The Train Doctors Train


                      Decoder mounted - speaker plug at left                               Rear of decoder mount showing earth lug and

                                                                                                                             socket to plug the body harness in


The Train Doctor’s Train (Well, the Train Doctor’s Loco, really) exists because I picked up this scrappy looking old V200 that I didn’t really want along with a load of bits and pieces that I did want.  So that left the question about what I was going to do with the V200:  Sell it as is? – No, it’s probably only worth about $20 tops;  Do it up and sell it? – No, it’d cost more to do it up than I could possibly get for it; Scrap it for parts? – No, I hate doing that sort of thing, besides which, parts for those old V200s aren’t so scarce that one needs to do that.

Chronic Märklin purists with high blood pressure had better stop reading about now, lest they have an apoplexy fit, because It seems that the only option left (apart from chucking it in a drawer and forgetting I had it) was to make it look more interesting and keep it, so what about giving the old girl a new lease of life by painting it up for a spot of advertising  (not so farfetched, surely).



 a spot of advertising


















Not being so much of a dab hand at that sort of thing myself, I removed all the body fittings (buffers, handrails, lenses, windows, Preiserling drivers’ old newspapers and cigarette butts etc), drilled a 3mm hole in each end of the roof to take a couple of LEDs, and entrusted the body to my friend Fred Spencer (who is a dab hand at that sort of thing) with only the vaguest brief to him about what I wanted.   Well Fred really turned up trumps for me and came back with a masterpiece, but more of that shortly.


Being a Train Doctor’s emergency vehicle? I decided it needed equipping with sounds and some special lighting effects.    I’d done something similar one other time using a loksound M4 decoder, but on this occasion elected to use one of the Märklin mSD decoders, (a) because I had several in stock; (b) because they’re slightly cheaper than the equivalent loksound M4 decoder but perform just as well; and (c) because I’d just received Marklin’s new sound programmer and wanted an excuse to try it out.

The Märklin sound programmer as yet isn’t as versatile as its ESU lokprogrammer cousin, and it appears that its ability at this stage is limited to compiling a sound file and writing it to a decoder, never the less it’s still considerably less painful than downloading sounds with the CS2.   Actually the makings of much greater functionality for the sound programmer are obviously there, if not implemented yet, but I digress – more on that some other time.  

On to the digital conversion piece then:  In this case of these old diesels and similar Eloks, the easiest way is to remove everything from the chassis (Motor bogie, reverse unit, and also the idler bogie.)  The motor is removed from the motor block, gears and wheels checked for tooth and spindle wear, and the new 5 pole motor parts fitted.  I normally check the motor block out with DC at this stage to make sure it runs nicely and there are no mechanical hang-ups, and also replace the tyres because it’s easiest to do when you have the motor bogie out of the chassis and in your hand. 

The idler bogie is removed to clean up a piece of the frame to solder a grounding wire on to.  (I bond both bogie frames to the chassis to make sure the finished product will have decent grounding.)  This is much more important with DC motors as used in digital conversions that it is in the AC (field coil) version of the motor.  I also remove the wire from the pick-up shoe contact at this stage, because I’ll be soldering the red wire from the decoder onto here when I get that far.



OK, that’s it.  Decoder next up.



















































Marklin supply a mounting bracket for the decoder motherboard with their 21 pin decoders and it’s pretty easy to screw motherboard and bracket onto the post that the reversing unit used to be fastened to.  I also put a grounding lug under that mounting screw at the same time.    Because I wanted a couple of alternate flashing lights on the roof (just an attention seeker at heart) I needed to come up with a plug and socket arrangement to allow the body to be removed.  An old 8 pin decoder plug (with most the pins chopped off) did the trick and a small bit of veroboard on the inside of the roof makes a suitable termination point for the 2 functions and a convenient place for a resistor to avoid smoking the LEDs













Motor Block.  2 lamps at each end glued temporarily into position.













Headlamps are standard 2 pin Marklin lamps via the “lichtkörper” and I also fitted a red lamp at each end (that utilizes the same lens as the headlamps) to allow for tail lights or a red warning flasher.  I used a separate output from the decoder for each red lamp so that I could have them separately switchable from the headlamps, and this also allowed me to have them flashing when the loco is stationary.  

Unlike the loksound decoder, the Marklin decoders don’t (at this stage) allow for alternative effects (e.g. flashing or steady state) on the same auxiliary output, but this can be overcome as the Märklin decoder has a couple of extra amplified aux functions over the loksound offering, so a little extra wiring can achieve the same effect.  The functionality of the lighting is such that in normal travel one can choose whether one wants the tail lamps on or not, and when the loco is stationary, one can have each end flashing red alternately from the flashing red lamp on the roof.  Confused?  So am I – I’m not sure that I explained that very well.















Back to the body, as can be seen in the pix Fred has painted the whole thing in a satin white, with my “Train Doctor” logo on each side, plus a "Marklin digital" logo and some other misc markings (“doctor” signs, phone numbers, web site etc.)  Before re-fitting I gave the hand rails a squirt of red paint instead of leaving the customary nickel silver plating, and cleaned up the windows etc.















All in all I think the effect’s pretty good and I reckon Fred’s done me really proud.  A big thank you to him.   















































The intent of this page is to describe a few interesting projects, conversions, or fix-it jobs I've completed either for myself or for other people.   

If this page fails to grow much it's probably not because I'm not doing these jobs, but more likely that I'm so busy doing them or other repair work, that I haven't had time to write about them.  

And here we go with the disclaimer again:  If you're attempting to copy any of this stuff or follow my "instructions" you're welcome to do so of course - but please remember that the risk is all yours and if you ruin some nice piece of model railway hardware I can't take any responsibility for that.








The Shunting Tractor ..............A project inspired by the Märklin model 46770

Have Propeller, Will Travel.....A digital/sound conversion on a 3077 Schienenzeppelin

ET87 Modifications..................A problem with the 37266 (ET87 ET501-506) - not staying on the rails.

The Train Doctor's Train.........Tarting up of an old 3021 V200.  Sounds, effects, and a new paint job.

Braking Module for your Signals.....Slow down your trains to avoid the "crash stop" at signals.








There have been many words written about Zeppelin conversions, especially around how to handle the propeller motor.  The easiest approach is simply to wire the propeller motor in parallel with the main motor, and although one could wire this prop motor through a diode to prevent it turning backward when the “electric jockey motor” is supposedly driving the zep in reverse, the totally un-realistic starting of the propeller at the same time as the zeppelin starts to move down the track was a show-stopper as far as I was concerned.     Because of this most “experts” seem to opt for a dual decoder approach - one for each motor - which overcomes this difficulty, and by tweaking the speed table of the main motor decoder or by programming a starting delay into it, one could satisfactorily get the prop spinning before the journey begins.  


If one was building a “mute” zeppelin, this would be a simple way to achieve an almost satisfactory result, but in the case where sound is required, it seemed to create as many problems as it resolved – one would need to use the sound decoder to drive the propeller to get the sound in sync with the prop, and the other decoder for the main motor.  But then what about sounds such as the squealing brakes?  If one wanted to stop the zeppelin but leave the prop motor running (to hell with Preiserling H & S issues) then the brake squeal sound wouldn’t work.  Also, I wanted to start the propeller with a jerky “starter motor” type motion (this is an Internal Combustion engine remember, not a turbine) and I couldn’t see a way to achieve that with the motor output from a decoder.   


The answer then had to be to run the propeller motor from a function output of a single decoder, and this line of thinking was ratified when I checked the ESU website to see how they’d written their sound project for MFX: - For load reasons they had actually allocated 2 auxiliary outputs in parallel for driving the propeller.


Have Propeller, will travel.


This is the story of a rather interesting and challenging digital/sound conversion on a Marklin 3077 Schienenzeppelin - you know, that funny cigar shaped thing with the electric finger nail trimmer mounted on the back.  Interesting in that it is definitely different from the usual run of the mill conversions (2 motors each doing their own “thing”), and challenging in that before the project was even half way through, Märklin announced their “new item” sound-equipped zeppelin, which meant that there was suddenly a benchmark set that I had to exceed (just to prove that I can do it better).












Just Needs a Pair of Wings

































OK. Methodology decided on - Let’s get started.


1st ingredient: Take one zeppelin, open up, remove power bogie, pull the main motor to pieces, clean out, install new 5 pole motor, check operation on DC power, Result: – horrible.  A worn gear boss with a bit of slop is causing lumpy running at slow speed.  Replacement of gear and gear pin cured that and now we’re all nice and smooth mechanically.  (So far this is all standard stuff that goes into almost any digital conversion).  The propeller motor as fitted is a small DC motor so I opted to leave that as is – maybe a mistake in hind sight because it is a bit of an el-cheapo, and is now due for replacement, but that’s another story.


2nd Ingredient:  Take one blank MFX sound decoder and load with Zeppelin sound project.  Plug it into the decoder tester and see what happens: - Yes, it sort of works as designed, but ESU seems to have done some funny things with the function mapping, and the project as downloaded also has the propeller starting up as the zeppelin moves off.  Never mind, that should be easily remedied with a few tweaks. 


Time to wire it into the Zeppelin and do it for real:  I decided I had to do away with the 2 functions in parallel driving the propeller motor – it seemed a little inelegant to say the least, and I could see I was going to run out of physical function outputs, so I used a single function output to operate a relay that switched suitably rectified and resistor-ized (if that’s a word) track power to the propeller motor.


If you read the first few paragraphs of this diatribe and didn’t just start in the middle somewhere, you’ll note I said I wanted a “jerky” propeller start-up and shut down, and so during these stages I drive the prop motor from a second function output operating in “random effect” mode. This generates a series of random pulses at the output.  I probably could have used a more regular (e.g. flashing) output which may have looked better if it could be timed to the motor start-up noise, but my customer’s requirement for a strobe light on the rear (fussy sod) foiled that plan – In the ESU decoder the frequency of the strobe flash has to be set as a common for all function outputs, and the desired strobe light frequency was about one per second – far too slow for the prop start-up. The strobe light itself is simply an SMD LED mounted on the body, and the wiring to this is via a small header plug so the body can be removed easily.


The propeller (once through its start-up routine) actually turns at a constant speed, but the sound of the motor revving up before “take-off” creates an illusion that one’s eyes seem happy to accept, and the overall effect is quite satisfactory. 


The other aspect of the physical bit worthy of mention is that I forsook the ESU supplied teensy weensy speaker for a large rectangular one pointing straight up. The resulting cacophony is tremendous, but happily the layout for which the Zep is intended isn’t populated with Preiserlings so there are no 1:87 protestors or Preiserling noise abatement societies to deal with.  (I might mention here that I did test the Zep on my own layout before handing it over to my client, and the raucous protests of the Preiserling rabble made Colonel Gaddafi’s recent problems look like kids squabbling over ice creams at a Sunday school picnic)














Noise maker extraordinaire
































At this stage I seemed to have a working zeppelin that did most of what was asked, but I struck a strange phenomenon:  Every so often the propeller would stop intermittently just for a fraction of a second – (although main motor and sound continued) and the headlights would also turn off briefly at the same time.  This had me head scratching, checking wiring, and doing all sorts of dismantling, until the discovery that the problem was to do with the decoder itself, and it only happened when sound was switched on.  I later proved that with sound switched on, all auxiliary functions exhibited this strange behaviour, but turn sound off and all was well.  Tests with a second decoder produced the same result and I came to the conclusion that the Zeppelin MFX sound project itself had some glitch in it.  I made enquiries on ESU’s forum describing how to test for the problem and ESU answered acknowledging the error and saying that they were re-writing the project so just wait.  Several months later at the time of writing this, the project on their website still hasn’t been replaced, so it’s just as well I decided not to wait, and I managed to make suitable modifications to the sound schedule that cured the problem.


Because I’d bashed the original sound project a bit to make it work satisfactorily, the actual synchronising of the sounds with the physical functions got a bit complicated, took me way outside the usual function mapping stuff that’s routine with these decoders, and won’t be discussed here save to say that it was such an educational exercise I even found myself discussing it with the Preiserlings (who were no help at all), the cat (who walked away before I had finished explaining), and the wife (whose contribution was to ask when I was coming to bed).


Aah well, we learn all the time and I hope the result is worth the pain although it turned into a labour of love in the finish, and I must have spent tens of hours on it.  One might think that’s the end of the story but there are apparently some other modifications coming up - Interior seating and lighting for starters.  I can see a massive re-wire coming up here because having used the f0 functions on headlight and rear strobe, and the f1 and f2 functions on propeller antics, I’ll need to use the aux 3 output from the decoder to drive the interior lights, (needs an amplifier), and I’ll also need to re-organise the whole of the physical bit so as to be able to fit the interior seating and lighting etc………….Watch this space






































Fingernail trimmer and strobe light                                                             Propeller motor and switching relay

































Not a lot of room in there



















The Shunting Tractor






















The ingredients


Here's a do it yourself Tractor Driven shunter for a small industrial backyard.


Ingredients:  One wagon chassis, one old style tractor model, one volunteer Preiserling to act as driver for the aforementioned, a decoder, a motor, several hours of spare-ish time, and a few other minor bits and pieces which you’ll see when we get to them. 


This wee project was inspired by the Märklin model 46770 (a Lanz tractor or bulldozer mounted on a frame and converted to a small shunting loco.)  My first thought on seeing the model in the 2010 New Items Catalogue was “pity they couldn’t find room for a motor”, but I thought I would probably buy one anyway and permanently attach it to a suitable covered goods wagon. The covered wagon would house the motor and a decoder so that the visual effect would be that of the “tractor-shunter” pulling/pushing the wagon, even though the reality was vice versa.  To that end  I purchased a suitably dilapidated looking second hand box wagon even before I really had any idea about how I was going to do things, and that was sufficient to trigger off a chain (train?) reaction and get the whole project moving.

Now, why was it I was going to buy that Märklin tractor model?  Surely any train fixer worth his salt could rig up something with an old wagon chassis and a model tractor that would look the part.   Aah yes, I’ll need a suitably old fashioned looking tractor then.  Easier said than done - There’s lots of modern HO tractors around in the “Siku” range and no doubt other brands as well, but models based on old looking prototypes seemed harder to come by.  Eventually I settled on a 1960s looking farm tractor from the Hornby range, and while being “00” gauge (1:76) rather than HO, I figured that the smaller prototype would make it look pretty much at home.   Next: A small chassis to mount the tractor on.  As it happened I had one of those old red tipping wagons that I’d come by as part of a “lot” with something else that I did want sometime earlier, so the removal of the broken tipping bucket and its brackets resulted in a nice short chassis just right for the job.

I wanted to rig up a gear train from the chassis wheels to the tractor so that the tractor wheels would turn while the wagon was being propelled around the yard, else it would just look like a tractor being transported somewhere on a flat wagon. Into my spare parts bins then, sort out 3 gears, one to mount on the axle, one to mount on the tractor, and one intermediate gear. I had to drill out the gear boss on one of the gears because I’d chosen a gear set that was meant for 1.5mm gear shafts but of course the wagon axle was 2mm diameter. I glued that gear to the axle up against the flange side of the wagon wheel, and with the dremel, cut a slot in the deck of the wagon to allow a top mounted gear to mesh with the gear on the axle. A wee scrap of thin metal from something-or-other provided the material to build a small bracket & axle housing for the intermediate gear, so that just left the fixing of the final gear to the rear axle on the tractor.   Initially  (as can be seen in the first tractor pix), I used a small gear on the tractor axle as on the prototype (if there is one) there would be pretty low gearing at that point.  I had second thoughts on this later because in the model version the gears would be driving the tractor and not vice versa, so a small gear would make the tractor wheels turn very fast, and the tractor axle and “bearings” weren’t really made to withstand that sort of thing.


With the tractor suitably mounted on a pile of old sleepers (ostensibly to give the tractor some height so the Preiserling driver can see, but in reality to make some space for the gear train) the tractor wheels turned quite nicely when the wagon was rolled along the rails.

With regard to the motive power for the Box Wagon destined to propel the thing, I originally had in mind an old 3092 or 3087 chassis that I could cut down and use as a powered chassis to hide inside the wagon, but at the time I didn’t have such a thing in my scrap bin so started looking for other ideas.  I did have a motor with worm drive ex a “Hornby” Thomas the tank engine, and considered ways of hooking that up to a gear train to propel the box wagon, then of course another one of those make-it-up-as-you-go-along thoughts struck me:   Why am I contemplating building gear trains and messing around with the box wagon when I’ve already done that with the tractor – maybe I can fit the Hornby motor on to the Tractor?  I looked at fixing the motor out of sight under the chassis of the tractor wagon etc, but clearances were going to be tight and leave no room for a pick-up shoe, so I decided to mount it straight on top of the flat deck of the wagon (suitably disguised) and sit the tractor up on top of it.  This added further to the height of the tractor, possibly making it look a bit odd, but it did all seem to fit (with some re-building of the just completed gear train), and the worm drive is introduced between the tractor and the intermediate gear.


Tested with some DC on clip leads, it seemed to waltz along the rails OK, so the next job was rig up a pick-up shoe mount (an old brass Nut with a drilled & tapped flat plate glued to the underside), screw the pick-up shoe on, and mount a decoder.  I opted to use a lokpilot micro decoder which fitted on the underside of the chassis without problem, reasoning that the tractor would never be called upon to pull more than 1 or 2 wagons at the most and it should be pretty light on current draw.  Rather than standard “headlights”, I fitted a small strobe light front and rear, as I reasoned that the tractor certainly wouldn’t be allowed on the main line and would spend all its days on an industrial siding or pottering around a small yard, and I wound the decoder max speed down to about  “Preiserling on a bicycle” sort of velocity. 


Because the beasty has only 4 closely spaced wheels I decided not to fit traction tyres to the driving wheels as I thought I would have trouble with intermittent power return, and this was borne out by some pretty erratic running during the first real “on road” run on my test track.  A bit of weight underneath improved the power return problem but in spite of this weight I found that I also had a considerable problem on curves with it losing traction and slipping, so figured that with the added weight on board, I might be able to get away with one rubber tyre after all.  Wheels off, cut a groove in one of the driving wheels, slip on a rubber tyre.  Result: - much better. No more slippage, and power return ok on clean track at least. 

   Axle with new gear 


Slot cut in chassis for gear wheel

          Gear mounted - sleepers ready for tractor

So now, what am I going to do with it?  For the moment it can potter about the branch line station, but what I have in mind is a small industry/factory with a single siding where a branch line loco can drop a wagon off.  The tractor will then couple up to the wagon, pull it back to the end of the siding then run it forward into the back of the factory or whatever.  Probably all to be done automagically via the “shuttle train” feature on the central station to add a bit of “watchability” (if that’s a word) to the layout.


I’m not at all sure I won’t have a problem with the Preiserlings protesting about the rough working conditions and the safety of the unguarded gears on the tractor, but I’m afraid they’ll just have to put up with it – in these troubled economic times they can’t afford to be choosy about employment and I’ve got the upper hand for the moment. At least I’ve given them some handrails to prevent them falling off the edge, and an access ladder or 2 to boot, just to show I’m not totally mercenary.


Tractor Mounted (1st attempt)




ET87 - Staying on the rails no matter what.

The train is driven from the centre car, so no matter which way it is running there is one coach being pulled and one coach being pushed.  The "pushed" coach, whichever one it happens to be at the time, is the one that has the de-railing problem.  The train set has simple couplings/drawbars in the usual position on the underside of the chassis, modified by a sliding arrangement to allow the inter-coach gap to change when running through curves.  There is also another connection about half way up the coach (vertically) and this one is a modification of the normal close coupler/guide arrangement which controls the actual adjustment of the coach spacing on sharp curves. (It also carries the wires for the interior lighting etc.)   Because of the "push" factor,  I presumed that the very small gap between the coaches was a contributing factor of the de-railing problem, and this was confirmed by running the train very slowly through an S curve and gently pulling the leading coach ahead of the motor coach to keep the coupling taught.  Immediate success and the train negotiated the curve without problem. 

Unfortunately, nothing in life is that simple and it wasn't possible to defeat the sliding arrangement of the lower draw bar altogether to prevent the coaches closing up too much, because on straight sections of track the job of the upper "close coupler" is to pull the coaches together as much as possible - closer than the necessary modification to the lower coupler would permit.  In the end I compromised by restricting the sliding action of the lower coupling by as much as the upper coupling would allow, and this helped a little, but the real root cause wasn't revealed until the bodies, lighting, and interior were removed.


I had an interesting problem a while back with a Märklin 37266 ET87 (ET501-506) that would come off the rails in sharp radius curves, a most definite and every-time problem especially in 360mm (R1) radius "S" curves.  After much diagnosis I cured it with some modifications and wondered at the time whether I was over looking something obvious.  Later when I was sent a 2nd, and then a 3rd unit with the same problem  my original thoughts were confirmed - that it was a design problem, and there was no way it was ever tested on a layout with tight "reverse" curves.  All the units I've seen with this problem have been the 37266 version, but  I presume the very similar 37265 would suffer the same problem, the difference merely being in the paint job.  

A pity, because in other respects it's a beautiful looking train and runs smoothly and quietly.



The modifications to one of the lower couplings to hold the coaches a little further apart: On the left there's a little hot melt glue been set into the slot on the coupling to limit the inward travel when pushed.  Similarly, on the right I've glued a stopper to prevent the "forked" end of the coupling from being pushed in too far.

At each end of the upper coupling shaft is the Y shaped piece that fits into the coupling guide and which should swing to either side as necessitated by the curve of the track.  When the train is travelling through an "S" curve there is a point at which one coach is on the "left" curve, the other is on the "right" curve, and the coaches are pretty much travelling on parallel paths. At this point the upper coupling is at maximum angle between the front of the motor coach and the rear of the "pushed" coach, and if the coaches are too close the coupling jambs up against the side of the guide block and de-rails the lighter leading coach by preventing it turning around its curve.

The photo below depicts the problem and one can easily see how it would have been before I re-profiled the outer edges of that upper guide block.   The guide block was removed from the chassis to do that job, but care was needed because, as one can see, there's not a lot of material left between the outer corner and the slot in the guide block.  In actual fact, to get the necessary clearance I had to compromise a little and take a wee shaving off the side of the upper coupling as well.  

A wee bit of lubricant on the outer edge of the coupling guide block was applied to allow the coupling to slip easily along the side, and the  job was done.  

Well, almost done:  At the re-assembly stage it was remembered that the coach interior (seats etc) rests on top of that guide block and I was worried that if for some reason the coupling should ride a little high, then the top of the coupling would rub against the interior fitting (instead of the guide block) and I'd be back where I started.  Luckily, that part of the coach interior is invisible from the outside once the body is on, so a wee shaving off the bottom of the interior at that end bought some insurance and peace of mind.


Upper coupling jamming against side of guide block when swinging on tight S curves. The out side of the guide block was tapered off and rounded on the corners to allow room for the coupling to swing fully

















































The key element in almost any braking circuit arises from the fact that most modern digital decoders are (or can be) configured to interpret a negative direct current presence on the track as a signal to slow down and stop at the decoder's pre-set deceleration rate.  Some years back, Märklin introduced a braking module (72441) which applied this DC (actually rectified and smoothed track power) to a section of track in front of a red signal.  The difficulty (there is always at least one) is that this Direct Current potential  MUST be kept away from the normal power the digital controller puts out at all times, otherwise the output stage of your controller can be damaged, and if repeated too often it will eventually cause the controller to curl up and die.  This applies to any digital controller, and the short circuit will occur every time a pick-up shoe crosses the boundary between the 2 zones. 

Märklin resolved this problem by having a “transition section” between the normal track power and the DC-supplied braking section, the power in the transition section being a half wave rectified version of the track power.  This reduces the current consumption of the “short circuit” when the pickup shoe crosses the boundary and the controller is able to handle this ok.  I found however, that some locos have trouble with this arrangement and experience a significant jerk as they cross the track power/transition section boundary.      

Note the stop section of track just before the signal which is to prevent an overrun if the train has its deceleration rate set such that it doesn’t come to a complete halt within the braking section.      Note also that the negative DC is derived from the track power in this instance, with a single diode for the transition section, and a capacitor added to provide smooth DC for the braking section.  While not being a “problem” as such, this arrangement does mean that trains stopped in the braking section will still consume some power from the controller.

The arrangement is shown below:


Braking at Signals


Braking (“Slow Down”) sections for your signals.

A recognition of the horrendous injuries suffered by Preiserling Passengers due to trains coming to abrupt halts at signals, and a discussion of some methods for prevention of same.

Probably everyone in the Märklin fraternity is familiar with the old Märklin Signals and how to hook them up to automatically control trains, but since the digital age began there has been myriad ACC claims from Preiserling Passengers suffering dislocated necks as a result of trains jerking to a sudden stop, and one recorded incident in which an unlucky Preiserling’s head actually fell off.  The latter in particular, has caused the coroner to recommend that attention be given to the various methods of bringing trains to a more gradual stop at red signals.

First then, why does it happen?   Well the signals themselves are really just a bi-stable (latching) relay with 2 or 3 sets of contacts on them – one set to change the colour of the light the myopic Preiserling train drivers are supposed to take notice of, one set to cut off power to the track, and usually a 3rd set that can be used to cut off power to the overhead catenary where applicable.   When we used to run analogue trains (surely you don’t, not in this day and age?) the problem was considerably less pronounced because the old Märklin open frame A.C motors tend to “run on”  when power is disconnected.  (I had an old 3047 BR44, a beautiful free runner, which would on occasion coast right through 4 dead sections of M track at one particular signal).  Alas, no more – when converted to digital, such locos receive a permanent magnet in place of the old field coils, and the magnetic field causes the motor to come to a dead stop virtually instantaneously, thus the cause of the hideous Preiserling injuries we are concerned with.

If one has moved into the truly modern age and has a really clever computer (that knows which train is where) controlling the layout, then this problem doesn’t really exist because the computer can send a command to the loco in question telling it to decelerate gradually and stop. Under those circumstances one doesn’t need the signals to do anything at all apart from look pretty, but for plebs such as I who like to drive our own trains rather than watch a computer having all the fun, the problem of the sudden “dead stops” is something that must be dealt with.

First of all, I need to say that this isn’t new stuff, and although I might’ve made some practical improvements to basic circuits, I didn’t invent any of it.  I’m merely pointing out a few recognized methods of dealing with the problem and highlighting what I consider to be the best answer.  In the diagrams and explanatories that follow I’ve shown the signals being operated directly by means of track contact sections ("analogue fashion" if you like), because it’s simpler to explain that way.  The principal of the circuits is exactly the same if operating them digitally with S88s and accessory decoders, but some detail will need to be changed.


Pix 1 shows the standard Märklin signal. “S” is the signal relay (or solenoid) itself and “S1”, “S2” etc. are the contacts the solenoid operates.  All contacts are drawn with the relay in the re-set  condition (signal green).  The relay solenoid has 2 coils, one to “set” and the other to “re-set” the relay.  (A bi-stable relay)


In my book the best way is to avoid the boundary-crossing pick-up shoe altogether.  This can be done by having the train well inside the braking section before instantaneously switching the power from normal track feed to “braking DC”.  This means that a pickup shoe never actually crosses a live boundary between track zones.   It does require an extra (braking) relay to achieve this, but I think the results justify the additional wiring etc.  The braking section can be as long as the track configuration allows (ideal is a whole block between signals) but the train doesn’t commence braking until it hits the “trigger” (a contact section on the track at an appropriate distance before the signal).

Contact sections (isolated outer rail) are better than the “switch track” versions in this iteration because a contact section will operate at the right moment even with a train pushing wagons (i.e. when the pickup shoe is at the rear of the train.)  One disadvantage of the contact sections for detection is that a stopped train will maintain current flow in the connected relay/solenoid, but if one uses devices with sufficiently high coil resistance then this is of no consequence.

In the diagram below I’ve shown the negative DC derived from a separate supply to save too much drain on the controller track supply, but it could also be derived from the same supply as used to operate the relays (shown as dashed line) if using DC for that purpose.

Note the relay contacts S3:  This contact is to prepare a circuit for the braking relay when the signal is red, (or to put it conversely, to prevent the braking relay from operating if the signal is green).    The braking relay applies the negative DC to the “slow down” section and also cuts power off completely in an “isolation section” behind the train.  The prime purpose of the latter is to ensure that any trailing pick-up shoes on a train (for lighting etc.) can’t create a connection between the negative DC in the braking section and the normal track power, but it could also serve to prevent a rear end shunt of a train already stopped at a signal by a second train which has taken upon itself to run amok (as these things are wont to do on occasion).   As mentioned before though, the best mitigation against both these events is to make the braking section as long as possible, ideally the whole length of the signal block. 

The diagram of it looks like this:

Another way of avoiding the short circuit is to do away with the transition section but wire a 12v 21w car brake light bulb (ballast resistor) in series with the power feed to the braking section.  The bulb has minimal resistance when “cold” but heats almost instantaneously when a pickup shoe crosses the boundary. When hot, the bulb has a high enough resistance to prevent the short circuit.  This is a nice simple circuit and I have experimented along these lines with reasonable success, but there are some drawbacks regarding what happens if a pickup shoe near the back of the train crosses the boundary when the loco is already inside the braking section.  

Note that the negative DC could be just as easily, and is probably better, derived from a separate DC power supply as long as the positive DC is at track ground potential. 


Having built block signalling on my own layout to this principle, and also done the same with the Auckland Märklin Club's modular layout, I can vouch that this does work well, and since installation, the Preiserling complaints have died away almost completely.  They must love it.