Turnout drivers
This appeared in Modelspoormagazine 67. More pictures and text can be found there ...
There are different way to set model turnouts: manualy or externally driven, hidden drivers, fast or slow, quiet or loud, with or without an external signal or switches, ... In a quest for the ideal turnout driver I finally came to a home-built device...
A turnout driver... what dou you mean?
There are many ways to "throw" a model turnout. There are manual and motorized units. Manually can mean a small box besides the turnout, but can also be a hidden switch rod mechanism. Manually throwing a turnout can surely have its advantages, but can be troublesome if you have a large layout with complicated station entrances. A visible box next to a turnout can be handy, but isn't very realistic.
Motorized turnout drivers are mostly electricaly driven, and can be driven by solenoids, memory wire or with a gear-box engine.
When is a turnout driver ideal?
Of course, this is a question everyone should ask for himself. In my opinion, this means the following list:
- Remote controlled
- H idden
- Electrical
- Suitable for different model scales
- Compact
- Solid construction
- Powerfull
|
- Low noise
- Slow
- Adjustable
- A correct feedback signal
- Extra switch contact(-s)
- Easy to connect, few wires
- Affordable
|
Is this list too demanding? Maybe it is, maybe not. Anyway, this is the list I started from while developing my own drive mechanism. Disappointing commercial drivers pushed me to home-building.
|
The schematic might seem complicated. The upper part holds the measure-and-controlling part. One IC and some diodes and resistors take care of polishing and logifying inputs and outputs (feedback to the railway controlling sysytem).
The lower part provides power to the engine by use of a so-called H-bridge.
|
This drawing shows both the print-layout and component positioning. As you can see, I tried achieve a compact design. Photographic etching is the best way to make this PCB. An experienced electronic craftsman shouldn't have any problem with the information so far...
|
After drilling the holes in the PCB, we can start mounting the parts, starting with the low-altitude components. In this design, these are the signal diodes (1N4148). Watch component polarity while placing them. Once mounted, the PCB is turned over, soldered, and the component wires can be cut off.
|
Slightly thicker - though not much different - are the resistors. The next step are the IC and some connectors. On mounting the IC, take care of placing it the right way.
|
Condensors and small transistors are the next step. Watch their polarity, too. The power transistors are a bit taller and placed after these. These don't need any cooling: some free air is sufficient. |
Our PCB needs a power supply. Because multiple turnout driver units might be placed next to each other, it is handy to have a connection method where different wires can be connected. Therefore, a common electric connector is dismantled and cut through. This
|
De mechanische opbouw
The chassis is built from various styrene parts, glued together. This task must be done only once, as we make ourselves a mold from this.
|
As we (most of the time) need more than one turnout driver, a silicon mould and resin castings are a good way to get things going. In this story, I made a transparent casting to get more illustrating see-through pictures. A traditional resin casting is not just better for the job: it is cheaper, too.
|
Some holes must be drilled in the casting to hold the microswitches. The second switch won't be needed often, but there is room for one.
|
The geared block should be adjusted to fit in the gutter of the chassis. Some drillholes should be made now: the first to hold a 1mm piano wire from the microswitches, a second to hold M3 bolts. You should cut the threads to adjust the bolts on both sides of this geared block.
|
This block of resin will move the microswitches. Therefore, a 1mm piece of steel wire is bent and cut to length. At the pivot point, near the microswitch, a 1m hole is drilled in the chassis.
|
At the other end of this steel wire, the geared block can be placed now to check its functionality. The microswitch should 'click' around the block's middle position. A little screw, mounted near the wire's pivot point should hold the wire in place while swiveling.
|
Now the PCB canbe mounted on the chassis. Some small screws in pre-drilled holes are enough for a solid construction. Glue isn't needed and shouldn't be used: we might need to dismantle the board in case of errors... |
The turnout engine's mechanical heart is recovered from a disco's rotating mirror ball engine. The numerous gears provide slow movement and sufficient power. The only setback is that this engine needs a 1,5V power supply, but that problem is solved electronicaly. We solder two new connecting wires, recycled from UTP-wire, to the engine.
A gap is sawed in the future bottom side of the engine to provide free access for the steel wire, previously mounted on the microswitch.
|
We connect the engine and give it a first try. On the PCB, some steps ago, we mounted a three-way connector. This part determines the default movement of the engine. A wire bridge is placed between the middle and the left hole. This position can be altered later (middle and right hole), but for now it should be placed this way.
|
The engine is wired to the circuit board using the connector, placed between the power transistors. When supplying power, the engine should rotate clockwise. If it doesn't, change the engine connecting wires.
Now the engine can be mounted on the chassis and secured with the cast clamps and two small vises.
|
We need four contact strips to measure the far ends of the engine movement. I recovered these from a relay from an old telephone switching central. From these relays, some pertinax-like material as ercovered, too.
|
Three pieces of this insulating material are placed at on side of a 40mm M3-spacer and held in place by an M3-bolt. Between the insulations, the contact strips are placed. After doing this on both sides of the spacer, we have a U-shaped bridge.
|
It's time to get the geared block moving, so we place a small gear on the engine axis. The 40mm spacer can now be glued in place against the engine's top side.
|
The strips are alligned with the positiong bolts, mounted in the moveable geared block. The hole through these strips allow us to adjust the depth of the bolts without any hastle.
Once the strips are alligned well, they can be secured in place with some glue. After all, they shouldn't be displaced while adjusting
the enigine when it is mounted below the layout.
|
Time to connect the contact strips, starting with the common connection: Solder a wire between the two middle strips. This wire is placed into the single-poled connector near the IC. Again, we us recycled UTP-cable wire.
|
The minus of the contact strips is connected between the strip that's the farrest of the PCB and the large minus strip on the PCB left hand side. The positive connection is made between the remaining contact strip and the positive power connection. We supply power again, and the geared block should slowly move towards the PCB board and halt when the contact strips are reached.
|
The wire, placed at right in the two-poled connector is the control entrance. When connected to the positive power supply, the enigine should rotate the other side until it reaches the other contact strips.
The left connection is the feedback signal. We can now adjust the distance the geared block should travel. In order not to damage the electronics: SHUT OFF off power supply
when calibrating!
|
Once our device is roughly calibrated, it can be placed under the subroadbed. The mechanical link can be made in several ways. By use of a 1mm steel wire, mounted in a drilled hole in the geared block, the device can be mounted flat with the steel wire sliding through the rectangular chassis base hole (the most common way) or straight (as in the picture) |
After mounting, the device can be finetuned, the default angle can be set and we are done.
Of course, this engine can be used for other purposes, too:
Mechanical semaphores, animation, ...
I know this isn't in any way a simple project. Hopefuly, I have provided you with some ideas on how to achieve slow motion on your own modeling projects. Have fun!
|
Some history, about an earlier prototype:
My first turnouts were controlled using a modified telephone relay, recycled from a small telephone central. These appeared later to be unsatisfactory, so I needed an alternative. I had to design them myself.
I never approved the standard turnout drivers, using solenoids. The electric built-in safety, if present, proved unreliable, causing the wires to burn. They perform poorly, and are quite loud, too.
above: telephone relay, modified and used a a turnout driver
When, at work, the telephone central was replaced, I managed to save about 150 relays from the dumpster, planning to use them for my future layout. A construction with a piece of U-shaped rod became a lever, the relay switches were useful for feedback to the control panel, and for wiring the turnout itself. They weren't that bad, but had some disadvantages: they were pretty noisy, used a dangerous 48V DC supply voltage and where difficult to install and fine-tune. Moreover, they where prone to temperature variations, needing fine-tuning every season.
Left the results of various prototype attempts, right the eventual prototype, with a view within.
The existing motorized turnout drivers (Bemo, Fulgurex, Switchmaster...) are quite expensive, even more so as I need about 100 of them on my layout. So, I would design my own switch engine. This seemed a very hard thing to do, as it was quite hard to find good and affordable motors, switches and gears. As you see on the upper left picture, lots of prototype attempts were made before I reached a good one. The electronics to drive it where designed as well. I found the engine, gears and micro switches at Opitec. The home-designed electronics are based on a logic pre-processing, controlled by a 40106 CMOS, the power driver is a H-bridge, using Darlington transistors.
Left: the working prototype. Right: the mould for a small series production.
During design, I took care to get a prototype that would be easily copyable, using a simple silicon mould. After construction of the mechanical parts, a U-shaped box is covered by the electronic circuit board. By carefully choosing the parts, I managed to get a relatively low-cost turnout driver, with some quite good characteristics:
Mechanical specifications:
- compact: only 35 x 50 x 80 mm
- movement adjustable between 2 and 12 mm
- stop points adjustable in both directions
- various ways for mounting and connecting
- slow motion (5mm per second)
- reproducible, solid construction
|
Electrical specifications:
- input voltage: 5v DC
- control using logic 0 and +5V
- real feedback with logic 0 and +5V
- possible to mount up to three micro switches
- low current: some µA's while waiting, 200mA while changing position
And ... cheap (about 8 Euro material costs) |
Left a nearly finished serial-built engine, right two mounted drivers
.... after connecting them to the layout power, I noticed the noise they made ... so far for the prototype ...
