Lighting the Faller Station

There's been some uncertainty about how to fix in place the Faller station, and also how to allow access to the lighting inside. The options were to either make the whole station removable from the layout, or avoid fixing the roofs (station + shed) so these can be removed in the future when needed. The latter was chosen, due to the disadvantages associated with the former, namely "breaking" the earth blend right next to the walls and the track ballast next to the corresponding track-side.

The lamp sockets and bulbs - Faller 180670 - are not included with the station (Faller 110096) as-is but I did purchase 3 of these a while ago. Since the sockets have each 2 holes, 1.5 mm in diameter, I've extended these to 2 mm, since I had plenty of 2mm bolts complete with nuts. The wires will have plenty of space, since the whole building will seat on beams, 2 cm in height, so the track terrain nearby is matched.

The Faller sockets come with pre-drilled holes
of 1.5 mm

Lamp sockets in place

Detail of the lamp socket in the station building,
slightly offset by mistake

Replacing a Viessmann Light Bulb

More than 3 years ago, a Viessmann 6384 railway lamp was bought. Besides aspects such as painting, positioning and wire placement, one thing of interest was how the light bulb is changed, in case it fails.

Removing the plastic hood is rather quick, although some force needs to be applied. Next came taking the light bulb out. This proved more complicated for me, so with help from Viessmann's support, it turns out the brown wire has to be gently pulled inside the mast. Using some sharp tweezers is best, in order to squeeze the tip between the mast's lattice, and pull the wire bit by bit towards the top. After the wire has some some space inside the mast, the light bulb can be further extracted. In first stage, this needs to be done with care until the wire that's not insulated is released from the tube. Then the whole brown wire needs to be extracted via this end (the 6229 replacement bulb comes complete with the wires - one long insulated and one short not insulated - just like the removed one). The red wire is actually soldered to the mast, and the electrical connectivity goes just up to the point where the bulb is inserted. The last picture shows the electrical connectivity between these points.

The lamp before disassembly

Removing the hood

Light bulb extracted so that the
live wire is released

View of the wire inside the mast

Testing that electrical connectivity
is in place between the edge of the
mast and the red wire

Problems Assembling Faller's 110096's Shed Roof

A while ago I've ran into issues assembling the roof of the shed for the Faller 110096 Klingenberg station. I've emailed Faller support at the time, and they were very kind and provided by post - free of charge - a new paper mask and a few elements as well. However it was my fault that I didn't explain my problem very well. Hopefully I will be able to do this below.

 At page 10 of the manual included with the station, the roof must be applied on top of the walls, after the M2 paper mask has been inserted in place. This doesn't work quite ok though. A bit of background: the roof has 2 shoulders underneath; these will "click" in place when applied to the walls of the building. Also, the adjoining wall of the station has 2 grooves, into which the edges of the roof will only slide into provided the shoulders click in place when applied to the shed's walls.The problems start when the shed's paper mask is put in place - once this is done there is no more room for the shoulders to be inserted so that they "click" in place between the shed's walls, since the paper mask takes up that space. Also the grooves in the station wall no longer match the edge of the shed's roof.

One way to solve this would be to cut the paper mask so that the shoulders reach enough depth, but this will mean the structure of the paper mask will be broken, thus losing strength. One workaround would be simply remove the black top of the paper mask, thus assuring the roof's shoulders are no longer blocked, however there is one problem by doing this - light will be installed inside the shed (Faller 180670); the shed's roof is plastic, but doesn't stop light completely, so will run the risk of the roof "glowing" in the night.

Roof shoulders

Shed roof sliding in the station wall's grooves

Shed's paper mask glued and inserted 

Shed roof's shoulders no longer able to "click"

Shed's roof too high due to shoulders hitting the paper mask

Shed's room no longer matching the groove in the station wall,
once the paper mask is applied

Electrical Modifications For Phidget 1012

  Sometimes last year, when I was about to order the remaining boards for input detection (for detectors & points status) I realized that the manufacturer of the board I was about to buy had a new product, capable of twice as many inputs. Namely - the new Phidget 1012 had 16 digital inputs as compared to the old Phidget 1018 board containing only 8. However there's a catch with the new board - it requires an external DC power supply.

  Even though I've previously believed that separate input boards are needed for optical detectors and for the switches' feedback, it turns out they can be handled by the same board. The key is using a ground level selected at the potential of 'brown' (of the Viessmann 5200 transformer). Going back to how the 1012 works - this will show an input as ON if the power supply has its circuit closed by the 1012's ground pin and that specific input (eg for input 0, this will be ON when the MRD1's relay is closed, namely when G and D1 are shorted on the figure below). Note that the DC power supply's polarity DOES matter. All in all, pretty straightforward for the optical detectors. Now for the feedback of switches, there's 'grey' and 'green' which get alternatively connected to the 'yellow' and 'pink' respectively, depending which way the switch is toggled. Since 'yellow' and 'pink' are actually outputs of the Phidget output board (1017), these are simply commanded by the input board's relays, which simply connect the "brown" of the Viesmann 5200 transformer to either 'yellow' or 'pink', respectively. So the key in getting the feedbacks to work with the 1012 as well is to have ground always connected to the same potential of "brown", which is made through a simple connection, as indicated in the picture below.

  Initially I had trouble understanding how this works, simply because both DC and AC notions apparently got together in the same figure. However the "brown" (BN) is a simple connection, there's no alternative current flowing through this diagram. It's simply that one can choose his arbitrary ground for the direct current.

  Special thanks go to John Parsons of Azatrax - the manufacturer of the optical detectors used on this layout - who was kind enough to point me in the right direction with the Phidget 1012.

More "Earth" Boards

In the end, 3 boards, each containing 8 squares, were built to get a sense of the resulting color during the tests described here. The last square of the 3rd board was rather dark, and - given blue pigment was added incrementally to each - I was confident each of the shades could be replicated. Then the day came when I actually wanted to produce a specific color, in order to patch some of the earth already fixed on the layout. I chose the most similar shade on the test boards, which meant a concentration of 4.2% blue ultramarine pigment in the standard base formula I've been using. However applying it in just one coat resulted in a color more similar to a 2.8% square.

Given things were totally different from what I was expecting, 2 more test boards were built. This time however, the increment percent of blue ultramarine pigment was increased from around 0.3% (in the first 3 boards) to an average of 0.7%. Still 2 coats were used. The sifting would be done until the new specs of formula barely get "wet". However this time the upper portion of each square was left soaked on purpose. It turns out that these soaked spots actually will dry sensibly lighter than the rest of sifted formula, which has an important contribution to the outcome.

  Thinking about what could have gone wrong in the first boards, it might have been some already prepared base formula, which could have been done without the electronic weighing scale. Also, given the final percentage of the last test patch went up to 10% ultramarine blue, I've also considered how this percentage is calculated. So far in the test boards done, formula was already mixed for the previously completed square. This had a known percentage of blue pigment, which has been gradually increased. In my calculations for the next square, the percent of added pigment will only be computed given the total weight of the existing formula. Adding the blue pigment's weight to the equation wasn't necessary since the difference in orders was quite important (30mg for 15g). However given that bigger and bigger percentages were being used - now reaching 10% - this had to be revisited. Simply think of having 10 grams of the formula already available; we need to get the final percentage to 10% blue pigment. The simplest way would be to consider adding blue pigment in an amount equivalent to 10% of the formula's weight. In our example - 1 gram blue pigment. However the resulting formula now weighs 11 grams. Hence the pigment's correct percentage is 9.09% - becoming quite far off from what we had in mind.

  So doing a little math, we'll first consider C the existing's formula weight, X the weight of blue pigment that needs to be added to get to the desired p percentage of blue ultramarine pigment. The equation thus becomes X / (C + X) = p. This becomes 1 / (C/X + 1) = 1/p. Which eventually becomes X = C / (1/p - 1).

  Turning to our example, the needed weight of blue pigment to be added to reach a percentage of 10% is in fact C/9, which amounts to 1.11g. Computing the resulting percentage yields 1.11g / (10 + 1.11g) which comes really close to 10 (9,999%).

Even More Electrics

Again there's been some work done exclusively on the wiring part. The original connecting board now has both its D-Sub connectors soldered to the corresponding wires (photo 1 showing work in progress). These wires consist mainly of individual signal LEDs and track / power / optical detector current. For the second connecting board, built a month back, the connector with the optical detector wires had about half of these soldered (photo 2 from when this was done). The short-term goal has become setting up 3 signals and the already connected optical detector in such a way that each signal "thinks" and reacts according to the various detectors located in front and after it.

This will most likely be a stepping stone towards the final computer-controlled code, yet should be simple enough to get some visible results fairly quick.

  Since the connector with the wires for optical detectors has been fixed to the connecting board, the control panel board had a corresponding connector installed, and wires placed for linking it to the Phidget 1018 interface. A second cable was assembled in order to link these 2 (photos 3 and 4).

  Finally the wiring plan for the layout itself has been entered in a Visio document (preview in photo 5). Holes cut in the foam underlying board are also present. The whole idea is to be able to successfully track a wire, since now there's information about its endpoints, its color and the tooltip entered (ShapeSheet->Comment). As for the control panel (preview in photo 6) - the wires that will be done in the future still have a black color, to differentiate from the ones already installed, that have been colored according to their real counter-parts.

  Tonight I've also connected the various power supplies and checked that everything works well. Last photo shows the temporary setup employed, complete with an USB hub, which was connected through 2 USB extender cables - just to be sure that the hub can be physically placed behind the layout - actually the spot where the control panel will be sitting, in a vertical position.

  What was unexpected was the sheer amount of time needed to complete this wiring part. Just tracking each and every cable on the layout itself and drawing the corresponding diagram took easily 30 work hours. Each 32-pin connector cable took more than 5 hours to build. What started as a very simple job ended taking much more that I would have ever expected - which makes me seriously think about DCC decoders for each signal, switch and other accessory in the future, given the wiring should mainly consist of 2 wires carrying the signals across. But for now - and this layout - things will stay the same. There are still a couple of  items to be purchased - namely some Phidget boards that will take care of additional optical detectors, switch feedback and switch operation; some D-Sub connectors; a small power supply for the optical detectors in order to stop using the current bulky one; and a few others.

More Electrics

Realized at one point that since there's no map of how each wire goes where, not to mention that some of the wires lost their purpose during the time since not all of them were marked - that it's gonna get complicated soon if this isn't rectified. So there have been quite a few hours over this months of mapping out what goes where. A Visio diagram was also employed, and each wire colored as it is on the layout, complete with tool-tip comments that show a wire's function when hovering over it.

Work was done to the electrical panel that will contain all the modules (Viessmann, Phidget, etc.) so I can finally move away from the wiring-in-a-box approach. 

Last but not least, since it turned out that quite a few wires will be required in the end, no less than 4 D-Sub connectors would be needed (37 pins/ connector, however because I'm using some 32-wire legacy cable, 5 of those pins will go unused). The required connectors, brackets and various other items were purchased;  so far a connecting cable was built, the original connecting board got modified so that it holds 2 connectors, and a new enclosure for a similar cassette plus the corresponding board were just finished today.