Rebuilding the Arnold G8 into a fiNe-scale machine
Introduction
Here we see the original loco on the right hand and the new rebuilt on the left. As you can see some small items are still to be done such as the new numbers on the smokebox etc.
The original Arnold model has the motor partly in the cabin and firebox. Therefor the firebox and boiler dimension are overscale and this probably leads to a nearly 2 mm overwidth cabin. All due to a much too large motor. Further the model has a Knorr preheater which was not characteristic in Belgium, only about 5% of the machines had this equipment. Apart from that there is the German lighting, the wrong buffers, old type of safety valve, superfluous bell and light machine, no light below the boiler because of the worm drive being there.
And above all this:
Pro's are: the wheels are in the correct position, although a bit small but it concerns the right amount of them (4 axles) and it is a Belgian machine. Together with the type 96 the only ones you can buy as R-T-R.
The running properties with a replacement motorising set (Verbeck, no longer available) using a 1016 motor with flywheel are reasonable to good. This depends on the friction in the gearing and this seems to differ from machine to machine. I know because I have several.
I wanted something better than the Arnold model thus decided to tackle this machine using some of the Arnold parts. I ended up using only the cylinders, the crossheads and the cabin less the overwidth.
tender and frame
Below we see the new drivework. The tender is based on 2mm practise, splitaxle technique. The 2 back axles form a splitframe bogie which has 1 springloaded central vertical axle. Pickup is by PB springs to the outerframe. The front axle has 0.5 mm extended steel axles which run in the outer frame plates. Thus we have a 3 point suspension and all wheels on track. The result of this is that the diameter of the flywheel can be maximised because it sits partly between the frame. A large flywheel means mass as working power. Also the weight of the tender is maximised ensuring more reliable pickup.
The front part shows a set of fiNe-scale wheels. These are newly turned steel tyres on Fleischmann half products. A different production method for these is shown elsewhere. Not all of them are entirely wobble free and I probably will replace 1 set somewhere in future. The gearbox uses a module .2 worm and 40 teeth wormwheel in the first stage. Total gearing is ~ 64:1. A compromise is driving the third axle and not the last. This is to obtain better running properties. There are 3 axles with gears, frontaxle is driven by the rods. The compromise is that, on 1 side only, a part of the gearbox is visible in front of the firebox. However I cannot tolerate a compromise in the running properties, trying to drive a geartrain via the last axle will lead to a highly sensitive loco due to play in the gears.
The wheels have the correct diameter, pickup is by PB springs on a pcb keeper plate which also houses the brake detail. Thus brakes cannot make shorts and are live. The frame is massive, built up with 3 layers of 2 mm brass screwed together after profiling the space for the gears in the middle.
On this photo we also see the testruns with dcc decoder. This is a DCX70, which is very good but too large. Its smaller brother DCX73 is now mounted but for optimal performance you need software version 14 or newer. Currently it is the best small footprint decoder for coreless motors on the market.
Building the loco body
This photo shows the new body. A question of simply starting out with some brass and soldering irons and files. And yes some parts were turned and others milled. The frontdoor is soldered using a resistance soldering apparatus. The cabin is glued to the firebox and spaces were later filled with Milliput. The boiler is 1 piece massive brass except for the gearbox space. As this space was oversized I filled it again with whitemetal (temperature had to be lower than solder meltpoint)
This is the other side. The ladders were soldered from plain wire. The chimney now has the characteristic setup for controlling the draft when not in use. Note the ribbed buffers with central hole. The new drivework was still missing above, but here it shows in close-up.
Blown up, an entirely new and in working order Walschaerts gear. The valve rod is moving and it can be put in backward and forward motion. However I didn't account for this when I started out. Thus with a massive boiler already finished I finally decided to fix it in backward motion. The rods are plain steel and show a bit too realistic rust here. Oh and yes the protecting tubes on the cylinders are still too short.
Here we see it in early stage in neutral position. And this photo also shows the compromise with the partly visible gearbox on this side. (There were machines with the Knorr preheater on this side!)
Current state, ready for heavy duty testing.
Extending the fleet
It took its time but finally a second machine reached a state that it can be taken into duty. So now they form a pair, not entirely twins as they differ somewhat in tenders but they are built along the same lines with identical gearing so that they can run as a coupled pair.
The new one has a slightly smaller flywheel of 13mm instead of 14mm as this Arnold tender body isn't as overscale as the tender body from Hobbytrain. The next photo shows a view from above before the new load of coal was fitted. 
The main difference is that the decoder is now fitted in the tender although in a rather experimental way that still has to prove that it is durable. This is a dcx75 that fits between the wheels. Its follow-up dcx76 is even smaller. It also shows the electrical connection between loco and tender with a set of wires into a tube and the screwed pulling connection. There is a bit paint on a wheel from touching up the sides.
But as the NMBS had 576 of these machine another one cannot do any harm, therefore the last remaining still intact Arnold loco is doomed to get a rebuild too.
Loco frame, wheels and tender frame already exist. This photo allows a view on the basic running parts. In this case the motor block in the tender is screwed with a M1 to a frame spacer that is stuck through openings in the frame, this makes it into a much more rigid construction. The last two axles are mounted in beams instead of a swivel bogie, this should be electrically identical but extended axles do run with less friction compared to inner frames. The flywheel is mounted a tiny bit lower which allows a dimension of 13.5 mm. The wheels are identical to the first one, made of Fleischmann half products ordered as spare parts but with new steel rims. Steel takes more care and thus time but it looks better than the NS used for #2.
Wheels mounted, wormdrive screwed to the gearbox with an M1 screw. Cardan drive is an home made UJ with 2 mm brass balls and steel shaft. On the back 2 new type 2212A0 wagons still unpainted.
Next phase will be the drawbar, pickups, test running, cutting and fitting of rods and Walschaerts moving parts and more test running.
copyright: Henk Oversloot
date: 3 maart 2002
extended: 17 june 2012
