Jump to content
  • entries
  • comments
  • views

Magnetic Uncoupler Mechanisms for Freshwater

Ian Morgan



18 July 2010: Following the Expo, I can now get on with adding the uncoupling magnets. I had always intended using permanent magnets, following their successful use on Brunswick. I bought some small magnets from Merg, but they were nowhere near powerful enough to operate the DG couplers from below the baseboard. Now I have purchased 50 'rare earth' magnets via Ebay. These are really powerful for their size (about 5mm diameter and 3mm long) and come with all sorts of warnings. 'Rare earth' magnets have a high iron content and would corrode very quickly if not protected by a special coating. Often this is nickel plating. It is therefore not practical to cut them, and you have to protect the plating from damage. If you let them bang into each other, small fragments can break off.


I planned to use some more servos, as used to operate the points, to move the magnets into position under the track, or away from the tracks when uncoupling was not required. The layout has several parallel tracks at various positions where I wanted uncoupling to take place, so I planned to have a number of magnets alligned to the parallel tracks, all operated by a single servo. I cut a length of paxolin sheet, and drilled it to take the magnets, which are glued in with epoxy. Working with the magnets is interesting, as they fly around the workbench attaching themselves to tools each time you get anywhere near them. Trying to get three magnets into their holes in the paxolin to glue them without them flying to each other was also quite fun. It is also no good trying to apply the glue with a metal screwdriver.


I made a paxolin 'arm' which bolts to one of the servo attachments, and the bar with the magnets attaches to the arm, using bolts and a paxolin block with threaded holes in it. Because of all the tracks above where the servo needed to be, I cut a sheet of hardboard to attach the servo to, which bolted under the layout away from the tracks. The photos show how it works, I hope. One shows the magnets moved well away from the tracks, and the other two show the magnets in position hard against the underside of the layout surface, below the three parallel tracks there. You can also see the Merg Servo4 board that the servo plugs into. Now I just need to build two more.






14 Aug 2010: Not much progress on Freshwater for a few weeks because another problem arose with the servos used for points and uncoupling magnets. As I was adding more and more servos to the layout, I started getting problems switching on the power. On switch-on, the servos would move rapidly, hitting the physical limits of the mechanisms they are attached to, and carry on trying to drive past them. The current drawn by all the servos was collapsing the voltage from 16V AC down to about 3 volts. All this was happening before the PIC microcontrollers had time to initialise and bring the servos under control, and the voltage then dropped below the PIC operating voltage, so they stopped doing anything.


Consulting the Merg forums, I was advised to add some pull-up resistors to the servo signal lines. Apparently this stops the switch-on spike which sends the servos trying to reach their end limits long enough for the PIC microcontrollers to initialise and start sending sensible control signals to the servos. The pull-up resistors are 10k ohm, large enough not to affect normal operation. A change to the PIC software to speed up the initialisation process was also suggested, but has not been found necessary. Having added the resistors, everything seems to be operating correctly again.


So, now I just have to build one more servo operated uncoupling magnet set and the electrics under the layout are done. Then I can start ballasting the track. Other jobs in the queue are:


- building a lighting gantry

- adding backscenes, and a transportation case

- replacing the temporary control switch box with a proper control panel (using CBUS kits to allow route setting)

- building a larger fiddle yard now the GJ challenge constraints are out of the way

- creating buildings and scenery

  • Like 1


Recommended Comments

That's a neat solution to moving the magnets, Ian.


Glad to hear that the pull-up resistors work. I've been wondering about this issue for a while as, although I've no problems with the three servos already installed, I have another dozen or so to add in the coming months.



Link to comment
  • RMweb Premium

I believe the Merg Servo4 boards are being redesigned to allow easy addition of pull-up or pull-down resistors. The type of servo used will dictate what is needed. Also, I think a change to a beefier DC supply might improve matters.

Link to comment

Sorry, for us non electronic types who also use servos, can you please explain what a "pull up" resistor is or does and where it is fitted in the system. Also would it not be possible to ensure that the voltage (or current?) does not drop off by using a higher rated supply?


Thanks, and sorry if this is too basic for you electronic wizards!

Link to comment

Thanks for this, Ian, I was going to pop the question on the board about using servos in place of point motors, etc, as point motors around here are silly beans and servos can be got for about 12€ and I was thinking about seeing if they were interchangeable in their roles.

Link to comment
  • RMweb Premium

Sean, yes servos are dirt cheap on ebay and have plenty of advantages over traditional point motors. If you remove the springs from Peco points, you can make a servo switch them in slow motion. See the point tie-bar entry on this blog for a mounting suggestion. The one disadvantage is that you need something more than just a switch to control them. They need a series of pulses to operate. The length of the pulse specifies the position you want the servo to be at. The servo has inbuilt electronics that will make the servo move to the specified position. Merg make a kit for a circuit board that can control 4 servos using 4 on/off switches. It does need a special set-up box, or a PC with a serial port, to do the initial setup, setting the extremes of movement required and the speed of travel between them. However, the cost, spread over more than a couple of servos, is still far less than Tortoises, etc.


ScRSG, from what I understand, if the input to the servo is left to float at switch-on, the servo sees a very long pulse and tries to move the servo arm as far as it can go. The microcontroller on the Servo4 board takes a finite time to boot itself up before it starts sending sensible pulses to the servos. If lots of servos are all racing to their endstops at the same time, the voltage of the supply can drop lower than the microcontroller needs to operate, so it never starts sending proper pulses. That is what I had, with supply voltage dropping below 3V. My problem was solved with the addition of two resistors for each servo. One 5k ohm resistor goes from the servo input to the positive supply, and a smaller 1k ohm resistor is between the microcontroller output and the servo input. The 5k ohm resistor lifts the input to the servo to supply voltage while the microcontroller is booting up, and the servo seems to not see it as a pulse and stays still until proper pulses start coming from the microcontroller. This fix does increase power consumption a bit all the time though.


Yes , a meatier power supply might help, but all the servos would still twitch on power-up until the microcontroller has started working, and the existing 16V 1A supply is fine once everything has powered up properly.

  • Like 1
Link to comment

Thanks for the further information, Ian, I will definitely be looking in to them for the next layout I'm going to build (only got one point on the current, and wire in tube works fine for it, as it's just a skills refresh more than anything).


I've a big box full of servos in the basement, so that's where the initial thoughts came from, so one of the set up boxes would put me ahead of the curve on cost to start as well.


Once again, thanks for the info and I'm going to try and put it in to operation sometime in the near future!

Link to comment

Sean, try a search for something like 'servo turnout motors'. There are several recent threads on the subject, including this, this and this. Be warned, though, some of these threads eventually descend into a tedious 'my way is cheaper than yours' or 'MERG vs the rest' arguments.



Link to comment

Thanks for your response, I think I understand but at some point would like to see a diagram or photo of where exactly the resistors go. As an aside did you see the circuit printed in MRJ No 205? Would you have any comment.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Create New...