:
  • Front of locos to be as realistic as possible, with all pipes, and hook-and-screw-link coupling.
  • Rear of locos to be as realistic as possible, while retaining a working hook-and-chain coupling and an NEM pocket to allow fitting of a coach-compatible coupling.
  • Pre-1970ish freight stock to have working chain-and-hook or instanter couplings
  • Passenger stock to be as close-coupled as possible, but I hadn't found the ideal system.
The locos and freight stock were therefore largely properly equipped, but the passenger coaches were fitted with a selection of couplings including all variants of the industry-standard "Triang" tension-link couplings (both NEM-fitted and non-NEM), Hornby/Seuthe close couplers, and the recent Bachmann "Vacuum-pipe" fixed close couplings. Interchangeability was achieved with some difficulty by retaining a selection of spares of each type and moving them around the NEM-pocketed coaches as necessary. Others had been tried and rejected.

A thread appeared recently in the "00 Garden Railways" forum entitled "Bachmann Coaches - Close Coupling System". This referred to an article in issue 204 of "Traction" magazine proposing the use of high-strength Neodymium magnets to make the Bachmann "Vacuum-pipe" couplings more readily seperable, and describing the writer's own experiments with the technique. For me, this sounded ideal for the coaches; all the advantages of the Bachmann "pipes", without the need to invert the entire train to uncouple!

I had previously carried out tests and calculations to establish the force required to pull trains up the 1 in 50 gradients on my railway, which showed that the horizontal force required from the locomotive for 12 coaches of average weight 150g up a 1 in 50 gradient is about 46g. This is beyond most RTR 00 steam locos, but for modern RTR diesels it is no problem. Using 2mm dia x 2mm Neodymium magnets from Magnet Expert, I lifted gradually increasing weights, and 55g was the highest that held reliably. This was with a sheet of gummed paper between the magnets to attach one to the weights, while the other was glued to the end of a shunter's pole.

So with a straight pull, 12 coaches up 1 in 50 is feasible. In practice, there will be misalignment, shear forces, and shock loads to contend with, so a stronger magnet would be desirable. But it was close enough that I resolved to carry out further experiments.

I acquired some 3mm dia x 2mm neodymium magnets with a specified 250g in-line parting force and 50g shear parting force. I found some 4mm heatshrink tubing and some Loctite 406 adhesive (for plastic & rubber), and bodged together a pair of magnetic couplers. They were badly made, and the magnets didn't make good face-to-face contact, but I tried them anyway. One went on a loco, the other on a coach, and they were coupled together. The other end of the coach had a cord attached, which ran over a pulley to a hanging weight tray. The weight was increased up to 87g, which is equivalent to pulling 25 150g coaches up a 1 in 50 gradient. I shook the coach and the weights to give a bit of shock loading, and still the magnets held.

I was now more than confident that the technique would work with plenty in reserve, providing I could find a successful method of attaching the magnet to the remainder of the coupling, which would hold when the couplings are repeatedly pulled apart to separate the coaches. The following technique seems to work. I first glue the magnet and pipe together using fast-acting cyanoacrylate adhesive, followed when that is dry with something thicker and more reslient such as Loctite 480. When that has cured, I slide a 5mm length of 3mm ID 3:1 heatshrink sleeve over the joint. Shrink the sleeve, squeeze in a little superglue, and hey presto! It's done! There's a picture below of a pair of coaches coupled up.

I measured the pull-apart force of a randomly-selected pair of couplings fitted to a loco and a coach. At 180g, it was still holding. That's equivalent to about 56 coaches up a slope of 1 in 50, and is much more than any of my locos could pull, quite apart from any other considerations!
There is one other minor matter to consider when making and using these couplings. To have a pair of magnets attracting each other, one must have the N pole of the magnet outward from the coach end, and the other must have the S pole outward. So we need to make equal numbers of each type of coupling (which will happen automatically if made in joined pairs), and make sure they are fitted in the appropriate places. Those who wants their coaches in a certain order in the train and pointing in a certain direction, may need to swap round some of the couplings to remarshall a train. Maybe not a problem, depending on how you use your coaches.
To resolve this N/S pole issue, I tried a variation on the design using two smaller magnets (2mm dia x 2mm) of opposite polarities side by side. In this way, the couplings at each end of a vehicle are identical, so the vehicles can be coupled in any orientation without changing couplings. I've tested a pair of double-magnet couplings, and they were still holding strongly at a pull-apart force of 190g! That's almost twice what my strongest loco can pull before the wheels start to slip. (Note that 190g is the horizontal force in line with the loco, or the weight it could lift or support using a horizontal cord running over a pulley. It is not the total weight of the train it could pull, which is much greater.) I'll post a full report when I have some pictures.

The advantages of this design (which I have designated Mk.2) over the single-magnet version (Mk.1)are:

- All couplings are identical and interchangeable
- Smaller magnets - looks neater in side view

I have also devised a simple production method which automatically holds the pairs of magnets in exactly the right orientation for assembly, without the need for any special jigs. Again, a report will follow when I have some pictures.

Further developments? Well, a better design variant is still needed for fitting to locos, and for use on vehicles without an NEM pocket. And it's possible that an even smaller magnet (2mm dia x 1mm) would still give enough strength for our purposes, while still retaining Mk.2 compatibility but requiring less force to pull apart. Watch this space!

Monday, 19 November 2012

... and In the Open

A few engines have been in the workshop recently for painting, weathering, numbering, detailing, performance optimisation, etc. The two closest to completion needed a test run on a longer track than I have indoors, so here they are!

The "Standard 5" 73054 has been modelled on photographs taken on the day she hauled me (and others) from Evercreech Junction to Bournemouth (West) (see previous post). She still has a few jobs awaiting, including fitting the front steps.

The diesel unit, known as a "class 121", has just been fitted with new wheels, as the flanges on the original wheels were so deep that they rattled along the rail chairs.

BR standard class 5 4-6-0 no. 73054 runs down the railway straight.

73054 further along the railway straight.

A diesel unit, cascaded from Paddington suburban services, on the quarry bank.

Friday, 16 November 2012

In the Paint Shop

I have spent a few hours this week working on nice, shiny, ex-works model engines and making them look scruffy. Am I going mad? Well, maybe, but my excuse is that this is how they really looked in service. Take the example pictured below. This sticks in my memory (assisted by my photographic records) as the engine that pulled me and some friends from Evercreech Junction to Bournemouth West on 27th March 1965. What made this engine memorable was that it was exceptionally clean and shiny, and it had a multi-tone "chime" whistle. However, looking now at the pictures taken that day, although most of the green paintwork had been cleaned, the rest of the engine still had a thick coating of grime and rust.

73054 on the work-bench during detailing and "weathering".

73054 at Evercreech Junction, 27/3/1965.

Saturday, 3 November 2012


Today the weather was reasonably dry, the railway was covered in leaves that needed removing before they turned into compost, and I needed some fresh air and exercise. I also had a few things to test on the garden railway which couldn't be done with the facilities I have indoors.
  • Move the wi-fi network extender to the summer-house and use it as a local-only network with a stronger signal everywhere on the railway.
  • Try out an old ipod given to me to use as a second walk-about controller with WiThrottle.
  • See how well Tornado performs after modifications to improve on its original 3 coach limit up the 1 in 50.
The trials were all successful. A train can now be controlled from anywhere on the railway, with no dead spots. The ipod controlled trains - for five minutes after which the battery gave up. At least I now know it's worth considering whether to spend the £59 needed to fit a new battery. And Tornado romped up the 1 in 50 quarry bank with eight coaches.

Tornado romps up the Quarry Bank, watched by the local wild-life.

Tornado, about to enter the tunnel. (Photo with permission of the railway authorities.)
As can be seen from the first picture, the trackside ground-cover plants are doing reasonably well - especially since some of them were not planted until the end of the summer. The fourth green patch from the camera is the most interesting; it's self-set moss, and it seems to be spreading nicely.

Having said that Tornado managed 8 coaches up the 1 in 50; it did twice, then the air suddenly got colder and damper as dusk approached, and from then on the engine started to slip as soon as it got a couple of coaches onto the bank, even when I reduced the train to 7 coaches. I can only assume that a thin film of water was condensing from the air onto the cold rails, and reducing the adhesion. Bachmann don't yet fit their engines with working sanders!