Mounting an uncoupler on an LGB 2070 loco

[curtis]

Over the last day or so, I've been fitting a DCC servo-controlled uncoupler to an LGB 2070 U-Class. I thought I'd share the process and my learnings should it help others considering something similar.

You can see a video of this in action on Instagram YouTube:

I should start of my clarifying this is not a "plug and play" solution. In order to get the uncoupler installed, it required some changes to the uncoupler, trimming the plastic of the arm holding the trailing wheels and drilling a hole to run the wires from the servo to the regulator & decoder. Given I intend to hold my collection until I can no longer play trains I didn't have too much concern about resale value and enjoy a challenge and learning something new.

This is my second loco to add a DCC controlled loco uncoupler..The first attempt was an LGB 2015 which also required internal "modifications" (read: dremmel and drilling the inside of the tender) to setup but is working well ( 65 1057 can attest he saw it live during our running session). For me, this flexibility made things a lot more fun so wanted to slowly expand the functionality to the rest of the stock.

Parts

• x1 Heyns Uncoupling Arm (purchased as a set from Modell-land.de including the wire)
• x1 SG90 Servo (bought as a set from Amazon)
• x1 Milliput (although I only used a fraction of it)
• x1 6v Voltage Regulator (from MD-Electronics)

These are the things I required as my decoder (MD Electronics Drive-M) already supports servos with some CV changes.

Placement
On the LGB 2015 (my previous attempt), the hook/loop coupler is attached to the body of the tender. This means the servo can be located in the tender and the fishing wire fed down to the coupler. However, on the LGB 2070, the coupler is attached to a pivoting trailing wheel via an arm. The servo is configured for very specific range of motion so mounting the servo in the body itself was a no go. Unless positioned perfectly at the point of the pivot (which was not possible) it would have meant a variability in the length of the wire between the servo and the mechanism. The only choice was to mount the servo itself to the arm of the trailing wheels.


First, I had to make this wouldn't interrupt the functioning of the loco. I took the servo and some blue tac and mounted the servo to the top of the arm just below the body of the loco (their wasn't room below). I then carefully ran the loco through various s-curves (R1, naturally) and a mess of points to find all ran well. The initial prototype, at least, worked well.

Modifying the Heyns Uncoupler
The Heyn's uncoupler functions to uncoupler wagons connected both with a single and double hook. It works by pressing a plate down into to the hook. This has the double effect of pressing down the hook from the connected wagon and the hook on the loco itself. Once depressed, the loco can simply move forward...

In theory at least. I found with both locos the range of motion isn't quite enough for the plate to push the hook of the wagon far enough down so I've had to pad the plate with additional thickness (example below from the unfinished Spreewald)


For the Tenderlok, I use PVC sheeting but for this I used some milliput. For those of you that haven't used it, milliput is a modelling material that has the texture of bluetac to work with and then sets solid in several hours (so solid, you can drill into it). I used this to create the additional padding and tested it with a couple of wagons.



Attaching the Servo
This was one of the more challenging parts of this endeavour. I tried using milliput to create a structure so I could swap our servos if (when..) I break them. That didn't quite work. I considered just supergluing the servo to the arm (definitely the simplest) but again, when I break them means I'd have to replace the entire trailing wheel arm which seems wasteful. In the end, I decided to superglue a yellow track insulator in the right position so I could screw the servo to it. This seems a reasonable compromise for maintainability. It was also the right decision because when initially programming the decoder I misconfigured it servo and damaged the gears on the first one )

Wiring and configuring the servo
This was the easiest part of the install. I already had a Dec+ running to the back of the loco for the rear and interior lights. The Dec- was running to the cab too because I've used the cab switch (the one in analogue that disables lights/motor) as a switch for the DIY Power Buffer. I simply extended a wire from the Dec- to the voltage regulator. The only additional wire I had to run was for the function output which was easy.

I drilled a hole in an accessible place to run the wires from the servo up into the cab and wired in the servo to the voltage regulatory. I then used computer software to configure the decoder and a basic range of motion for the decoder. I tested this without any attachment to the uncoupler to avoid damaging anything.

Once I had a range of motion that looked reasonable, I set the servo to the uncoupled position and attached the servo to the uncoupling mechanism wire the attached wire. I then disabled the uncoupling function to restore the servo and the uncoupling mechanism to the "normal" position.

You can see the configuration in the first image.

What's next?
I want to do a few sessions with the attached prototype before I consider it a success but from a 15minute play this evening I'm happy do far. Once I'm confident it works, I'll clear up the modified arm and paint in back (not that it is noticeable from the outside) but makes me feel better. I also need to paint the white milliput clay either in the gold/brass colour or black.

Let me know your thoughts or if anything was unclear!

Curtis

 

[Dagnall]

Curtis.
Very interesting .. I did not see any mention of modifying the hook "vertical" part to adjust how far the hook depresses when the lever comes down...It might be interesting to comment on that?
Out of interest, I independently modelled a plastic version of this "Heyn" decoupler for 3D printing, just to explore the geometry.
Its available here Heyn type decoupler 3D model The "real" metal version will of course be much stronger, but this one is cheap to print and could be useful for experiments.
My preferred hook and loop decoupler is this one that I designed -Its a modification of one I designed some years ago. For you it would avoid the servo mount issue as it uses a tiny linear RC servo internally mounted. - BUT these servos are only marginally powerful enough for reliable use. - Which is why I changed to 3D knuckle couplers of my own design, and they have their own electric knuckle decoupler . The advantage of the knuckle decouplers is that the servo is only needed to pull back the wire locking lever- which requires far less force than pushing down the (potentially two) hooks. Also, decoupling a waggon from the center of a train only requires you to lift it off the track! Much easier than hook disentangling!. Also on my rough track, the knuckles stay engaged more relibly than single hooks, which were forever disconnecting on curves and bumps.

I hope I have given you some ideas ?

Dagnall

 

[curtis]

Dagnall , good catch! I forgot to mention it. Yes, I had to clip the vertical part of the hook after adding my milliput adjustment to ensure I got the range of motion from the plate and the hook. Honestly, I wasn't too scientific about it, I simply trimmed a small amount off the plastic with snips and checked. [For those wondering what the "vertical" bit is, check the second photo above of the loco and the carriage connected. On the right you see a vertical bar that sticks up into the hook]

Wow, looks like you're much further along in loco mounted uncoupling than I am with a lot of learnings! Your designer for the 3D printed one - does the plate push forward to almost pry the opposing hook out of the loop? Your setup is much more discrete than mine - I think my approach only works on the U-Class because of the coverings around the rear.

Honestly, if it was a different design that wouldn't be viable. Regarding the micro servo you used - is it similar to the one Mattoth uses for their uncoupler? I did check out their offering but the price seemed too steep for something that only appears to handle its own hook (and not the mated one too).

The knuckle coupler is clearly the most advanced form - kudos on the development path. Looks like it works perfect! Do you have any troubles with it? Also, I'd love to hear more about how magnetic (?) lighting connectors your mounted!

 

[PhilP]

I am also interested in the magnetic couplings for power.

I might have dreamed it?
Wasn't there a plastic moulding available, to slip the wires / magnets into?

PhilP

 

[Dagnall]

Hi all.
Magnetic power connector
This was designed to be an easy way to take power from the "accessory" socket (the plastic moulding that perhaps PhilP is remembering? ) on the Stainz and into the coaches for illumination.
You can see it in this video about the more recent decoupler.
It is designed to clip on the "standard (?)" LGB single central buffer mount. I do not glue it, but it would probably be a good idea to use some light adhesive.
I published the stl files and instructions in thingyverse here The key thing to understand is that the Magnets themselves actually make poor connectors (I think they have oxide surface layers?) so the actual contacts are the exposed tinned copper wires that you push through the holes in the 3d printed magnet holder. Make sure that they will touch their corresponding "opposites". The magnets just hold the connectors together tightly.
They are not suitable for any significant power transfer, but fine for lighting.

Massoth decoupler.
I do not have a proper massoth decoupler, but did look at pictures!. I think its just got a small motor with a lead screw. I have heard that it is easy to "jam". I think this may be because it screws tightly to each end and may not have the strength to "unscrew" itself. This is certainly a problem with the micro linear servo, but the servo (usually!) avoids the endstops as it is a proper servo.. I think the Massoth design just screws until the motor stalls. I have no idea if it has any electronics, or how it is driven by the decoder. (I would love to be educated by someone how uses them!!)

My original "magnet spring" decoupler used magnets to "bring the hook up", and also to centre it, I have a couple of versions of the "pusher", but you are right, it pushes backwards and closes off the "loop", hopefully pushing down and disengaging any hook from the "other wagon". The technical difficulty is that it must also disengage its own hook from the "other wagon". If the magnetic force is too strong, the hook can easily stay "up" and the servo then does not have enough strength to break the attraction. What happens then is that the motor overheats and the servo is usually toast.

I did revisit this design at the end of last year, and when (eventually!) the crucial 0.2x4x10 springs came it showed some promise. In this new design, the spring is mounted at the "bottom" of the pivot, and the pivot is part of the hook. The selected spring has just enough force to keep the hook nicely up. (A biro spring is far to powerful here and you must get these quite weak 0.2mm springs.)
I have not actually tested this on a train as I still think the nano servo is too weak. (also I burn out another servo trying!)

I am currently exploring using tiny stepper motors - which are reasonably priced and about four times more powerful. The current idea is to have an external circuit to drive them from an RC servo signal, as shown in the proof of principle video. But I have not integrated this into a decoupler yet- and my first look shows that even this the stepper motor drive is probably still a bit big. But the good news (at least in comparison to the Massoth) is that it can reliably drive itself to each end, stall the motor and STILL reliably start in the opposite direction!.

Thats quite enough links for one post!!.

Cheers.
D