ejstubbs

In a similar vein, every issue of Meccano Magazine in PDF format:

OMG, nostalgia central!  I remember the "Spirit of America" one from April '68 (IIRC my brother set about building the model but I'm not sure he ever finished it).

I'm sure that I also remember (though I can't locate) an edition from the early 1970s which had plans for a wooden, rubber-band-powered submarine which I did build.

Pretty sure that I bought some of those MAP plans booklets, too.  Whether I ever built anything from them I rather doubt.  I did, though, build a KeilKraft Cadet balsawood & tissue glider, accidentally smashed it to matchwood with a football, bought another kit and built it up all over again (ah, happy memories of tautening the tissue skin in the steam from the kettle, and the heady whiff of dope...)

But what was going on with the September '72 edition?  There's a girl on the front cover!  (Actually I think I do remember that one as well.  Can't think why.)

Oh, and the advertisements for air guns...happy days...

And yes, if someone could be persuaded to put archive editions of RM online, er...that would probably destroy the market for second-hand copies on eBay!

If I chose not to the use the peco accessory switch and instead a micro switch. Would the same switch as the one that changes the point motor be used in one circuit or would the two be on separate circuits / need two separate switches?

 

Or as the post above states does it need to be a mechanical link?

They need to be separate switches.  As I said above, the point motor power circuit and the frog polarity switching circuit are completely separate electrically.  The microswitch is the mechanical linkage: it translates the movement of the point blades in to the switching of the frog polarity.  The Peco PL13 is just a microswitch built in to a chassis that fits together with the PL10 point motor: the PL13's microswitch is triggered by the movement of the actuator rod that moves the point blades.  If you use a separate, bog-standard microswitch of the type that you can buy from Maplin or RS, you just need to fit it in a way that gets the mechanical movement of the point motor or point blades to trigger the switch itself.  The point motor switch switches ~16V to power the point motor.  The frog polarity switch switches the correct "side" of the track power feed to the frog.  Different circuits, different power requirements and totally different jobs.

By the way, don't be fazed by the use of the term "microswitch": all it is is a physically small switch which needs only a small actuating force to make it switch a circuit.  They're used where the switch is required to be operated by a small mechanical movement rather than, say, someone's big fat finger.

The other very important reason why they have to be separate switches is that the switch that actuates the point motor and the switch that controls the frog polarity function in significantly different ways.  The frog polarity switch must be a "stay put" switch like your average light switch ie switch it one way and it stays switched that way until you switch it the other way.  If you try to use that kind of switch for the point motor then you'll end up with a burnt-out point motor.  The point motor switch must be momentary (sometimes called a "passing contact" switch)*, giving just a burst of power to one or the other of the point motor's solenoid coils.  If you use a switch like this to control the frog polarity then you will end up with a more or less permanently 'dead' frog, which will prevent you trains from running through the point.

Suzie's suggestion of future-proofing - by modifying the points but wiring the frog polarity control under the baseboard to use the point blade contact with the stock rails - might be the right way to go if you are still struggling to grasp what the microswitch achieves.  In fact I might even do something similar myself when I come to fixing my track down permanently.

(At the risk of confusing things: another approach is to use the point motor power pulse to also trigger an electrical device called a latching relay to control the frog polarity.  This is not a mechanical linkage to the point mechanism, but achieves the same result in a similar way - basically, by having a second solenoid motor hidden inside the relay enclosure.  However, if frog polarity control using a microswitch is beyond your grasp then you probably don't want to go down this path either.)

* I am ignoring stall point motors like Tortoise and Cobalt: these do use "stay put" type switches but they are still powered by a separate circuit from that powering the track.  They also have built-in switches to control frog-polarity and other functions eg LED control panel indicators.  They are also more expensive on a unit basis than bog standard solenoid point motors like Peco and Seep/Gaugemaster.

The Peco PL-13 switch can be used with the Hornby R8014 point motors in exactly the same way as it is used on the Peco PL-10 motors.

Thanks, I didn't know that.

To the OP: It might help to be aware that the wiring for the frog polarity and the wiring for the point motor are completely separate electrically.  They are usually linked mechanically so that the frog polarity switches at the same time as the point blades move - these have to be in sync to avoid short circuits.  But the actuating power for the point motor doesn't - in fact it really mustn't - go anywhere near the rails.

It would, in theory, be perfectly possible to operate manually an electrofrog point that has been modified for external frog polarity switching, so long as you had some kind of mechanical-to-electrical linkage (eg a microswitch operated from the point blade tie bar) to change the frog polarity at the same time that the 'hand of god' changes the points.  (I wouldn't be at all surprised if someone on RMWeb has a layout that works this way.)

If you do this method am I correct in thinking that a polarity switch is needed to be added into the circuit with the point motor and points switch itself  such as the Peco  PL13 (are there other options?) 

If you modify the point then you need some means of switching the polarity of the frog.  The PL13 is one way to do it - if you are using Peco point motors, since it is specifically designed to work with them.  Or you can use a point motor with a built-in switch, such as the Seep (aka Gaugemaster PM1) point motor.  Note that some people do not rate the reliability of the switch in the Seep point motor.  Alternatively you can add a microswitch yourself, or look in to using a latching relay.  Bottom line though: if you modify the point, you need something to switch the frog polarity, since you will have disabled the switching that is done automatically by the contact between the point blades and the stock rails.

If you read around this subject eg in other threads on RMWeb such as this one then you will find plenty of discussion on the pros and cons of the using Peco electrofrog points unmodified.  The two main issues seem to be:

• Poor electrical contact between the point blades and the stock rails, preventing the frog polarity from being switched properly.  In theory this can be avoided by ensuring that the contact areas are always kept scrupulously clean and free of debris eg loose ballast.
• Shorting between the other (non-contacting) point blade and its stock rail.  This can occur because both point blades will be at the same polarity as the frog, meaning that the non-contacting point blade is at the opposite polarity to its adjacent stock rail.  The actual short can then be created by a metal wheel passing along the stock rail and briefly contacting the point blade.  This is more likely with older rolling stock whose wheels do not meet the closer dimensional tolerances that more recent products achieve.  This is also more of a problem with DCC, since DCC controllers are more sensitive to momentary short circuits and tend to shut down temporarily to protect themselves.

FWIW my DC layout (which is only in the early stages of construction atm) currently uses unmodified electrofrog points and I have yet to experience any issues with them.  I am, however, planning to modify the points - and thus use external frog polarity switching - when I commit to fixing the track down permanently.

I would suggest that if you're not confident with electrics, if you are using rolling stock produced within the last dozen or so years, if you are not using DCC (which AFAICS you aren't) and if you are prepared to do the necessary maintenance ie keeping the points clean, then you might be better off not trying to make external frog polarity switching work.  On the other hand, as others have said, if you don't do it now then it's not easy to retrofit it once the layout is built.  Your call...

Thanks everyone for the suggestions.  I think I'm going to go with the pluggable terminal strips, and automotive multi-core cable between the strips on intermediate boards so as to keep it as tidy as possible.

Something not mentioned in the OP: is this wiring for DC or DCC?

It's DC.

Occasional use, not regular dismantling and re-assembly.

I'm trying to decide on plugs and sockets to use for distributing power and point control wiring between baseboards.  An obvious choice would seem to be full-sized DIN plugs and sockets.  Can they handle the sort of current that solenoid point motors draw, albeit only fleetingly?

I think I've also read of people using computer connectors such as RS-232 25-pin or VGA 15-pin, but to me the pins on those look much less likely to be able to carry the sort of currents that model railway equipment can sometimes draw.  Am I underestimating their usefulness?

My layout is being built with Streamline electrofrog short turnouts (with one exception), with Peco Setrack for everything else.  See photo:

The track separation in the station platform loop on the right is standard Streamline 50mm.  In the fiddle yard on the left the spacing is Setrack 67mm - the main reason being that it allows more room for 'fiddling' with my pudgy digits.  (The one solitary Setrack turnout is the curved point at the top left - but that could still have given closer to Streamline track separation if I'd wanted, by leaving out the short straight between the curved point and the 11.25° curve which leads on to the straight on the left-hand edge of the photo.)

The double track curve near the camera has a track separation of ~60mm.  This is sufficient to allow my longest bogie stock - LMS 57ft coaches - to pass, see below, taken at about the tightest point on the curve:

As it happens, I don't envisage running trains simultaneously on both tracks, but I could if I wanted to.

I have run stock on all  the continuous run paths on the layout as well as the majority of the sidings.  Trains ranged from Duchess of Hamilton plus four coaches (!) through a 4F 0-6-0 with half a dozen or so wagons, a 2-6-4T with two coaches, and a couple of Jinties with a mish-mash of anything that came to hand.  All have run smoothly and without problems once the track had been de-bugged; both hauling and propelling, and for extended running-in periods.

I did have to cut a few sections of Setrack to work with the Streamline turnouts. The curves are cut down from 45° to 33°, so that when combined with the 12° angle of the turnouts they give 45°.  The remainder of the curves are laid with unadulterated Setrack.  The inner curve is second radius Setrack, the outer is third radius (apart from the final 22.5° curve leading in the the fiddle yard which is second radius).  There is also a fourth radius curve cut down from 22.5° to 12° just in front of the signal box.

There are numerous non-standard lengths of Setrack straight in there, to accommodate the difference between standard Setrack geometry and Streamline.  The trickiest bits were the ones nearest the camera, which have four and five sleepers respectively.  You need to be careful cutting Setrack because the sleeper configuration is not uniform along the length of a straight, so you have to make sure not to end up cutting in a place which leaves the sleepers disconnected.  Cutting short sections from long straights you also need to beware the fact that the Setrack sleeper base is not continuous along the full length of the straight - they use multiple sections of the sleeper base from the shorter straights.

Streamline turnouts don't like the fishplates fitted to Setrack.  Setrack fishplates have a rolled lip at the end which won't fit in the gap between the rail and the sleeper base on Streamline turnouts.  Unadulterated Setrack, in contrast, has gaping holes under the rail to accommodate the lip.  So where you are making connections from un-modified Setrack to a Streamline turnout you need to swap the Setrack fishplates for standard Peco code 100 rail joiners .  When doing this you will often find that the Setrack fishplate is ridiculously hard to remove: I once had to resort to filing the edge of a fishplate down until it was thin enough to tear apart!  Most of the time the Setrack fishplate will not be re-usable (and why would you want to anyway?)

With straight sections of Setrack especially, pulling on the fishplate while holding the track can sometimes lead to the whole track section flexing, or even the rail pulling out altogether.  Gripping the rail itself with pliers risks damaging the rail head.  Basically: care is required.

To accommodate a fishplate on a cut end of Setrack you need to trim away part or all of the moulded chairs on the first sleeper, and it also helps to skim off a thin sliver of the sleeper from immediately under the rail.

As others have said, the main advantage of using Setrack is that you don't have to worry about mucking up tight curves.  I could have laid all the straight sections using Streamline flexi, but as I had a fair amount of straight Setrack I decided to be a cheapskate in this instance.  A significant drawback of the Setrack straights is fairly obvious in the first photo: the sleeper base at the ends of the track sections just looks ugly.  Hopefully it won't look so bad once I get to the point of ballasting.  (I'm not inclined to be too precious about details like that, given that both Setrack and Sreamline are compromised anyway in terms of gauge, sleeper spacing etc.  At the end of the day the objective is to run trains.)

The layout was designed using AnyRail, and then "tweaked" as I did the trial laying of track.  The only bit that was significantly 'out' from the AnyRail plan was the short straight at the far end of the layout.  This was supposed to be an ST-202 79mm straight, but turned out to need to be quite a bit shorter.  I still don't know why.  The two sides of the layout do seem to be parallel.  I'm not stressing about it, though.

I've still got a few decisions to make about what is finally going to go where, in terms of buildings and sidings.  I have found the process of trial track laying very useful to see what works, and what should end up looking acceptable when the layout starts to get properly populated with stock and buildings.

HTH

in terms of incidents at crossings, sorry I don't agree that in EVERY case we should dismiss the cause as just a driver who has deliberately acted like a selfish idiot

Neither do I.  Add I don't think I suggested that (although I'll admit I did drift off in to a bit of a rant).  However, the tone of stevex59's post to which I was replying seemed very much to be suggesting the opposite ie EVERY case has to be assumed to be caused by a deficiency in the infrastructure and that's the only thing that needs to be fixed.

Quite frequently on RMWeb people wax nostalgically lyrical about the old public information films warning people of the danger of various foolhardy behaviours on or near railways (and elsewhere - remember "don't be an amber gambler"?)  I can't remember the last time I saw anything like that broadcast.  Admittedly a lot of people these days watch recorded TV and FFWD through the ads; I just offer it as an example of what would appear to be a fairly straightforward mechanism for reminding people about how they are would be wise to act in the face of everyday risks no longer being deployed for some reason.  Relying on enforcement alone - camera vans and the like - apart from being expensive and therefore almost never being able to achieve sufficient coverage to have a reliable deterrent effect (viz speed cameras) can also create an adversarial feeling ("us against The Man'") which almost encourages the more 'rebellious' and dimwitted to to break the rules.

And the point is.....improvements...simples

Such as?

Improvements in driver behaviour generally would make the whole world a more pleasant and safer place, so I'd be strongly in favour of those.  Actually I'd go further and say improvements in the behaviour of all road users (including pedestrians and users of pedalled and powered two-wheel vehicles) would contribute to greater overall wellbeing.

These last few days I have personally witnessed:

• A taxi speeding up to ~40mph in a 30 limit order to run an amber traffic light - and predictably failing, passing it when the red light had been on for several seconds.  (You've never heard mechanical pain until you've heard an Edinburgh black cab being thrashed like that!)
• A private car being driven through a traffic-light controlled pedestrian crossing with the lights at red and other traffic stopped, as if they hadn't even noticed that it was there.

Do you fix all those sorts of behaviours by "improvements"?  Why not be honest and admit that some people really are too stupid and dangerous to others to be allowed to drive?

In my opinion there is far too much acceptance that road users are unable to do anything other than act like selfish idiots, so everyone and everything else has to adapt to allow them to continue doing so.  I can't readily think of any other transport system which is allowed to operate that way.

I believe there is a problem with the chassis detail with one of the non-corridor kits, and I think I remember reading that the brake kit comes with the roof for the composite, perhaps because the mould for the brake's roof is u/s.

Fairly certain I read ask this on here, so worth having a search for it.

I ended up buying a pair of the original Airfix/GMR ones on eBay. They seem to be fairly readily available; Rails were selling off quite a few when I was looking a couple of months back.

Edit: this is the earlier thread I remembered on RMweb on this topic: http://www.rmweb.co.uk/community/index.php?/topic/64375-Dapol-ex-lms-non-corridor-lavatory-coaches-a-review-of-sorts/

Just as a matter of interest (I've not yet seen the reincarnated FS in action) is the whistle really that weedy or did the BBC tone it down to avoid startling viewers? 

 

John

http://www.rmweb.co.uk/community/index.php?/topic/103992-flying-scotsman-back-on-the-mainline-2016/page-22&do=findComment&comment=2218576

My baseboard supplier (building them myself would have taken far too long and likely given a poor result!) recommended grey primer.

Also, have a look at some of the answers on this thread: http://www.rmweb.co.uk/community/index.php?/topic/107478-birch-pine-does-it-need-sealing-before-use/  There seems to be a consensus on that thread to use an oil based (or at least non-water based) medium, and to do both sides of the board to avoid warping.

In my search for materials I found that B&Q do "solvent based" grey undercoat for reasonable money.  If you prefer to leave a plain wood finish then Wickes do a solvent based varnish - all the varnishes in B&Q seemed to be water based.  IMO the term "solvent based" seems a little vague.  After all, water will dissolve many things - but I think what they mean is not water based.  I doubt many of the "solvent based" products are actually oil based in the sense that people of your or my generation would think of it.

FWIW I haven't actually got round to painting my boards yet; I'm still in the final stages of making detail adjustments to the track layout.  I will take it all up again when that's complete, and paint/varnish before re-laying the track on closed cell foam underlay and getting stuck in to the full wiring.  What I have had running on it so far (including a Hornby 'Duchess of Hamilton' which is ridiculously large for my layout but which looks very handsome and was at an irresistibly good price second-hand in my local model shop) seems to be going very well, I'm happy to say.

If this is all good how do you do physically it?

Cut, dismantle?- do they come in two bits?? 

It very much depends on the make & vintage of each model.  The hook on Bachmann NEM couplings can just be popped out of the body of the coupling (and popped back in again if you choose to later).  I can't speak authoritatively re other manufacturer's NEM couplings as I've gone all-Bachmann (although I am in the process of experimenting with Kadees).  I know that some older tension-lock couplings - especially the big ones on old Hornby stock - have the hook attached using a loose rivet which you can drill out, or just monkey to pieces using a combination of wire cutters, pliers and stupidity (that'd be me, then).

I re-watched "4472m - Flying Scotsman" last night (the BBC film of the 1968 non-stop KGX to EDB run - available on iPlayer).  The whistle definitely sounded a bit feeble at times in that.

Edit to add: The film also includes some rather fine footage which can only have been shot from a helicopter (eg around about the 16 minute mark - example screenshot here).

Just sayin', like...

Kadee sell a drill and tap set that includes a through-hole drill and a drill for tapping (#246). The sizes are Imperial, 2-56, but suitable screws can be obtained (I just did a Google search). I got the type that uses an Allen key, much easier to handle.

 

http://www.ebay.co.uk/itm/361384057282?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

 

Screws from

http://www.accu.co.uk/en/

According to this table, 2-56 is a coarse thread (56 tpi) #2 diameter screw.  #2 diameter is 2.18mm major diameter, so an M2 will fit - and I can confirm that an M2 is a very slightly loose fit through the mounting hole in the Kadee draft box.  It looks like an M2.2 would be a snug fit, but I don't have one to hand to try.

Are these the nylon screws in the pack. What works is an 8BA tap if you already have one. It shouldn't work, but it does!

 

Edit to add.

Try it out on a thick piece of Plastikard.

8BA is 2.2mm major diameter, 59.17tpi, so the diameter is OK but the tpi is smidge finer than the 2-56.  So long as the tapped material is reasonably soft - eg polystyrene or white metal - then it should be OK, assuming that both the screw and the receiving material will 'give' a bit.  If the receiving material is significantly harder than the nylon screw then I expect you'd end up trashing the thread on the screw 

EJ Stubbs, Self-Appointed President of the Society for the Prevention of Cruelty to Screw Threads and Heads (SPCSTH)

So does the #206 insulated gauge.

Talk about jumping to conclusions! In fact you seem to be jumping so high you must be in a low Earth orbit by now!

 

...

 

It is also very poor Level Crossing design where all four barriers lower simultaneously - increases the risk of trapping vehicles as happened here. If the exit barriers had lowered just a few seconds after the the entry barriers the accident would very likely not have happened.

I'm still firmly grounded on planet Earth, thank you.

There is a link to the Wikipedia account of the incident in one of the comments under the video on YouTube.  The Wikipedia article states:

"The driver drove through a rail crossing near the Studénka train station. The crossing was protected by signal lighting and gates, and the driver entered the crossing when the red lights accompanied by acoustic signal were on for at least 30 seconds. The driver drove into the crossing despite warnings and as the gates went down, he stopped the truck on the tracks.^[4] Instead of ramming through the gate, which is constructed so as to allow even a motorbike to penetrate the barrier and automatically sends a stop signal to all approaching trains if breached, the driver moved the truck only a few metres forward in order to avoid a direct impact to its cab, with its trailer carrying aluminium sheets remaining on the track."

The Czech news article in linked from the Wikipedia article states that the operation of the warning lights was confirmed by the preliminary investigation.

You seem to have jumped to the conclusion that my comment was based purely on viewing the video.

I hope they throw every book they can find at that homicidally dangerous lorry driver.

(Seeing as how he doesn't seem to have cared much about anyone but himself, once he'd broken the law by passing the flashing lights why didn't he just drive through the barrier to get off the crossing?  Or is he one of those clueless idiots who thinks that "the train will stop"?)

I suspect that as you now have it, if point 2 is set towards point 3, and point 7 towards point 2, 2-7 will be live, which still isn't the way you want it.  The top rail of 2-7 will always be fed from the original feed on the left, and the lower rail will be fed from the new feed to the right whenever point 7 is not set for the crossover.

I agree with this.  In fact I was going to raise it earlier, but I got confused about halsey is actually trying to achieve in his branch terminus: is it the ability to operate two (or even more) locos in that area, or is it making sure that there won't be any power dead spots?  If the latter then the more power feeds the better (a la DCC).  If the former then KISS should apply - but then KISS to me would be mean full cab control (because that way you design out 'phantom' routes for power and you know which sections are powered and which aren't), and IIRC halsey has already said he doesn't want that.  Which of course is entirely his choice.

I habitually feed points from the heel end, it can save yards of wire, just connect + and - to the two outer rails of the four, it is exactly the same as feeding from the toe. 

That's not what I meant by powering from the heel end.  What I meant (and what halsey's trackplan in post #236) was powering from both rails of one road on the heel end of the point,  Doing it that way means that, in an insulfrog point, power to the toe end from that power feed will be interrupted by setting the point against the road to which power is being applied.  (And yes, I know that in halsey's track plan the goods yard can also be powered from the other power feed, provided points 2 and 3 are set appropriately, but that's by no means going to be the case with every track configuration.)

not sure I completely understand the wire saving comment but assume this is electrofrog related?

No, it's nothing to do with electro vs insul frogs.  David is simply pointing out that you can apply power at any location along the continuous (ie unswitched) rails which run through the each side of a point, even a good ways further along each road from the point in question - hence potentially saving wire if the point itself is a long way from the control panel.  Personally I wouldn't do it that way, since I would prefer to spend a bit more on wiring in order to make debugging easier at a later date - but again, each to his own, rule 1 applies etc etc.

(By the way: in case anyone takes exception to the tone of any of the above, please be assured that no offence is intended.  It's just that it's Friday, I still have a cold and I'm struggling to be as cheerful as I would normally try to be.  Sorry!)

Chris

Your diagram is correct, and it proves what I agreed with you on before: that to prevent 2-7 being live when point 2 is set against it, you need to operate points 6 and 7 together - just like they would have been in the prototype.

The single IRJ is to prevent 2-6 being live when point 2 is set against it.  As I said in my loooong post last night, ensuring that points 6 and 7 are always operated together will prevent this, too.  The single IRJ is just another way of achieving the same thing (I think - as I said, my brain was full of cold last night so if you want to draw that one out and prove me wrong, I'd be more than happy to retract the idea!)  Call it failsafe, if you like - particularly relevant if the points are being worked manually rather than by point motors.

Looking at point 7 from the bufferstop ie the toe end, the single IRJ should go on the left-hand of the two rails emerging from point 7's frog.  No additional switch would be required, so the control panel can remain unchanged.

I agree with you that a pair of IRJs between points 3 and 4 might be a good idea, together with moving the power feed for the goods yard to between point 4 and 5.  The main reason I think this would be a good idea is that it would avoid any unexpected and difficult-to-diagnose behaviour arising from power being fed to points from the heel end.  A possible side benefit is that it would allow the branch terminus to be run with more than one loco (although not simultaneously).  For example, a loco could be shunting in the goods yard, then isolated in a siding (or even the headshunt) while a passenger train arrives and runs round.  Then, while the passenger train is waiting to depart, the passenger loco could be isolated on its platform and the goods loco could finish shunting the yard and depart with its train.  Again, this would only need the new IRJs and one power feed moved; no additional switching would be required.

I'm going to go a bit Pierre de Fermat here and say that I had a beautifully expressed explanation all written up and ready to post, but then I pressed the wrong button and lost it all 

I'll try to re-create it over lunch - I am at work today, despite the cold and the lack of sleep...

think if point 6 is set to the crossover and point 7 is not, 2-7 will be live regardless of the setting of point 2....if 6 and 7 are either both normal or both set to crossover, you shouldn't have a problem

I agree with that, at least in respect of 2-7.  I can't remember whether halsey is working the points manually or by point motors.  If the latter then it would be straightforward to work 6&7 together from one switch - which would, as you say, be the way they would have been worked in real life.  If the points are being worked by hand then the operator would just have to remember to follow the rule.

If you do put IRJs on the crossover, I think you would need another feed to the right of point 7

I agree with that, too. With IRJs on both rails of the crossover then there would be no way to run a locomotive from 2-6 into the spur beyond 7 (or vice versa) without it - in other words, you couldn't run round!

I can't find a way (yet) of making both rails of the lower line live with point 2 set against it

Ignoring point 3 for a moment, if point 2 is set for the 2-7 road, and 7 is reversed then 2-6 would be live.  What happens is that the "upper" rail of 2-7 is always live (because it's continuous all the way to the bufferstop beyond point 7).  With point 7 reversed there is also a circuit from that "upper" rail out of the frog of point 7 towards point 6.  Since the corresponding rail on point 6 is continuous, that makes the "upper" rail in 2-6 live as well.  Now looking at point 3, if that is set towards the 3-6 road then the "lower" rail of 3-6 is also live (since the "lower"rail of point 2 is continuous).

One way to get round this would be to make sure that point 3 is never set to the 3-6 road when point 2 is set to the 2-7 road.  Following on from the answer to keeping 2-7 isolated, as previously described, you might think that the way to do this would be have points 2 and 3 operate together - but if you did that then you never be able to access the goods yard (oops).

In fact, though, operating points 6 and 7 together fixes this problem too, since you would (should) never leave point 7 reversed if point 2 is set to the 2-7 road.  Alternatively - or even additionally - you could put a single IRJ on the rail which emerges from the from of point 7 towards point 6.  Making it a single IRJ would avoid having to have the additional power feed to the toe of point 7.

if you put [iRJs] between 3 and 4 you would need the headshunt feed moved to between 4 and 5.

I agree with that, too.

In fact I'd go further, and say why not run the upper branch as cab control?  IRJs between points 3 and 4, with a separately switched power feed between points 4 and 5.  More IRJs between points 6 and 7, and after point on the 2-7 road, with a separately switched power feed to the toe of point 7.  That way at least nothing would be live when you didn't expect it to be.

Come to think, wasn't halsey planning to run the branch "one engine in steam"?  If so then all of this is largely moot...

MAJOR CAVEAT TO ALL OF THE ABOVE:

I have a stinking cold so my reasoning may well be faulty.  I only decided that I had to join in with this particular debate because the cold woke me up, and then couldn't get back to sleep because of thinking about the problem!

Oh, and another thing: I'd have drawn diagrams to illustrate what I'm saying in all the above, but (as anyone who's seen my post on the current Kadee couplings thread can attest) my drawing skills are minimal to non-existent.

And finally: sorry about the unnecessary blank lines in my quotes of Chimer's post.  I can't seem to make them go away

Right, now I've got all that off my chest it's time to take some more drugs and head back to bed!

The characteristics I see are that the original Kadee magnets both repel from the centre and attract at the edges.

The uncoupler magnets definitely have opposite poles at the two long edges.  That's easy to see by laying two of them side by side: one way round they will repel each other, the other way round they attract.  If you have a magnetic compass you can easily find out which is the north pole of the uncoupling magnet, and which is the south, according to which end of the compass needle each side attracts.

It's simply not possible to make a permanent magnet with one pole in the middle and the opposite pole at each end, which is what you seem to be suggesting, and what the diagram on your blog suggests.  If you could take one of the uncoupling magnets and magically split it lengthways, the newly-exposed faces of the resulting thin magnets will be opposite poles to the corresponding outer edges ie you'd end up with two thin magnets each with one north pole and one south.  That's how magnets work.  (What you suggest would actually be a physical marvel: effectively a magnetic monopole, which has never been observed experimentally).

Fundamentally, the trip pins need to move sideways, which means that the magnetic field needs to be at right angles to the direction of the track.  If the uncoupling magnet had a north pole at one and a south pole at the other then the magnetic field would run parallel to the track and the trip pin would not move.  This can easily be demonstrated by taking a single Kadee coupler and lowering it by hand towards an uncoupling magnet at right angles to the normal direction: the trip pin will not move, even when it is in contact with the surface of the magnet.

Since this can be demonstrated using a single coupler, it disproves the theory that some people put forward about induced magnetism in the trip pins causing them to repel.

It is worth keeping in mind how the stock uncoupling magnets work when experimenting with using small, discrete neodymium magnets in place of the ferric slabs that Kadee sell.