Ian Allen and Peter Reynolds bring us up to date with progress on the control system for the SR 7mm Group’s layout

Control panels

After the short over-view of the control panels in the last article, it is time to have a look at them in greater detail. As mentioned earlier, we have looked at improving the overall operational control of the panels by reducing the amount of controls on them. This is beneficial in two distinct ways:

1. The panel is less cluttered and thereby simpler to operate.

2. More space available for LED indication giving the operators more information.

The plethora of LEDs we have installed in the new panels may seem like overkill, but every LED has an important job to do and they can be broken down into three distinct types. The larger 5mm diameter type, are used to indicate which power controller is routed to each line, or section. Therefore, unless something is drastically awry, there should only ever be one colour lit, per section/line at any time. Controller outputs are fed by a form of route selection, so as a route is set by the signalman the LEDs will change to show which controller is to be used. At certain locations these LEDs will flash. This is to indicate to the operator that although a route is set and a controller is also routed to a particular section, the signal controlling the section is not pulled off as this switches the traction power to the section.

The second set of LEDs are the smaller 3mm type, and these indicate the lie of points, in the same way that a panel on the 12in/ft prototype would. White LEDs indicate normal route, with the red LEDs indicating the reversed position. Red/white bi-colour LEDs have been used for some of these which helps reduce the total number required. It was deemed important to utilise this method of indication so that the operator can see exactly which routes have been set, as well as being a confirmation that the signalman has set the correct route too. We haven’t though gone so far as to show signal positions. That will be up to the operator to keep an eye out for.

Moving on to the guts of the panels, as can be seen in the accompanying photographs, there isn’t a lot to them, apart from a large amount of wiring and the microprocessor controllers. As all the power supplies and associated electrics have been kept separate it means they are lighter than previous types we have used.

The panel tops themselves have been manufactured for us by Parc Signs in Cornwall, https://www.parcsigns.co.uk/ after speaking to them regarding similar work they completed for the Mid-Hants Railway installed at Alton signal box. The design was carried out on Microsoft Publisher, as this allowed easy modification to a custom page size, specific for each particular location.

Once complete, the files were converted to PDF format and e-mailed to Parc Signs. Paul Bryant at Parc Signs was very helpful in discussing options for material and finish with us. The tops are made from 3mm acrylic with the layout printed onto the rear which allows for a clean look to the top face. LEDs are rear mounted on a supporting structure which means no drilling is required. Another advantage with this design of panel top is that rear covering layer can be cut, which allows us to adjust the position of the OLED displays.

The NX panel

As mentioned in the previous article, this has been built from redundant parts secured from going to the scrap heap, as Network Rail closes more and more signal boxes in advance of the new rail operating centres being built. It is a Western Region version with rotary entry switches, and is made of parts from Reading and Slough panels. Routes and track occupancy are indicated by 24v bulbs, 118 of them. It works on the eNtry-eXit principle, (hence NX). To select a route, the signalman operates a switch at the entry to a route and then selects the exit by depressing a button. The setting of a route also clears the signal applicable to the route, provided all points are detected correctly, and all track circuits are unoccupied. The system requires track circuitry to detect when a train both enters and leaves the area controlled by the NX Panel, which allows a route to self-normalise once a train has left the section. This is something the signalman will have to remember as switches will have to be cycled to select a route again. As with the other control panels, the NX Panel has a visual confirmation of the route set by illuminated displays.

Relays

On the original part of the railway which has been replaced with this new build, twelve 4-Pole Change Over relays were used to allocate controllers to varying sections/lines, with the restriction that a lever frame could not be returned to normal, until a train had reached the next operational section, thereby curtailing other movements until a line was clear. Our intention with the re-build was to use Picaxe microprocessors to do the same job, however, it was found we could achieve the same using 4PCO relays again, which would cut down on the complexity of programming and keep it all fairly simple from the point of view of fault finding. The erroneous task though was working out the wiring.

The first priority was to figure out which controllers would need to be routed to particular sections, and then working backwards from the outputs of the relays to the input side. Complexity was increased with some relay inputs being the outputs from other relays, all due to route setting requirements, with bi-directional lines providing an added headache, as the system has been derived around Up and Down line controllers. In these instances, the use of signals to switch relays provides a neat solution. Although the wiring diagrams may look like knitting, if each individual relay output is taken separately it is simple to follow through. The diagram is fundamentally a flowchart, expanded to cover four layers. The other advantage of this re-design is that now we can clear a lever frame behind a train, allowing for more flexible operation, and allowing, in theory, an increase in train movements. The first few diagrams only showed how the relays switched the live feeds, and not the returns, mainly for clarity, but for the new build at Bexhill they show both due to the complexity of the pointwork and how we were switching polarity on crossing vees etc.

The relays themselves have been housed in 19 inch rack mount cases with one case per location, making three in total, with up to 25 relays in each case. In total that’s 64 relays just for controller switching, excluding those for the Bexhill goods yard and platform crossovers, and we’ve also used the same relays for switching crossing vee polarity, although the complex station throat of the final station has been configured differently to the through station.

In addition to the relays, AND/E-OR (Exclusive OR) gates have been used in a couple of locations. An E-OR gate is switched by more than one signal at a particular location, as each signal needs to control the switching of a relay to energise the section it protects. One is operated by one of three signals, whilst the other by one of four. Once a signal is pulled to clear the E-OR gate allows the section of track in front of that signal to be energised. The AND gate relies upon two actions occurring to switch a relay. In this instance it is the correct switching of a point and a signal to allow a section to receive traction power. This forms the basis of flanking protection so that two trains cannot collide where two lines converge.

Destination describers

PICAXE An explanation provided by Peter Reynolds

In order to ease the complexity of the various elements of electrical control for the SR 7mm Group layout, use has been made of the various PIC microcontrollers to allow miniaturisation, specifically the PICAXE variant. As the PICAXE has essentially been developed for educational use, there is a large number of additional modules available that are specifically designed to interface with the PICAXE microcontroller itself. These range from servo and infrared modules to LCD and OLED displays.

Another great benefit is that ready-made printed circuit boards are readily available, and all items are generally relatively cheap. The programming software is also free, but an inexpensive programming lead is required to upload the programme from a PC to the microcontroller. Starter sets are also available to purchase. Programming software also includes an emulator, so you may try your programme prior to transfer to the PICAXE to see how it works, and is quite easy to pick up for ‘those of a certain age’ as it is very like the BASIC of the early days of home computing.

The actual PICAXE microcontrollers are essentially an 8- bit microcomputer on a single integrated circuit, with directly usable input and output lines. The number of available I/O lines varies between 6 and 33 depending on the actual PICAXE, but these can be used (with certain documented restrictions) with all sorts of devices; from switches to stepper motors to LEDs to memory wire.

They are very versatile; currently we have used them to interface lever frames to servos to operate semaphore signals, control the traction power to sections on a scissors crossover, drive the Plan Y control panels LEDs, and, eventually, control the OLED displays and the genuine NX panel we have.

With the semaphore signal use, a unit provides interlocking so the signals will only operate in a prototypical order. This sends a command down a single wire to a servo driver that can operate up to three servos/signals (in the application we have used them - although it is possible to control many more than three). It sounds a lot more complicated than it actually is. For anyone interested in further reading, have a look at the very informative website www.picaxe.com, where there are links to documentation, programming tools, forums, and hardware distributors.

Effective Communication

Since the early days of the SR7mm Group, effective communication between the many operating positions has been a constantly evolving issue. The logical way of carrying this out was to give the trains a four character head code, as on the real railway, and I set about designing an electronic means of communication between locations. This had to necessarily be portable, compact, and relatively simple to use. Eventually, after a considerable period of trial and error, I did come up with a pair of units that did fulfil the above requirements, but they were limited in operation and couldn’t be easily amended. This was finally completed just as the railway was about to be demolished outside and partially rebuilt inside, for ‘Plan Y’.

Once ‘Plan Y’ was in its genesis, we had a number of discussions about how we could include a visual means of communication between operating locations and it was decided that more modern means of communication would be required. Block instruments were ruled out, mainly due to size and noise issues, although I did design and build a pair of electronic block bells. The new control panels were designed to accommodate OLCD display panels as train describers. Ian designed a control panel to drive these displays, and I set about designing the electronics to get everything working. It quickly became apparent to me that, although it was perfectly feasible to carry this through, the wiring between locations would be complex. The next option would be to use wireless.

I had now reached a stage where I was looking around for a means of simplifying everything; I wanted to have something that could be turned out relatively quickly, and meet the requirements, and an Android tablet computer seemed ideal so I proposed that we could manage all the traffic flow using tablets. Once I started to look into this in greater detail, there didn’t seem to be a negative whatsoever; a decent unit could be bought for the same price as the displays, it was compact, it had wireless, clocks, Bluetooth, internet… Well, you get the picture. ……..or here.

Android tablet version

When I started designing the tablet display, it was necessary to get a general consensus about layout, functions, and means of communication. This was agreed, and I started work using Eclipse, which was the usual means of Android programming at the time. I bought books, a number of books, a quite large number of books, if fact, to guide me. Then after this work had to necessarily be put aside to get the railway working, I returned to the programming to discover that Eclipse had been superseded by another approved programming environment, so it was necessary to start again. Things were also delayed as the new edition of a book I needed was only published towards the end of last year.

So this is the stage we are at now. Luckily, the new Android programming environment is very ‘visual’. You can instantly see what your design will look like on the screen, and you can even simulate its operation without using an actual tablet. This suits me extremely well as I don’t need to carry too much equipment around with me. Although, I need the books for guidance, the programming environment helps enormously; it is intuitive and it indicates errors as you go along. Program modules can be tested as they are written to see if they do what I want, and the watchword is simplicity from the users’ perspective. Ongoing development of this will continue for some time to come, until it will do everything we need it to. The usual limiting factor, of course, is time, as this all is necessarily bespoke. I can’t see anything commercial being available to fulfil our requirements.

Has it been worth it? That is quite a wide question and I am not really in the position to answer it either. So far, from a purely personal point of view, it has opened up new possibilities for me, and provided a good means of exercising the ‘little grey cells’. However, the ‘proof of the pudding is in the eating’, as it is said, and that will be easier to gauge success once I have a pair of units finished and they can communicate. After all, this work and project is to enhance the experience of those operating the railway from the operators’ point of view. The use of modern materials is not a bad thing; is it…?