New Zealand Railways Wa Class
Deep-End Learning
A modelling odyssey hidden within an innocent-looking S scale loco kit
J.G. Gardner Models Wa Kitset in S Scale
July 1999 - and still going

Building this kit has been a long journey for me, laden with frustration. While there were some things "wrong" with the kit in terms of how it's designed, as well there being a few missing parts, most of the problems were down to my own inexperience.

My Dad bought the kit in 1999, but never started it. He'd always wanted to make a model locomotive and thought this would be an easy first project perhaps, but just never got around to opening the box. The kit has a special meaning for me because I started building it while caring for him in his final weeks of life in 2005. I assembled the frame and drivers but then packed it away. I couldn't get it to run smoothly, and gave up.
What the hell?
“The thing that really ripped my nightie over the years I worked on it, was that the whole mechanism ran very smoothly, until under load. Then it hitched along like a ruptured tortoise."
Looking back, I know that Dad would have been sorely tested to make it. The way the mechanism is designed, how it all goes together and the missing parts meant I ended up re-making a lot of it. There were the usual beginner mistakes because I'm an idiot, and because some of the kit design just encourages running problems.

When I mention this kit to modellers I usually hear, "Yeah, that one's a bit challenging..."
The thing that really ripped my nightie over the years I worked on it was that the whole mechanism ran very smoothly, until it was under load. Then it hitched along like a ruptured tortoise.

The actual kit construction seemed fine, but something was obviously wrong. I pretty much rebuilt every part of the mechanism trying to get it right.

Bright and Shiny
It all went together easily. Until I tried running it.

Mad Scientist
One of my many gearbox attempts. This was an attempt to combine gear reduction and moving power to the centre axle. It worked, but was very noisy and didn’t fix the problem.
The kit has the motor and gearbox rigidly mounted both together and to the frame with no universal joint, which is bad engineering. I added a universal joint from NWSL, but that didn't fix the poor running. I had all sorts of crazy theories. I even thought that maybe sending power to the front axle could exacerbate misalignment. I made a whole new gearbox to drive the centre axle instead. It was an interesting project, but didn't fix the problem. I made several versions of this...

After a bit more testing, I also found the motor was faulty with a slightly jammed armature, causing a very high current draw.

I ditched it and got a nice new one from NWSL.
By now, I was completely at a loss with what to do. I read up about model loco construction and endlessly checked the mechanism by hand. The equalised suspension on the kit bound up very easily under load. The back axle seemed to jam and lift up. After a lot of fiddling I just ditched the whole thing and put independent springing on the second and third axles, with harder springs on the rear. The front axle is still unsprung.

I rebuilt the frame and suspension twice, but that didn't fix it either! In the background, Philippe, my good friend in Vancouver was sending words of encouragement. Most of the pictures in this post were originally sent to him, with captions along the lines of, “well, that didn’t bloody fix it either! Now what do I do?”

Missing Bits
The Crosshead castings for the Wa version were missing, so I had an early introduction to scratch-building. I made the crosshead guide clearances too tight and that gave a lot of trouble.
For an extra challenge, the Wa cylinders and crosshead castings were missing so I needed to make them. Not understanding how valve gear all works together, I didn't know to allow enough crosshead travel. I also made the clearance on the guides too tight.

Some of the valve link castings were not completely formed as well, which meant I had to make them, too.
It was only after years of tinkering, during one dark Norwegian winter's night (and they're pretty dark) that I found the core of it. By taking each driver crankpin screw out in turn and rotating everything by increments, I noticed that one driver had a different crank eccentric distance. The measurement from the axle centre to the crankpin hole was about 0.8mm smaller than the other five wheels. A remade injection tool, maybe? Doesn't matter, but it meant the mechanism ran like it was out of quarter, even when it wasn't.

The Culprit
That rear driver was the cause of my problems, even though, when I took this photo, I didn’t know it yet.
If I'd checked them at the start, the rest of the kit would probably have gone together fine. On the good side, it forced me to learn many of the fundamental rules around model loco construction. But, the fun wasn't over yet.

By the time I discovered the faulty driver, they were all mashed and reworked beyond repair, the result of years of unscrewing crankpins and generally fiddling about. I managed to get a replacement set ordered from New Zealand but, in a hilarious twist of fate, one of them was eccentric.

I was pretty grumpy, so I thought I'd try something else. I searched on the web and found some lovely HO brass drivers from Greenway Products in the States.

I also needed to take a fresh approach, and throw out my assumptions. The thing I love the most about railway modelling is the diversity of skills needed, and the willingness of modellers to share knowledge. A guy in the North Vancouver modelling scene I'd met once through Philippe had built many amazing models and gave me, via email, hugely valuable advice about running tolerances and the critical points to check. Thanks!
I had learned to hand-file brass to close tolerances, but didn't know how tight was too tight, or too loose. There didn't seem to be much on the internet about it, so these clearances and notes are from his advice and based on my own experience. They’re up for debate, but worked very well in getting this model running:

1. Bearing to axle clearance should be a close running fit. No more than +0.02mm, otherwise the axles seemed to drop out of line.

2. The bearing to frame-slot clearance controls how square the axles will be held, and is critical. I filed mine to a sliding fit of +0.05mm - more clearance makes the axles loose in the frame (same as the bearing clearance) and any tighter jams them up.

3. Crank pin to siderod hole clearance should be about +0.1-0.2mm - I'm not really sure about this one, but any more than 0.2mm clearance seems to indicate something else is not right. Find out what’s wrong, don't make the hole bigger!

4. All axles must be
exactly parallel with each other, and square to the frame. This is where I made the biggest mistake of all at the start. I filed the bearing slots in the frame oversize, then decided to solder shims to the bearing blocks to fix it. I wasn't accurate enough, which made each bearing a custom-fit. It also changed the axle spacing which meant I had to make new side rods. Duh.

5. When I went right back to basics, I needed to fix point #4 above, with new bearings and new brass bearing guides. For alignment, I used 3mm x 300mm diameter hardened steel dowel pins threaded through the bearings and located in a jig. The long dowel pins exaggerated any misalignment and allowed the bearings to be held securely during soldering. I did the front axle first and used it as my datum.

The Right Tools
As well as buying some driveline parts from NWSL, I got their quartering tool. It made a world of difference.
By the time we were due to leave Norway the Greenway drivers had arrived. I had just enough time to rebuild the frame and bearings before we left.

I fitted the new brass drivers with a NWSL quarterer, which is an awesome tool, if you're thinking of getting one. The chassis ran beautifully!

But, it wasn't over yet.
I took the kit back to NZ with me on the plane, ahead of our other stuff that was in a shipping container, hopeful that a few hours spent adding the valve gear would see it running. In Christchurch, I was fortunate to meet up with a very welcoming group of modellers who meet once a month to chat about trains. When I went to shown it to them running, I found I'd mounted the drivers on one axle backwards, causing a dead short. That was embarrassing, let me tell you. However, they encouraged me to carry on and fit the valve gear. I did so, only to have it jam up solid.


I could see that the crossheads jammed against the cylinders. Maybe the crank eccentric on the new drivers was too big? I checked the drawings and saw the new brass drivers were actually too small for a Wa anyway. The driver diameter of 3' was written on the W drawing, but on the web I found Wa loco drivers were slightly bigger. I'd been very lucky to find another Gardner Wa kit a few months prior and decided to pinch the drivers and cylinder castings from that.

After checking the plans further, I spotted that the cylinders and valve bracket were actually mounted 1.5mm too far back, when compared to the drawing. I never noticed it before and dunno why this was so, considering both mounting points are fixed in the kit. That was the reason the crossheads jammed up when the driver cranks rotated all the way forward. It wasn't to do with the drivers. There was only a fraction of a millimetre interference, but that was enough.

To add to the fun, I then found the front crankpin screws hit the inside of the crossheads. The new cylinder centres were about 0.7mm too narrow, or the new wheels too fat, or the crankpin screws too long, or some damn thing.

Either way, I fixed it by thinning down the screws and widening the whole cylinder chest. I'd had a guts-full by this time and just threw a bit of brass and epoxy at it. The cylinders are now a couple of scale inches wider than prototype but I don't care. I'd much rather have a loco that runs.

With a new gearbox, universal joint, torque arm, new motor, completely new suspension, replaced bearings and guides, new springs, cylinders (widened and in the correct location), reworked crosshead guides, new drivers and the reversing bracket in the right spot, it ran ... like a dream. Surely not, I said to myself. But I looked again and it obstinately refused to run like crap. I kept going out to make a cup of tea and sneaking back in, as if to catch it by surprise.

After fourteen years of head-scratching, rebuilds, redesigns, threatening to smash it with a hammer, de-soldering, burned fingers, swearing and frustration, I had the finished chassis of my NZR Wa class loco running like silk on the test track. It was getting juice from two AA batteries but I wasn't fussy. The bloody thing was running.
To celebrate a mechanism that actually ran, I pulled some of the old brake piping off the shell and added Precision Scale Co. cast brass elbows, as a treat.

I've decided to snip the coupling pins from the Kadee couplers so they can be mounted properly under the buffer beams instead of sticking out over the catchers, which was my first solution but looked stupid.

Adding a separate universal joint to the original 16x32 motor meant I needed to push the backhead rearwards by about 5mm. With the shorter (but more powerful) 20 x 25 motor I can return the backhead to where it should be, and add a bit of detail too.

When we're back from our trip and I have my workshop set up again, I can finish the body and reconnect the decoder. Then I can finish this post and make something else. Until then it will have to wait.

I can't wait to see it painted and weathered, trundling along the rails with some La wagons rattling along behind. I can sense Dad smiling over my shoulder.
Sign up to my blog
I'll send you a message whenever there's something new to read
I do not share personal information with any third-parties, nor store information collected about your visit to this blog for use, other than to analyze content performance through the use of cookies, which you can turn off at any time by modifying your Internet browser's settings.
I am not responsible for the republishing of the content found on this blog on other Web sites or media without our permission. This privacy policy is subject to change without notice.
Made on