BR89 (90% BB, 10% TLG) and TurnTable (99% TLG)

[Toastie]

OK folks,

this is so on-topic, because @zephyr1934 as well as @dr_spock and @Lok24 and @Ts__ and @HoMa, and, and (HA! Now I have you all spammed ...) replied ...

Why on earth - and I need to understand this before making a decoupler - and I need to make one because otherwise my Level 3 project is prone to utter failure (OK, yes, only if I use more than one car to shunt ... but you know ...) - WHY IS THIS working:

Let's assume, there is one engine (right) and two cars, coupled by magnets (1st row). Next, the last car (left) is bolted to ground (or blocked by some means) and I do accelerate the engine (2nd row). I have watched countless videos ... 9V/12V/FXTrack, home brew stuff): Why is it, that always the blocked (here last) car is decoupled (3rd row, position 1) and not the engine from the two cars (4th row)??? As far as I see it, the force F is acting equally on the coupling of the first car and loco, as well as on the coupling of first and second car. What is it, that I am not getting? Particularly when the magnetic strength between the connection points is even a little different? Let's say a little less between cars 1 and 2? 

None of the videos show I watched any faulty behavior, and I believe it is not "take lucky" out of 10 bad ones.

Man. I am a chemist. I like to blow up things - but this physics stuff drives me crazy. Well it does, because I am supposed to be a physical chemist, but maybe I should revert to solely chemistry.

Any ideas (I need to grab another beer, so take your time )?

Best,
Thorsten

 

[Ts__]

Because the videos show only successful attempts ;-)
I had also tried it first without 2nd clamp. Has often worked, but just not always. Sometimes (maybe in one of ten attempts) the wrong car was uncoupled. Therefore, I have in my hump yard two holding clamps. One is moved away by a carriage and pushes the waggon away, the other remains in position and holds the other waggons. This ensures that I uncouple at the desired point.

Thomas

Edited June 6, 2022 by Ts__

[Toastie]

22 minutes ago, Ts__ said:

I had also tried it first without 2nd clamp.

Saw that, that is entirely rock-solid, also Werner's proof of concept video showed that "clamped" car and then the push action on the other.

With your design, there is a 100% success rate.

Hmm, but even 1 in 10 is rather favorable. And this is not only owing to video clip selection. I am still thinking about it. There is something I have wrong with the partitioning of the forces, I guess, because there is this fixed point in the distribution. Don't worry, I like to think about such things - as my colleague over in Analytical Chemistry always says: If we were all PChemists, we still would be thinking about why the hell that fire lit up.

They eat, we think. Guess who survives.

Best,
Thorsten

 

Edited June 6, 2022 by Toastie

[dr_spock]

How would you decouple the cars by hand?   Hold one and pull/push the other one away.  Reproduce that action without human parts. 

I guess decoupling can be thought of as breaking covalent bonds by mechanical means. 

 

 

[Toastie]

6 hours ago, dr_spock said:

How would you decouple the cars by hand?   Hold one and pull/push the other one away.

Yes, that's the way I do it And that is the way Thomas is doing it with his setup.

However, when I have three coupled cars, I usually hold one and pull the very next car and not the third on the far end. That's my "problem" and that's the way many if not all other LEGO train decouplers I am aware of work: In a consist of 10 cars (#1 = pulled by the engine, #10 = end), if you want to split that consist in two halves, put a block at car #6 and simply pull cars #1-5 away.

Yes, it is like breaking covalent bonds - in this case thought, the weakest will break - and if they are of equal strength (lets say in a long CH2 chain - statistics always wins . Clear difference is: None of the atoms is bolted to some stop ... 

Actually, that's my problem - and I must miss something. I guess the force balance (pulling engine, coupled magnets, blocked end) sketched into my scrap book is wrong - and it may have to do with magnets loosing contact (on the differential scale). The fact that the last car is blocked, but the other(s) can move, affects the distance between the magnets in a way that the unblocked cars stay together/recouple with much higher probability than having the last car grabbing the next to it. The magnetic force between two magnets of opposite polarity with this geometry scales highly non-linearily with distance (https://www.supermagnete.de/eng/faq/What-factors-affect-the-adhesive-force-and-strength-of-a-magnet?). This must have to do with it, particularly because the adhesive force is >> than the force at even a very small distance.

Yes, I am a nerd.

Best,
Thorsten

 

[Lok24]

you can move the magntes between the two cars to the side or upside, that weakens the magnetic power .

You should define the preliminaries.....

The solution shown in my proof of concept and build and enhanced by Ts_ if suitible for every car.
That was the requierement.

If you have your own waggons only you can of course use special parts and constructions on the bottom   or something like that, you don't even have to use magnets.....

[JopieK]

Great to see how Powered Up can really improve our Train experience. Of course it can always be better but especially interoperability with ESP32 / Pi / whatever is really great!

[zephyr1934]

19 hours ago, Toastie said:

this is so on-topic, because @zephyr1934 as well as @dr_spock and @Lok24 and @Ts__ and @HoMa, and, and (HA! Now I have you all spammed ...) replied ...

My idea of an uncoupler is simply a bar that slides out and pushes the magnets apart (not all the way apart, just in a "V". Here's a bad picture using a technic beam for the pusher,

or a slightly different version with a pointed plunger in this video. (or see  switch-yard-1a.avi in brickshelf)

 

With your precise color code positioning you should be able to line things up to stop right where you need to for the plunger. If you come up from below, you can rotate a swing arm instead of laterally inserting a plunger

[Lok24]

8 minutes ago, zephyr1934 said:

My idea of an uncoupler is simply a bar that slides out and pushes the magnets apart (not all the way apart, just in a "V".

 

That's what I proposed this morning..
But cars with less weight are pressed out the rails, that's my experience.

 

[Toastie]

9 hours ago, Lok24 said:

You should define the preliminaries...

See, that is my problem: Instead of going after a solution for my setup, I am thinking about why the simple blocking type works most of the time.

Non-nerds go for the solution and solve the issue - I am still way behind and try to figure it out. Never engage with a basic science guy. They'll ride every horse to death and then claim to having solved the problem - in the middle of a desert.

Thinking of horses ... wasn't the 12V system using exactly that approach? Or did the rotating part of the coupler actually weakening or even pushing the magnets apart? I believe @Reker1000000 knows for sure, as he is currently going crazy on documenting/demonstrating the 12V trains system here on EB for us.

But thank you very much @Lok24 and @zephyr1934 for guiding me to a working solution!!!

All the best,
Thorsten

   

[dr_spock]

How about switching to Kadee couplers? 

 

 

[Reker1000000]

16 hours ago, Toastie said:

Thinking of horses ... wasn't the 12V system using exactly that approach? Or did the rotating part of the coupler actually weakening or even pushing the magnets apart? I believe @Reker1000000 knows for sure, as he is currently going crazy on documenting/demonstrating the 12V trains system here on EB for us.

I have been summoned...  glad to know I'm regarded as a 12v maniac/expert now! 

The rotor in the 7862 Remote-Controlled Decoupling when activated raises a bit and it interferes with the wheel blocks- the 2 prongs on the rotor prevent a wheel block from moving over it. At the same time, they also guide the pins on the coupling parts into position to decouple smoothly. For it to cleanly decouple it's vital that you are using new rubber rims on the motor, since they provide traction and the motor can then successfully pull away from the rest of the wagons. See about 5:20 in this video for an explanation, and there are some clips of the decoupler in action at the end:

 

[Lok24]

5 minutes ago, Reker1000000 said:

The rotor in the 7862 Remote-Controlled Decoupling when activated raises a bit and it interferes with the wheel blocks- the 2 prongs on the rotor prevent a wheel block from moving over it

Yes, and that's why it might not be suitible for all MOCs.

[Reker1000000]

1 minute ago, Lok24 said:

Yes, and that's why it might not be suitible for all MOCs.

Of course- it's a very niche set and has very limited functionality outside of the 12v system. But, one could draw on the mechanism of 7862 to make their own decoupling unit!

[zephyr1934]

2 hours ago, Reker1000000 said:

The rotor in the 7862 Remote-Controlled Decoupling when activated raises a bit and it interferes with the wheel blocks- the 2 prongs on the rotor prevent a wheel block from moving over it. At the same time, they also guide the pins on the coupling parts into position to decouple smoothly.

Interesting to know, I always assumed it simply guided the couplers apart and nothing more.

 

14 hours ago, dr_spock said:

How about switching to Kadee couplers?  

That is probably the most space efficient solution, and since it is only a couple of couplers involved, probably not that expensive

[Reker1000000]

5 hours ago, zephyr1934 said:

Interesting to know, I always assumed it simply guided the couplers apart and nothing more.

 

Yeah, I thought that was how it worked as well before I actually got my hands on the set.

[Toastie]

Thank you all for your replies, tips, and solutions!!!

On 6/7/2022 at 11:56 AM, Lok24 said:

you can move the magntes between the two cars to the side or upside, that weakens the magnetic power .

Yes, this seems to be a very good solution - although it may not work for all situations, as you already said.

On 6/8/2022 at 1:43 AM, dr_spock said:

How about switching to Kadee couplers?

Very nice video and so elegant! Thanks a lot!!!

On 6/8/2022 at 2:04 PM, Reker1000000 said:

I have been summoned...

Heehee ... thank you very much for your reply. I should have watched your video more carefully. Maybe one last question: Do the magnet come a little apart, when operating the switch?

Best
Thorsten 

Edited June 9, 2022 by Toastie
the or not to the, that is the question

[Toastie]

So I finally figured it out – I mean, why “simply” blocking the wheel set of an arbitrary wagon in a consist leads to decoupling at that very position. Remember that I had headaches because of not understanding that? They’re gone.

I am not talking about the “weakening” of the magnetic coupling between two magnets by pushing them slightly apart (as discussed and shown by @zephyr1934 and @Lok24) or even blocking one car and pushing the other away, as shown by @Lok24 and @Ts__. Pros and cons have been discussed here and elsewhere. These are very smart solutions and render the section below meaningless. However …

I am just talking about a simple one-side blocking – realized by others in many ways. Even the 12V train system does that (cf. e.g. @Reker1000000's videos in the corresponding thread here on EB as well as on other YouTube channels); also the 4DBrix decoupler does it that way: It simply blocks one wheel set to move any further when the exact position of magnetic coupling between two cars has been (automatically) detected:

 

However, a quick experiment with a consist of 5 different cars, some heavier, some very light, clearly shows that simply blocking does not decouple the blocked car(s), but the weakest magnetic coupling of the consist. As I expected. And always. When you have some more cars, and you find pairs of magnets, which create (almost) identical magnetic force between them, then decoupling occurs statistically along the consist. The latter experiment took me a while, but eventually I found some cars behaving like that. And again: As expected, see my little drawing somewhere above; the pulling force of the engine (or my hand) is equally distributed over all couplings – and naturally the weakest will decouple. Or if they are all identical, statistics rules.

Now, why on earth do all the “simply blocking the car/set of cars you want to decouple" designs work, then? We speculated it was: “Shown only the attempts, when it worked”. Hmm. I doubted it. And why did my experiment demonstrate, that it does not work? Further investigation strongly suggests: It is the acceleration (and mass) of the entire consist that generates the required surplus force acting on the coupling to be decoupled.

This is the difference between my experiment and “all the others”, favorably shown in the 4DBrix video, as here you can clearly see what happens:

• In the 4DBrix video one can see, that when the coupling is precisely located (man, this is a nice way of doing it!!! Using a Hall sensor – very cool), there is still some considerable "slack" (distance) between the wheel block and this position. Then the engine is fired up (and in many videos out there people point out to use new rubber bands on the wheels, to quickly move the engine away, etc.) accelerating the entire consist, and only then one car runs into the block. At that moment, the car(s) to be decoupled instantaneously stop, while the remaining consist has gained some speed, caused by the acceleration, and the latter leads to an additional force into the direction of the moving consist, i.e., at the coupling of the blocked car, but not at all the other couplings – as they are still moving.
• In contrast, in my experiment, I blocked the last car at one end, then pulled carefully at the other end and moved the consist into “position”, i.e., the last car touched the block. Then I carefully pulled with increasing force the car at the other end. Decoupling was either at arbitrary positions or at the weakest coupling. Always. And hardly at the blocked car.
• When repeating my experiment, adding some “slack” between block and wheel set, the expected behavior of a decoupler results: It decouples where it is supposed to decouple. The smaller the acceleration is, the less reliably it works. Success rate is 100% when the acceleration of the entire consist is large enough to generate sufficient additional decoupling force via F(add) = m a, where F(add) denotes the additional force acting on the critical coupling, m is the mass of the remaining consist, and a is the current acceleration when running into the block. A sufficient distance between block and “decoupling position = beginning of acceleration phase” is required to allow the engine actually building up some acceleration (note that velocity is not critical and that a of the consist remains zero when there is no distance between blocking and starting position).

In other words: All is in harmony.

Best,
Thorsten

Edited June 16, 2022 by Toastie

[Ropefish]

Bit late to this, but i will add on to kadee couplers, however.... most of my experience in O scale is that people just use a skewer of some kind to decouple them as the magnetic system sometimes just doesn't work right or is to messy to set up. I still love the kadee's though as they are just perfect, and i would say bit more affordable then the lego magnets now.

[Reker1000000]

On 6/16/2022 at 12:26 PM, Toastie said:

At that moment, the car(s) to be decoupled instantaneously stop, while the remaining consist has gained some speed, caused by the acceleration, and the latter leads to an additional force into the direction of the moving consist, i.e., at the coupling of the blocked car, but not at all the other couplings – as they are still moving.

Well said! As you mentioned, I believe 7862 uses this concept because you can decouple wagons other than the one directly behind the locomotive.