Sunday, February 8, 2015

Wollar Station Yard Upgrade

Several months ago as part of signalling my layout I decided that I would finally do something about the non-prototypical track work at the UP end of Wollar. Way back in the mists of time (mid-1980's) I was given some Shinohara points by a friend. As money was tight I built a station yard with what I was given which led to the not quite right track work.

I recently ripped out the offending points and replaced them with some Peco code 75 points in a correct arrangement. Another reason that the points had to go was that they weren't DCC friendly at all and also caused the occasional derailment.

I won't bother you with what the track work looked like but here is the upgraded yard layout. The main line is the loop track through the station platform and the actual refuge loop is the straight track, a bit different but it was used occasionally by the NSWGR. The back platform road is the one closest to the aisle which is at the bottom of the diagram.

This diagram is the one I have mounted at the Wollar 'Signal Box' otherwise known as the Wollar control panel.

About the same time that I was rebuilding the track work I visited Train Works at Thirlmere with Chris and my youngest grandson Jesse and I took photos of the plates on the lever frame there.

As a result of this outing I decided to do some work to upgrade the control panel to at least partially replicate a signal box arrangement.

This led me to draw up small versions of the plates which I had 3D printed. I have just completed painting and installing these above the key switches of the points and signals (see below). The plates are 25mm x 10mm and it was interesting applying the brass colour to the numbers, letters and edges of the plate that were raised only 0.4mm from the plate surface. I used a gold coloured Uni-ball Signo fine gel ink roller ball pen that I found at Officeworks. If I ever re-draw them I will look at raising them higher from the surface. I had to do a bit of touching up with black paint.

The numbers on the plates are replicated on the track diagram in the photo above and will assist the Signalman at Wollar the next time that we operate the layout. When I have time I will print up a 'Pulling List' that will tell the Signalman which levers (switches) to pull (throw) to move through the various routes in the station yard. There is an area for the list to the right of the main bank of switches. The group of four switches at each end are the UP and DOWN Accepts and the three UP and DOWN Starters. The Accept switches are the two grey ones on the outer ends. The Starters are the red ones and the points are black. I actually have one switch, number 8 which is marked NOT IN USE as it doesn't control anything. This was a decision I took when I decided that I would have the bank of four switches at each end of the main bank so that it was balanced. It originally controlled one of the removed points. The two red switches on the ends of the main bank are for SHUNT AHEAD signals for each direction out of the rear platform road and are used to control leaving the sidings and moving onto the main to shunt other areas of the station yard.

The white thing hanging on the brass cup hook at the UP (left ) end of the 'Signal Box' is the staff for the Coxs Gap to Wollar section, a similar hook is at the right, the staff for this section being at the Gulgong staging yard at the moment. We use these staffs when operating. I also have 'Tickets' drawn up but we haven't used those as yet.

The Wollar sign on the layout fascia is also 3D printed along the same lines of the lever plates using the NSWGR 'font'. At the moment I have only done these letters of the font. I will eventually do signs for the other stations.

There are more signals than there are controlling switches as a number of the signals are linked to points and are controlled through JMRI and its inbuilt logic. A signal can't be thrown for a train to enter or leave unless all the points are aligned correctly. At the moment the logic is only in place at the DOWN end. The semi-automation is part of making the signalling easier and hopefully fun. The Signalman will control the entry of trains with the Home signals which are linked to the Accept switches. The Home signals are also linked into the JMRI panel logic an example of which is the Flour Mill siding at the DOWN end that has a manually operated point which if thrown to the siding will stop the Home from being cleared or change it to stop if it is at clear. The main, loop and back platform road will have occupancy detectors feeding back to JMRI as well. The Home signals are brackets which will also show which road is clear to enter.

It will probably be another month or two before I complete the rest of the signals and do the JMRI logic. The signal posts are in place and painted but no arms, servos, etc. So an operating night is a while away.

I just hope that the rebuild and 'Signal Box' go down well with the two Ramblers that often act in the role of Wollar Signalman, time will tell.

Of course the observant will notice that there isn't a signal box at Wollar, something else to build. I am working out where it will go at the moment and I am currently favouring one on the Down end of the platform.

Saturday, February 7, 2015

Installation of a Soundtraxx Tsunami ALCO 251 v12 sound decoder in an Auscision 45

Given that the Auscision 45 class comes with a 21 pin decoder socket for the ESU Loksound v4 decoder this presented me with a problem as I already have seven diesels fitted with Tsunami decoders.

Why was this a problem?

Here are some issues that I had to think about:
  • The Loksound v4 momentum factor issue - ESU has moved away from the standard NMRA momentum factor of 0.876 and introduced a factor of 0.25 in the Loksound v4 what this means is that for the same momentum input for acceleration or deceleration the Loksound v4 will only have about 30% the momentum of any other brand decoder. This becomes a real issue when consisting and using the momentum button on a NCE throttle to set the load behind the train. As you can imagine, locomotive consists with a mix of Loksound v4s and other decoders will have trouble as the locomotives will appear to fight each other. The Loksound v4 equipped locomotive(s) will slow down or accelerate faster than the other(s). The momentum can of course be adjusted in each decoder type to match but then that stops the use of the NCE throttle momentum button for more realistic operation.
  • There is a Loksound v4 decoder that is available with recorded 45 class sounds that has manual notching whereby a press on the appropriate function key will start the locomotive notching up and there is another key to notch down. While this can give an appearance of prototypical operation it also means that you can drive your locomotive in speed step one up a steep grade with a long train. It can be a bit like playing a musical instrument, you have to think about what you are doing with the diesel motor to match how you are driving. My seven Tsunami equipped diesels have automatic notching and a speed curve set up so that in notch eight the diesel is going at 15 miles per hour. I did this as the NSWGR rated their diesels at what speed could the maximum power be achieved (notch 8) on a 1 in 40 grade. The speeds ranged between 8 to 13 mph so after some experimentation I chose 15 mph as the average of approximately 10 mph was too slow. I also set the Tsunami decoders to notch up/down at 10 speed steps so in 128 step mode you can tell which notch the diesel is in by looking at the speed step (10, 20, 30, etc.). The ruling grade on my Bylong layout is 1 in 40. Incidentally the Loksound v4 decoder does have automatic notching but the notches are spread over the full speed range. I believe that this could be adjusted but the learning curve required is very steep and the available 45 class Loksound v4 is locked so this can't be changed, maybe one day.
  • The price of a Loksound v4, enough said.
Since I had an AT1000 Tsunami ALCO v12 already, the choice was made to install it.

I removed the existing board with its 21 pin socket and proceeded to wire up the AT1000.

Edit: I should mention that there is little vertical room for the AT1000 and I had to cut off the two short posts that located the original board at one end. I then glued a piece of 0.005" styrene across the two mounting posts of the original board at each end to insulate the AT1000 from the chassis.

Mostly, everything went well until I found that I couldn't get the headlight/white markers/number boards to light up. The red markers worked and I tested the outputs on the AT1000, the headlight outputs at both ends were good, so a problem with the 45 class. A call was made to Marcus Ammann who quickly sorted me out. It appears that the white wire to the headlight output at each end of the AT1000 is correct but the purple wire has to go to the function common along with the blue wire! The wiring of the Auscision 45 is back to front for DCC, something to do with DCC using a +ve function common instead of the more usual -ve common found elsewhere. Perhaps Auscision need to inform their factory about this anomaly and change it for future models.

Wiring Diagram

I wired the front red markers to the FX5 output on the AT1000 and the rear red markers to the FX 6 output.

So what I have is the headlight on F0 (Headlight button on NCE throttle) and the headlight/markers/number boards will turn on at the front or rear depending on the direction of travel, not perfect unfortunately.  The red markers are on/off as required.

This allows me to have the headlight/white markers and number boards on at the front when moving forwards and not have the red markers on at the rear of the locomotive when pulling a train.

Also, when the locomotive is standing in a siding or shunting it can have both front and rear red markers on at the same time as per NSWGR practice.

The headlights, white markers and number boards should all be on separate outputs but this is about the best we can do with the restricted number of outputs on decoders. Although I should add here that the QSI Titan has 10 outputs which almost gets us to the 12 required to handle marker lights etc. prototypically. Unfortunately at the moment they are hard to get.

I used a Soundtraxx 810113 Small Oval Speaker in the 45 class speaker enclosure but I had to cut the sides of the enclosure to fit the speaker (see photo above).

The sound from this speaker is quite good and here it is in the video below along with the same sound decoder and speaker set up in the second 45 class, an AR Kits version. Unfortunately the microphone in my video camera is not that great. I am sure that the Tsunami ALCO 251 v12 doesn't sound as good as the available 45 class Loksound v4 version but it does match my other ALCOs and it doesn't cost as much.

I actually have four Loksound v4s in steam locomotives, I just don't like them for diesels so I am not trying to stop anyone from using the Loksound v4, this is just my implementation for my reasons although I do think that modellers should consider the momentum issue against their existing diesel roster and layout operational ideas.

Approaches have been made to ESU by some well known model railroad identities (as well as Matt Herman of ESU USA) to have a CV added so that the standard NMRA momentum factor can be chosen if required but to no result. We can only hope that ESU will wake up, especially considering their current inroads into the US OEM market.

Wednesday, January 28, 2015

Lighting Signals with Surface Mount LEDs

Sorry, but here is another signal related post although it is also related to progress on Bylong.

The instructions on the Signals Branch blog for my 3D printed signals discuss using 0.8mm SMD LEDs to light the signal aspects so here are a few photos showing how I did a bracket signal.

Firstly it is best to buy the SMD LEDs pre-wired, they can be bought without wires but they are so small that soldering very fine wire is difficult.

The lamps on the signals have a slot in front and back. The rear slot is to allow the wires to come out and the front is a slot as the minimum wall thicknesses make it difficult to have a lens or hole.

The 0.8mm SMD LED will be a slight push fit into the hole in the top of the lamp and the square LED sticking out of the front of the tiny circuit board will fit into the front slot.

The 0.8mm SMD LEDs are to be gently pushed vertically into the body of the lamp. I found that I had to use my finger nail to push it in but a toothpick would also work. Don't use a metal object to push the LED home as damage may be done to the fine wire on the top end. The other wire is at the bottom of the LED. Ensure that the wires align with the slot at the rear of the signal as it is pushed into place.

The wires from the small outrigger dolly post of the bracket signal were threaded under the landing and then through the angled bracket supports under the rear of the landing and then down the main post.

The wires from the lamp on the main post were run down the back of the post around the end of the landing and then down the back of the post.

Glue the wires to the signal with Super Glue progressively stopping at each change of direction.

A small circuit board can be made from a printed circuit board (PCB) sleeper or with a piece of PCB cut to size (4mm x 15mm). The copper on the PCB is cut down the middle of its length with a fine stone in a hand motor tool (Dremel, etc.) and then one side of the copper is cut at the half way point. A 1K Ohm resistor is soldered across the half way gap, this is the +ve side of the circuit. The other copper strip is the -ve side of the circuit.

A small 2 pin plug and socket is made by cutting up an integrated circuit socket. solder the long pins of one half to each side of the circuit board as shown on the diagram and photo below. Solder the wire from the LED to the two connections at the other end of the circuit board. The longer wire of the LED goes to the +ve terminal and the shorter wire of course goes to the -ve terminal. I have used a 1K Ohm 1/2 Watt resistor so that the light shows through the spectacle plate of the arm during daylight but any value from 1K Ohm to 10K Ohms could be used, your choice. 10K Ohm will probably not show during 'daylight' (room lights on).

Glue the circuit board to the signal where shown in the photo below and tuck the excess wires of the LEDs out of the way up in between the two halves of the mechanism.

The wires on the male plug are soldered across the two holes rather than into the holes, this is needed from clearance from the main operating crank of the signal (see photo).

Signal Circuit Board Diagram
Please note that the plug and socket doesn't have to be on the circuit board or even used for that matter. This method is presented in case the modeller wants to disconnect the wiring from the signal.

Signal showing small homemade circuit board
Lit LEDs during 'daylight'
LEDs lit during 'night' - No signal arms.
Once it has been confirmed that the LEDs work then the wires can be painted over or even smoothed over with a filler. It has been found that Selleys Spakfilla can be applied and once dried the signal can be flexed without it falling off. Spakfilla is an interesting material, it has no real weight and must be pushed onto the surface then carefully smoothed with a metal spatula or similar tool.

One thing I found while doing this is that posts glowed internally from the LED so I had to paint the posts black from the top of the counterweight brackets to the cap on the post. Of course I then had to paint the post white again.

Glowing bracket posts

Lit signal arms during 'night'
The light showing on the posts in the photo above is reflecting back off the rear of the signal arm lenses, perhaps some matt clear might reduce this.

I found that the LED was not yellow enough to give a green glow through the Tamiya X23 Clear Blue so I give the blue lens a coat of Tamiya X25 Clear Green.

Now it looks like I will have to light the Brakes Landmark, what have I started....

Installing lights though is a long term project for Bylong as I would like to operate eventually at 'night'.

Wednesday, January 21, 2015

Painting the Clearance Post and Lamp

Well, I painted two posts again, the first with two coats of black and then white and the second with silver, black and then white.

It seems that either is sufficient to hide the glowing light from the LED so I think I will recommend two coats of flat black and then the final flat white coat. There did not seem to be any difference in light between the silver and the black painted posts.

Flat or matt paint should be used for coverage as the matt effect is caused by the pigment protruding through the surface of the paint and diverting the light rays in all directions. A gloss paint just has a nice smooth surface and as such does not have as much pigment.

While the paint was drying I went to Jaycar and bought a selection of 1/2 Watt resistors in 2K Ohms, 3K Ohms, 5.1K Ohms and 10K Ohms. I made up a 10K Ohm resistor by putting several in series.

The end result of swapping the resistors around gave the following effects.

1K Ohm resistor - Light flare from lens
1K Ohm resistor - Light shows in 'daylight'
10K Ohm resistor - No light flare from lens
10K Ohm resistor - Lamp appears to be off in 'daylight'

I will recommend that the modeller chooses which resistor to use depending on the preferred effect.

I prefer the 10K resistor as the lamp appears to be off although it is actually on. So back to Jaycar to buy some of the 10K Ohm resistors that I didn't get today, oh well, that's how it goes.

Tuesday, January 20, 2015

Lighting a Clearance Post Lamp

I have had a chance to try lighting a clearance post lamp as discussed in the last post.

I may need to amend the installation painting instructions on the Signals Branch blog as the light from the LED is much brighter than expected and as can be seen in the photo below the light is showing through the paint.

Here is the clearance post lamp lit by a warm white LED. As this was a temporary mounting I added loose ballast around the base for the photo.

This post was painted in black first then white.

Another post painted in silver and then white was actually worse so the silver has less covering power than black.

At the moment the instructions state to use silver then black and finally white but I don't think that the silver adds anything and isn't needed for light transmission either.

I am currently painting another clearance post and lamp with two coats of black and then a white top coat. I will report on this in a day or two as each coat needs to dry thoroughly before the next coat is applied.

The LED is powered from a bridge rectifier running off the DCC track bus. I used a PO4 6A 400V Bridge Rectifier (Jaycar Cat. No. ZR1314) to convert the DCC power to a nominal 12 volts DC. This bridge rectifier has the AC and the + and - terminals clearly marked on it. Also it has a hole through the centre which allows it to be screwed to the layout frame near where the lighting is needed.

I bent the legs out at right angles and screwed them into two terminal blocks at each side as shown in this photo.

The resistance I used for the LED was 2K Ohms which was made by using two 1K Ohm 1/4 Watt resistors in series. I will rewire the LED and resistors and use shrink wrap tubing once I have decided about the resistor value.

The light level could still be a bit lower which could also help to reduce light leakage through the paint. I will have to get a few 5K Ohm 1/4 Watt resistors and see how the light level looks.

I will post the results and amend the instructions accordingly.

Monday, January 19, 2015

NSWGR Clearance Posts with Lamps

I have just added some instructions for clearance posts with lamps to my Signals Branch blog.

Clearance posts with lamps were placed at fouling points of sidings and loops and were painted white for visibility. The lamps were lit at night usually by the Station Assistant and spares would have been kept along with kerosene in the lamp room attached to the men’s toilet.

I have made these in two material types, the White Strong and Flexible for those who may be worried about the posts breaking if hit while cleaning track with a track rubber and Frosted Detail and Frosted Ultra Detail for those who want more detail and who also may wish to have the lamp lit.

If printed in Frosted Detail or Frosted Ultra Detail material they may break if knocked hard (I don't know yet as I have only just installed one), if so then just pull them out and replace with a new one.

These clearance posts with lamps are to be mounted in a 3mm hole drilled through the baseboard at the clearance location of a set of points (turnouts).

The Frosted Detail and Frosted Ultra Detail are an acrylic material which can transmit the light from a 3mm golden white LED placed into their mounting hole from below the baseboard.

Clearance post and lamp in Frosted Ultra Detail - Not lit yet.
Just a quick photo, ballast not yet reinstated around post, being white the detail doesn't show well but the black dot is the hinge of the lamp top and the catch is on the other side.

I have tested it with a torch from below and it gives a small circle of light that would only be seen with the room lights off as you would expect, this is no glaring headlight.

Further details can be found on the Signals Branch blog.

Wednesday, January 7, 2015

Signal Assembly Instructions Update

I have just updated the assembly instructions on the Signals Branch blog.

I did this as a result of currently assembling 16 signals (single posts and brackets).

I amended the drill bit sizes for the fine holes in the White Strong and Flexible (WSF) material and the Fine Ultra Detail (FUD) material. The  drill bit sizes were amended because the WSF expands slightly when drilled so you end up with a smaller hole than the drill bit. As a result the next drill bit size larger than for the FUD is used.

I also added something that I forgot to mention which is that the signal lamps have been designed to take a 0.8mm golden white LED (the ones with 150mm/6 inch leads). Some thoughts are included on wiring.

I haven't installed any 0.8mm LEDs in the signals as I am using Tam Valley Singlet Servos which have Green and RED bi-colour LEDs on the small control board.  The control boards are all under the layout near the servos. I won't be using the push buttons on the control boards to change the signals they will be run from a layout panel on JMRI using some logic. I will probably extend one of the bi-colour LEDs from each control board out to the fascia opposite the signal to show the status.

One problem with semaphore signals on a model railway is that we view the signals from some unusual angles and can't always see the actual arm position. The bi-colour LED will show the arm position for the driver.

If I ever get into night time operations then I will probably install the 0.8mm LEDs in those signals that can be viewed from the front.

Lastly, there are instructions on making clear lenses for the signals and colouring them with Tamiya Clear Red and Clear Blue.

The prototype signals had blue lenses so that the yellow light from the kerosene lamp would show green. This is why a golden white LED should be used, probably at a low light level which will depend on the value of the resistor used.

A statement and an amended date is at the top of the instructions and if there are any further amendments the date will change and I will list the amendments.