This Blog describes the ongoing development of a 16mm scale 45mm gauge garden railway situated in the North West of England, UK from 2004 to the present day.
In Part 1, I described how I added a simple recordable sound module to my double-diesel loco to give it the basic sound effect of a diesel engine in action. However, there was no way of adding a horn sound to this set-up so I decided to add another sound card to the circuitry to provide a horn sound effect.
A suitable horn sound was tracked down on the freesound.org website and downloaded. It was then edited in Audacity to tidy up the sound and lengthen the sound of the second horn blast slightly (using the 'Tempo" tool).
It was then exported as a WAV file and uploaded to the new soundcard.
The wiring inside the main diesel loco was modified - basically, adding a second set of positive and negative leads to power the new card. A speaker was fitted into the cab roof of the other diesel loco ....
.... and a new micro JST connector wired in to link the speaker to the output from the second sound card. This was wired-up in reverse to the first connector so it would be impossible to connect the two systems incorrectly.
To create space for the second sound card, the underside of the bonnet lid was hollowed-out. As it was 3D printed, this was a relatively easy job - the inside of the domed bonnet was already hollow but with the 3D default infill structure.
The cards were shrouded in insulation tape to avoid accidental short circuits.
The second sound card was wired up to one of the channel 5 output pads (Pad 6) from the Deltang Rx61b so the horn would sound when the bind button was pressed on the transmitter.
The loco was then given a test run to check all was well.
Using these sound modules is a relatively straightforward process but as only one track can be played on the card it is not very versatile. Being able to select different tracks stored on the card would enable a series of sounds to be played in sequence (eg engine start, idling, accelerating, running, decelerating, stopping, etc.). However, this requires a more sophisticated control system such as an MP3 player or a micro-processors such as Picaxe or Arduino. For more information on using an Arduino-based sound system see - How I added sound to my Bluetooth controlled loco.
In the past I have used a greetings card sound module to provide low coast sound for my IP Engineering plate frame Simplex ( see How I used a greetings card module as a sound card). It was cheap (c£1.50 GBP) but quite limited in scope. The sound file was restricted to around 20 seconds, it had to be recorded via a microphone so the quality was poor and the volume wasn't great.
When I saw another recordable module on eBay for a reasonable price (£12.00 GBP), I thought if give it a try, particularly as it was also available on AliExpress for less than £3.00 GBP. They are also available on the UK Amazon website for £13.99 (including a battery box and a suitable USB lead).
There was no documentation but one of the photos accompanying the listing indicated the module could be programmed to operate in different modes.
Further research on t'internet revealed the modes were fairly limited but it was possible to loop the sound file, which is what I needed for a basic soundcard.
This chart may at first look baffling but, with help from a YouTube video, I managed to interpret it - see below for more info.
I had already tracked down a suitable sound file of a noisy diesel tractor on the Soundsnap website (see How I added sound to my Bluetooth controlled diesel loco), and so my first job was to edit a suitable clip from the file to upload to the sound module.
Since I downloaded this file, there has been a price increase on SoundSnap so you might want to seek out alternative tractor sounds from one of the free online sound libraries - eg https://freesound.org/people/aUREa/sounds/268779/
I used the free audio editing program Audacity to create a clip of the running engine lasting about two and a half minutes ( clips of up to four minutes can be uploaded dependent on sound quality).
The file was exported as a WAV. Unlike MP3 files, WAV files don't suffer from that annoying half-second break in sound when the file loops back to the beginning.
The module was connected to the computer with a micro USB connector, bearing in mind that some USB leads are not fully wired. The first lead I tried must have been wired only for charging a device as it failed to show the module's memory store.
With the right lead, the module's memory opens like any other USB drive.
The MP3 file which comes with the module was deleted....
.... and my edited tractor WAV file was dragged and dropped on to the module's memory (files up to 8Mb can be uploaded).
At first, I tested the sound was OK by connecting the two power leads to a 4.5v battery box.
I then got to grips with changing the playback mode using the above chart. It was a lot simpler than I expected. The zeros and ones indicate which of the three resistors labelled A B and C needed to be retained or removed to change the mode.
For example, 0 0 1 means keep resistors A and B and remove resistor C. This will make the module play the sound file when the button is pressed and continue until the the button is released (or the sound file ends).
The mode I was interested in was the penultimate one - press the button to play and then cycle or loop the file being played until the button is pressed again. This was coded 1 0 1 which meant I needed to remove resistors A and C but retain resistor B.
To remove the resistors, I applied heat from a soldering iron to one end of the resistor and gave a small amount of pressure on the iron towards the opposite end of the resistor. As soon as the solder melted, the resistor slid off.
The process was repeated for resistor A.
A quick test play showed the operation was successful.
I now had to fit the unit into the loco. Fortunately, the loco in question is powered by two li-ion cells wired in parallel to give 3.7v output to the Deltang Rx60b receiver. In addition to the 5v input terminal, the sound module has 3.7v input pads.....
... so the wires on the 5v input pads were removed and two wires were soldered to the 3.7v terminals.
These were then connected to the 3,7v input to the Deltang Rx60b so the sound module would be powered up when the loco was switched on.
I removed the leads to the push-button. At first I simply soldered across the terminals to the push button so the sound module would start playing immediately the loco was turned on.
However, as the sound file I used was of the engine running at speed rather than ticking over, I decided I wanted some way of turning the sound on and off from the transmitter.
The designers of this module have fortunately anticipated this requirement as the push button connects the trigger for turning on the sounds to ground (0v). Fortunately, they also include a trigger which is activated when it is connect to V+. The Rx60b is a very early Deltang receiver, with outputs which give only 3.2v rather than the 0v of later models and so, via a 1k resistor, I could connect the output pad from the Rx60b to the V+ trigger pad (labelled PH - P High) on the sound module - the other pad is PL (P Low)
The small speaker which is supplied with the module was housed in an enclosure which I designed in Tinkercad and 3D printed.
This was mounted, as is my usual practice, on the underside of the cab roof using double sided sticky pads.
The loco was then given a few test runs
This soundcard solution is not the most sophisticated - it plays only one sound which is independent of the speed at which the loco is running. There is also no horn or whistle. It is akin to the sound cards which used to be marketed by Acme (no longer in production). However, as the modules can be bought for under £3.00GBP which includes the speaker, it is a cheap and reasonably effective way of providing small internal combustion-engined models with a sound system.
Having designed, drawn and printed most of the Southwold Railway goods stock (eg see How I designed and constructed Southwold long wheelbase van, short wheelbase van, two plank open, and four plank open wagons), it was inevitable that I would attempt to design, draw, print and assemble at least one of the Southwold Cleminson six wheeled coaches.
Background information
The Southwold was one of the few narrow gauge railways in the UK (apart from the NWNG Railway and the Manx Northern) to use six wheeled coaches using the Cleminson system.
6-wheeled coaches at Southwold Source: https://lightmoor.co.uk/BDLbooksample_pics/L8429_samp2.jpg
The design, whilst ingenious, never really caught-on. Long term analysis showed that wear on track and wheel flanges was quite excessive and so four wheeled bogies became the norm - though in the UK some standard gauge six wheeled wagons (eg milk tankers) and vans (eg parcels vans) persisted well into the 1970s and 1980s.
The principle of the design is quite simple. The axles on the end of the wagon or coach swivel around a central pivot - while the central axle slides from side to side. The trucks on which the axles are mounted are connected to each other with a simple linkage.
As the wagon or coach rounds a curve, the centre truck slides sideways and the linkages force the outer trucks to pivot.
In reality, the centre truck does not normally slide outside the sides of the wagon or coach frame - the above diagram is to emphasise the effect. However, as we shall see below, this projection of the centre truck beyond the side frames did need to be incorporated into my design.
Developing a working chassis
Creating a working chassis was achieved very much by trial and improvement. My aim was to try and replicate the original Cleminson mechanism as faithfully as possible. However, the original design must have been developed to cope with broader radius curves and far less undulating track than that which is found on the Peckforton Light Railway.
Sharp curves
To deal with my tight radius curves, I needed to allow the centre truck to swing out far more than the design of the prototype would have accommodated, as indicated above.
I experimented with various designs of centre truck but eventually I had to compromise between prototypical appearance and performance. The centre truck is wider than the the coach body and so projects slightly beyond it when travelling along straight track.
This is to allow the truck to slide further than it would on the real thing when taking the tighter curves on my railway.
Undulations
Once I had cracked the design of the centre truck to tackle tight curves, I quickly realised that I also needed to compensate for the ups and downs of my trackwork. Because the wheelbase of the coach from one end truck to the other is quite large (364mm), there can sometimes be quite a large variation in the level of the track beneath the centre truck between them.
For example, when a coach is breasting the summit of a slope or travelling over a hump, the centre truck can act like the pivot of a see-saw, causing the wheels on the end truck to leave the track.
By contract, when a coach is traversing a dip in the track or when starting to ascend a climb, the wheels of the centre truck can leave the rails as the truck is suspended between the two outer trucks.
Clearly, the centre truck must be allowed to rise and fall to cope with these two scenarios.
Not only must the truck be free to slide sideways and move up and down, it needs also to be kept captive in some way, otherwise whenever the coach is lifted from the track, the centre truck would fall off. This was achieved with four hooked brackets, which contain the centre truck vertically by also restrain it when sliding from side to side.
In addition, the pivots for the linkages between the trucks also need to allow the centre truck to rise and fall whilst also allowing the linkages to operate unhindered. This was achieved with 15mm long pivot pins on the linkages.
Compensation
Another issue which needed to be addressed was to allow the wheels on the outer trucks to rock from side to side to cope with sections of track where one rail might be lower the the other. Normally, this isn't a major difficulty, provided the wheelbase of the vehicle is short. However, the Southwold coach is, in effect, a very long four-wheeled wagon which just happens to have an extra pair of wheels in the middle. Without compensation (ie suspension), there is no way the coach would remain on the track - the slightest discrepancy and one wheel would lift clear of the track.
Fortunately, I had already solved a similar problem with my Ford style railbus.
This too had a long wheelbase - though not as long as the SR coach. However, the principle could still be applied - keep one axle rigid and allow the other to rock from side to side (see How I added suspension to my Ford railmotor).
To achieve this on the SR coach, one of the pivot mounts for the end trucks was a simple ring bearing on the flat surface of the truck .......
....whilst the other truck had a longitudinal half-round bead along its length - thus allowing it to rock from side to side on the flat pivot mount.
Summary
Although the above may seem self-evident, they are the result of several months of designing, printing, prototyping, experimentation, re-design and re-testing to hone the way the coach performed. Even so, I discovered a couple of places on my railway where there was a hump and a dip which were too severe for the coach and so had to iron these out. I am pleased to say that I now have a rake of three Southwold Cleminson coaches and they perform well with no more derailments than any of my other stock.
Construction
Once I had finalised the design, construction was relatively straightforward, though the overall length of the coach means it has to be split in two (or three) to fit on the print bed (even though mine is 300 x 300mm).
As with most builds, I started with the underframe or floor - in this case the frame is separate from the floor. The two halves of the frame are identical .......
... and were glued together at the centre - with lugs and slots to ensure they matched up.
The central section forms the mount for the central sliding truck - the slots are for the retaining brackets.
... it was flipped over and the three floor sections were glued in place. The reason the floor was made in three sections rather than two was to provide some additional reinforcement for the joint between the two halves of the frame. The floor sections sit on a rebate on the upper edge of the frame. The two holes are to gain access to the bolts used as pivots for the end trucks. They will later be blocked off with removable plugs.
The sides of the coach were then glued into place. These are split into two halves. The two halves are slightly different - one has an opening toplight (see below) .......
..... while the other half doesn't. There is only one opening toplight on each side. The side walls fit into another rebate on the top edge of the frame to provide a slight overlap at the base of the side wall.
Once both sides had been glued into place, the ends were then glued in between them. It was fairly obvious where they needed to go - as you can see the ends are notched at the top to allow the curve of the roof to meet the upper outside edges of the side.
The support ring for one of the end trucks was glued into place next. A short piece of rod was put into the holes in the frame and the ring to ensure they aligned while the glue was setting.
The centre truck was assembled next. The W-irons/axle boxes are slightly different to those for the end trucks and are labelled 'MID' on the inside edge.
They were glued to the centre truck frame - trapping the wheelsets between them. I use Bachmann 24.5mm diameter metal wheelsets. I find they are fairly reliable and also provide a bit of additional weight to help keep the wheels on the track.
Before the end trucks were assembled, a 20mm long M3 bolt needed to be positioned in the pivot hole (it's almost impossible to insert it once the wheelsets are in place) .....
..... the W-irons were then glued on trapping the wheelsets between them. One truck was mounted on the bearing ring - the bolt being passed through and a couple of nuts threaded on to to act as a lock-nut.
Before the other truck was attached, two lengths of half-round bead were glued along the centre line, to provide rudimentary suspension/compensation.
The second truck was then bolted on to its mount. However, the nuts were not over-tightened to allow the truck to rock freely from side to side.
With the two end-trucks in place, attention was turned to the centre truck.
Two brackets were glued into place in the slots on the central truck mount.
The centre truck was then put in place ........
..... before the other two brackets were glued.
An LGB self-tapping screw was passed through one of the holes in the linkage, and screwed into one of the arms from the centre truck.
A long pivot pin was then passed through the other hole and glued into the hole on the arm from one of the end trucks.
This process was repeated for the other arm on the centre truck.
The roof was made from four sections - two outers and two inners. The outers have a curved batten along the outer edge.
These were glued together and smoothed off with a file.
The internal ribs were then added - not only to provide some additional strength but also to help locate the roof correctly on the walls of the coach. I decided not to glue the roof in place as I like to be able to gain access to the interior - to add passengers and also to be able to reach the nuts holding the outer trucks in place.
The coach was then test tun through some of my pointwork.
I added some strips of lead flashing to the centre truck to help keep it on the rails.
The steps were now glued to the ends of the frame beneath the balconies. I believe the SR coaches had steps only on one side as all their platforms faced the same way. On the PLR, with island platforms, I need steps on both sides.
Before the benches were added to the interior, the coach paid a visit to the paintshop. The trucks were removed to be painted separately. The wheels were masked with tape and the body and trucks given a couple of coats of Halfords grey primer from an aerosol rattle can.
The interior was masked and the exterior given a couple of coats of Halfords Vauxhall Burgundy Red from a rattle can. The railway's crests were printed on to self adhesive vinyl sheet, cut out and attached to the middle of each side.
The trucks and ironwork were painted black and the floors of the balconies and steps painted brown.
The interior was then hand-painted with brown acrylics,
The maker's plates were then painted and added to the sides of the frames - they also help to mask the joint between the two halves of the frame.
Glazing was then added to the sides and ends using 1mm clear acrylic sheet. I fixed this in place with clear Bostik adhesive - superglue has a nasty habit of making clear plastics go hazy.
The bench seats were assembled, ......
..... painted brown and glued in place inside the coach
Passengers were then added.
Balcony rails were made from 75mm long pieces of 3mm wide brass strip with a 1.5mm hole drilled near the end.
A point 40mm from one end was marked and the longer end filed down to a point and the other end rounded off with a file.
A 40mm length of 1.5mm diameter brass rod was soldered into the hole and the strip was folded in half.
A small hook fashioned from some 1mm brass wire and soldered to the end.
The assemblies were stained black and glued to the ends of the balconies and blackened brass chain glued between them.
Grab rails were made from 38mm lengths of 1.5mm diameter brass rod. 5mm at each end were folded through 90 degrees and then the handrails were slotted into 1.5mm diameter holes drilled beside the balconies.
The trucks were refitted and my own versions of hook and look couplings added - with integral Binnie medium Carmarthen buffers which can swivel with the couplings.
The coach was then test run.
I am pleased to say that it ran almost perfectly. There were a couple of places where the centre truck had a tendency to derail - but closer inspection revealed that one was a place where there was an unwanted hump and the other had a rather severe dip. Rather than modifying the design of the chassis yet again, I re-laid these sections of track to eliminate the problem - see Progress Report 90.
I have now completed a rake of three of these coaches and do like their appearance, especially when pulled by my model of a Southwold Sharp Stewart loco (see How I constructed a Southwold Sharpie).
You can see all three coaches running in this video charting progress on the railway in 2022 (3m45s into the video).
As I worked on each of the three coaches, I improved the components slightly each time and so, the final version is as good as I'm probably going to get it. I have provided the parts as a download in the 3D files section on the gardenrails.org website. However, bear in mind that my model is constructed to 15mm/1ft scale and I designed to parts to be printed on my printer with a 300mm x 300mm print bed. I have not (as yet) produced parts for 32mm gauge (as the Southwold was three foot gauge). I might consider it if there is sufficient demand. The download is free of charge though you will need to create an account on the forum to access the files.
