A day of modelling progress

Thanks to Mark Tatlow for hosting today, and to Oly & Chris for great company. I think all four of us made progress on various projects. Mark was making holes in the side of a nice looking Gresley Buffet, and had made them disappear again by the end of the day.

My own projects were to try and complete the build of the MERG DCC system for SLAG’s Bankside project. I had to cut a new support for one of the PCBs and jigsaw it all into the case. I succeeded at the first task! Re-assembling the boards (it is a bit of a jigsaw puzzle), I discovered one of the internal connections wasn’t mating correctly, so that will need replacing in another session.

Here’s one of the boards, captured to record which colours I’m using for connections (note to self: green: 0vDC, Blue, +5vDC):

PCB and wiring harness

This one replaces the buzzers the system uses by default to tell you when there is a short or other fault. By translating that information onto CBUS, we’ll be able to have buzzers or LEDs at any control station. Hopefully those will be more useful in noisy exhibition halls.

For the afternoon, I dusted down the Tornado chassis I’ve been looking at for a while, and assembled two more links for the Walschaerts valve gear.

It’s not about the gauge

P4 is not about the 18.83 gauge. Side on from the track, I find it hard to spot the differences between plain hand built track in 16.5mm, 18.2mm and 18.83mm. P4, for me, is about the other dimensions in the standard.

Firstly, there’s a nicely defined wheel profile that doesn’t look like you’re building a steamroller. This profile also helps when squeezing all the parts in a chassis together so it looks like a scaled down version of the real thing.

Secondly, the variously named check rails found in crossings are close to the scale distances away from the rails they lie next to. This makes the track-work look like a scale model, rather than a toy. Once the other details possible with hand built track are used, it can look almost like the real thing.

I think it’s a shame people often label layouts built to P4 standards as “18.83” layouts – to me it misses the point.

Why I chose P4

I’m going to be giving the “Getting Started in P4” demo at Scaleforum next Saturday. That’s got me thinking about why I chose to get started in P4.

For me, what started it was this picture of my first completed model railway:

Trackwork on my 'Phoenix Yard' Layout

I was taken by the view of model railway track, and also struck by the crudeness, in scale terms, of the mass-produced track. I’d always been aware of ‘finescale’ model railway standards, and started to look into them with more interest. Could I build track to those standards, and achieve something that would satisfy me more as a model?

If I chose to stick with the 4mm/ft scale used by the OO models I already had, I had three choices – build finer track to OO standards, move to the EM standard, or use the P4 standard.

I’ll hold my hand up, and say that at an initial glance, it can be hard for me to tell apart good examples of each. There are tell-tale signs, but I have to look for them.

I was struck that once I ventured into hand-built track, there would eventually be a desire on my part to use finer wheels. In OO this seemed to be a can of worms – there are many subtly different tolerances between track and wheel, and choosing which to adjust to achieve the finer look I desired seemed to be a bit of a minefield. There are several documented OO ‘fine’ standards, and between them modellers I observed seemed to occasionally run into surprising difficulties with things that should work, but didn’t. Sticking to a well documented standard seemed to be the solution that people had most success with.

Considering EM, the appeal seemed to be that it has high enough tolerances that some mass-produced models would ‘just work’ on the track after a minimal easing of the wheel gauge. Additionally, those larger tolerances might prove more forgiving to my ham-fisted modelling skills. As I looked further though, it seemed that there was very much a gradual process of rebuilding embarked on by the more successful modellers, to finer wheels, more flexible chassis designs, and therefore fairly tight tolerances in their models.

So, to P4. Here I would probably need to worry about flexible chassis immediately, and would certainly have to work to the fine tolerances. I rapidly established that this was a matter of learning to use hand tools & jigs well, which didn’t seem beyond my ability. There was also the advantage of a single standard, rather than several interconnected ones, and that appealed to me.

Given that both the other ‘fine’ approaches to 4mm/ft models seemed to end at hand built track and stock – with the better models having flexible chassis – there seemed to be no particular advantage to either a ‘fine’ OO or EM, when compared to P4. Either way, I’d need to learn to build track to fine tolerances with jigs, and eventually need to build quite complicated chassis. Given those conclusions, choosing P4 meant my models would be closest to scale. Additionally, P4 had the easiest route into hand built track, with the ridiculously easy to build Exactoscale Turnout Kits – so P4 seemed the best choice for me.


Over the last couple of posts, I’ve been building up to completion of a DCC system assembled from MERG kits.

Last week, I had the pleasure of taking it to the South London Group’s clubroom, to trial it with our new layout. It worked!


What we can see here is (starting at the loco on rollers), a class 20 on the ‘layout’ output of the system. To the right is the programming track, and alongside that we can see the command station. This provides the DCC signal for the layout, and a direct connection to the programming track. The command station has no directly attached controls, and cabs (one is at the bottom) are connected over CBUS. Since CBUS is a publicly documented standard, I intend to develop custom handsets over time. I’m intrigued if exhibition layouts will be easier to use if the controls are designed with the operation of the particular layout in mind.

To the right of the picture, the top board is a 5A DCC booster, which provides the actual layout power and DCC signal broadcast. Below it is a USB to CBUS connection. Not directly needed for the system to function, it will facilitate future firmware upgrades to the command station and handset, and permit me to prototype the custom handsets on a PC before committing them to electronics.

Surface Mount Practice

MERG provide a couple of kits intended to be used as practice for surface mount soldering. This is my second go at this, and the first working attempt:


My first attempt failed because I wasn’t familiar with how pin 1 on the IC was identified. I now know it is the chamfer along one side. My first attempt had the IC the other way round, and rapidly burned out without flashing the LEDs, as this circuit is designed to do. For a £3.60 kit, it seemed like a good value way to learn the lesson. The instructions did explain all this, but somehow it still didn’t sink in without the practical lesson!