Bloodnok

They are to minimise RF interference generated by the motor and transmitted via the track as an antenna. That's a DC specific issue. If you are using DCC, there's no benefit to them, and only the potential downside if they fail. So I removed the one from my stock NDM while I was in there, and never fitted one to the NDM I motorised from Hornby spares.

As delivered, the 2021 Hornby APT wouldn't climb out of my storage yard. The single NDMs sold out faster than I could get hold of one. There's one critical piece missing from the kit of spare parts required to equip the dummy NDM with a motor. Here's how I solved this problem:

Let's take a 37 or 47, and three coaches. They want £239-£249 RRP for the loco at the moment, and even recycled Mk1 tooling from 20+ years ago is now £65 a coach. That's £434-£444. However, that has un-lit coaches and no driving cab on the back. If we compare to a loco+3 rake with a rear end cab with head/tail lights and lit interiors throughout, that'd be a fairer comparison. Current RRP for Mk2Fs is £99 each, and the DBSO is £149. All of a sudden the comparison price is £586-£596. That makes the 4-TEP not uniquely high priced in Bachmann's lineup -- it is in line with Bachmann's other recent releases. We can debate whether those are also too high, but that just tells me Bachmann have higher overhead costs than say, Accurascale/IRM, KR Models, or even Auscision to point well into left-field -- all of whom have 4-car multiple units available for somewhere between half and two-thirds of the cost. Hell, even the upcoming retooled Hornby 4-VEP looks like a pretty good deal now. I'm kinda tempted by that Southern one. I travelled on that unit quite a few times.

It actually makes the revised Hornby 4-VEP look reasonably priced. I *might* have been tempted by a 3-CEP ... if they'd done one in South Eastern condition. Still priced out of my reach though.

The four-way ABC PCBs are now available. Order PCBs from PCBWay here: https://www.pcbway.com/project/shareproject/DMR_ABC_4_DCC_Asymmetric_Braking_controller_4_channel_9d298a48.html PCBWay referral code (which will give you $5 free credit, enough for 10 boards): https://pcbway.com/g/4z4dro You can also download the Gerber files there, or get them on the MERG knowledgebase if you are a MERG member. How do you build it? 20 1n5401 diodes (can be any "1n540<something>" really, any reverse voltage is way higher than what we need) 4 x SPST relays -- there are two common footprints on the board so a variety of relays are supported. If you want to drive it from 12V, get 12V relays. If you want to drive it from 5V, get 5V relays. 4x 1n4001 diodes 4x 5mm pitch two-way screw terminals ... and a control connector. Can be 2x3 pin (typical for IDC ribbon style connectors), or 1x6 pin (more useful for other things). How do you use it? Isolate the right rail for the entire block. Install it in the track feed for this block where you want your trains to optionally stop, e.g. approaching a signal. Top connection of each four to the track, lower one to the bus wire. You can use it as a power distribution board if you want by adding jumpers into the holes provided, this way you don't need to wire four separate wires back to the bus. Enable your DCC Decoders ABC braking feature. When the relay is not powered, the train will stop. When the relay is powered, the train will not stop. For reliable stopping, put a detector a set distance from the stop point, and turn ABC on when this is hit. Enable constant distance ABC braking in your decoders. Tune each loco using the constant distance features to stop in the right location. Decoder brand does matter -- some are better than others. I recommend Zimo decoders, these are the best I've found. ESU are OK provided you create a long enough braking section -- nothing less than two yards from the stopping point will do, and you may find that if you want to stop from an unrealistically high speed (above real travel speed for typical trains), three yards is better. Please avoid DCC Concepts decoders -- these are the only ones I've encountered where the loco won't stop if you turn it around (it heads into the braking zone "B end first").

I did contact the seller, and in this instance they sent me a refund ... without asking for the model back. I'm sure there's lots to read into that. A lot of repairs sit in the "easily possible but not commercially viable" zone, particularly when you factor in paying someone to look into an issue, and the unknowns about how long diagnosis takes. People can do lots of things that businesses can't.

A DCC controller will shut down when it detects a short, yes. Some are a permanent shutdown, others will intermittently retry to re-establish power. That's a surprise -- DCC decoders are very well protected from track side shorts -- as you've noted, these happen quite frequently no matter what the layout control system is. You only need something with pickups to run the wrong way into a live-frog point and it'll happen. Where DCC chips are known to be sensitive is accidental bridging of track power to motor outputs. Consider a split frame loco where the motor contacts are directly connected to each side of the chassis. Let's say you DCC fit this by separating the pickups from the chassis, and wiring pickups to the track connections of the chip, and the chassis to the motor connections of the chip. Done like this, It's ridiculously easy for a wheel to touch the chassis from underneath, and that kind of short (which would go unnoticed on DC as both would have been electrically connected anyway) is a big problem for DCC decoders. Honestly though, leaving any part of a loco (or coach) frame live to anything (track or motor contacts) is asking for trouble eventually regardless of how it is controlled, and something to be strongly discouraged. The best way of wiring that loco is to also isolate the motor from the chassis, and wire the motor separately, not using the chassis for anything. Far slower than a DCC controller or booster will do so. And it's normal for a DC controller to not completely remove power when there's a short, just limit the short current to a low value. I've managed to harm a DC loco. It stalled, and sat (still powered) without me noticing. The loco never ran smooth again after that. I presume I'd cooked one or two of the motor poles. Not enough to make it not run at all, but enough to introduce a noticeable stutter at low speed. Depends what you mean by "mix the systems". If you mean run a DCC chipped loco on a DC layout, that's fine. DCC chips are designed to allow running on DC. If you mean have both DCC and DC power on the same layout at the same time ... that can be done but needs careful management to ensure the supplies stay apart. Things like a coach with pickups rolling over a gap between both systems would be bad news. That's definitely the best way of building all metal vehicles -- only the bits you need to be live are live, and everything else is electrically dead. Also, paint or cover internal metal surfaces to minimise accidental contact. And for wiring, if it's loose, make sure it's insulated. If it can't be insulated, make sure both ends are secured so it can't go anywhere. That should be the advice for all control systems though, not just for DCC. Re-phrase that as "RTR plastic-bodied locos with a plug are more suited to DCC than poorly built metal kits" and I think you'll get universal agreement. It's certainly possible to build a metal kit well, such that it works brilliantly on any control system. But it's also possible to mess one up such that it causes problems on every control system. While some people consider DC tolerating more faults than DCC does as a virtue, that's not a universal opinion -- the alternative view is that it encourages continued poor design, workmanship and maintenance that continue to produce electrical faults...

No! That's the fault of the person who installed that previous decoder. When I convert a pre-DCC-ready model to DCC, I don't hard-wire a decoder, I fit a socket. That way a decoder can easily be removed later, whether that's being done to upgrade the loco to a new decoder, or to return it to DC (using a standard blanking plug), e.g. when it moves on to new ownership. And now it has a socket, so anyone wanting to re-convert it back to DCC can just plug a decoder into it again.

Interesting. I wonder if these will come up as spares for purchase for earlier models. (Yes, I'm aware that Hunt couplings exist, but they also bind on 2nd/3rd radius curves on this model). Capacitors were always hidden in the DTS and TBF vehicles, it was the TS/TU/TRSB/TF that had them in view.

The class 50 moodel is 20 years old, the next two at least 17 years old as I have one of each that I bought second hand ... in 2006. While they were definitely good models for the time, standards have moved on, and 8 pin sockets with basic head/tail lights aren't what the market is after any more. Combine that with a 2023 RRP over 4 times the launch RRP with little change in specification since, and other manufacturers have decided these classes are opportunities to target with better models at cheaper prices. For the 66, that "new" model is actually not new tooling -- they bought the tooling for a well-received earlier model from someone else and gave it a warm-over with parts-bin electronics. I can understand why the transition era might be seen as "traditional" ... but it's really not. Before it was "traditional" to model the mid 1960s, it was "traditional" to model the early to mid 1950s. Before that, it was "traditional" to model the big four. What is the popular era appears determined by a combination of what today's middle-aged hobby returnees (with disposable income) saw on the real railway when they were kids and what they had on their train sets when they were kids. In practice, it's roughly 30-40 years ago, but tends to stick on the early side of any specific railway related traumatic events. It appears things like grouping, nationalisation (and likely also privatisation) don't tend to count the way you'd think they would, but things like war, the Beeching axe, and the death of steam certainly did. The death of steam was such a strong event it has caused the "popular era" to stick in that place for a much longer time than previous events, which is why there's a perception that the early 1960s is the "traditional" period. But we are far enough past that we are reverting back to the 30-40 year rule now. There is a lot of pent up demand for era 7 and 8 models, stretching from the mid 1970s to the early 1990s, and we are currently seeing companies deliver to this market. It won't stay here forever -- it's going to keep moving forward, and we'll see it move on soon. I'm not sure what the next traumatic event in railway history will be that will cause a sticking point. End of loco hauled passenger trains perhaps? That has potential...

When they are hosted internally, the size of the image is known to the code generating the page. It is put into the HTML code. That way the browser can lay out the page with correct size 'holes' for the images, and then knows how far down to scroll to the current link, skipping over images that haven't loaded yet. When they are hosted externally, the site doesn't know how big they are going to be once displayed, so it can't tell the browser. The browser makes a number of requests first, but only up to some number. For pages with many external images (like this thread), that's not all of the images. And the browser will jump to where the new post is at that point, but as the images continue to load it then ends up in the wrong place.

The commercially available working Chinese signals, sure. There's no reason "working" has to mean crude and overscale though, it can be done. Some of the modelling I've seen on this site (including in this thread) is first class. That ... is a completely fair point. It's not so much the work to make the signals light up, it's to make them light up appropriately. If they light up, they draw attention to themselves. If they light up wrong that's arguably worse than not lighting up at all. Doesn't have to be operator input if you have the right data from the control system. IIRC you use a form of route-setting on this layout, which is one of the two things you need. The other requirement is to detect the movement of trains. Retro-fitting that would be a challenge, but possible. I was assuming this was the way you were going to go rather than suggesting a human signaller using a Mk1 eyeball as train position detection equipment... The simple three aspect signal would be for plain line with no junctions (and that's never the interesting bit of the railway people model!). There's a logic to how the signals work though, and the correct aspect can be derived from the route set and the number of blocks to the train ahead ... so this need not burden the operators. But it's far more difficult to integrate it into the control system of the layout after it's already built, so I understand why it would go that way. On the other hand, you also said: ... and here I totally get it. If you are looking at this project in terms of it being closer to the end than the beginning, it makes far more sense to do a 'light touch' to complete the scene, and look instead at future projects. If you are looking to finish the book on this layout with some exhibitions and move on to other things, starting a whole new chapter for working signalling is probably the wrong move as it will make the book so much larger...

And knowing the level of detail you have put into this layout so far, there's exactly zero percent chance these are unlit, scenery-only signals. Looking forward to finding out how you have chosen to operate them :-)

I bet that's not a consideration most modellers have considered -- protecting the arriving train from a conflicting movement *from the headshunt*. Okay, so with the track plan most modellers have, a single yellow ground can be replaced with two red grounds, one for the route out of the yard, and one from the headshunt back to the yard. You only go to a yellow exit signal when that second one can't be provided for some reason -- which could be a spatial issue or a workflow issue. Presumably the signaller can at least pull off once for a series of shunt moves in and out of the headshunt, rather than having to clear the signal for each one?

<glances nervously at yard with headshunt and planned yellow ground signal> Is this a signalling misunderstanding with modellers, where they are missing something different they could do on the same track layout to avoid using one? Is it just that people typically think about signalling too late in the plan, and have selectively compressed out a potentially critical bit of track, which then means they then have to use a yellow shunt signal? Or are Headshunts and Private Lines really rare IRL, and are just over modelled, because that track plan happens to fit conveniently in a shelf sized space that real life isn't constrained by? Is there a better (but still shelf-baseboard friendly) way of arranging the track that people are missing? I'm also wondering if modellers are too allergic to stopping and reversing on a mainline, insisting on running straight into a siding or headshunt when a reverse move would work just as well...

Of course not. He'll just be after his cut to look the other way while it happens...

Has anyone else seen flickering lights on one of these? I picked up a 35-301 (37034) recently, fitted a shiny new Zimo MN330P22 decoder to it, and the marker lights and tail lights are constantly flickering. If I turn the headlights on, they are brighter in the direction of travel, but e.g. the tail lights at the front stay flickering when travelling forward, even though the lights are on. If I usee the "lights off no.1 end" / "lights off no.2 end" functions, the end does go dark. But that prevents having the headlight on. Anyone seen anything like this before? Edit: For anyone finding this in future and saying "I also have this problem" ... , setting CV124 bit 7 to ON will fix this.

Sorry for being quiet for a bit, but I got back into the country to: * 9 hours of jetlag * A car with a broken windscreen * A leaking tap * A dying server hard disk * Several unhappy desktop PCs Then two days later I fell ill with something I assume I picked up on the plane. That was a week of being out of it, and another week and a half afterwards to properly get over it. When I did try to use the layout, loads of stuff there turned out to be broken, too. I tried to set a DCC address in the NR class, and the DCC system completely failed to work And when I tried to put the lift-out bridge back in, it wouldn't go back in either. Temperatures weren't that different, but I'm guessing the humidity was different enough that the gap had closed up several millimetres. It was enough that I managed to dislodge two track ends (one on the bridge and one on the layout) while trying to persuade it back in. Oh, and my phone died when I tried to capture some footage of the NR class running, too...

How exactly is DCC relevant to this? Do you mean it should be triggered by the setting of points (feather lit when points are set a specific way) and movement of trains (returns to red when block section in rear of signal is occupied)? Or do you just want to set the signal aspect from the DCC controller, but not automate it to work with the movement of trains? Or ... is it just that the wiring diagram for it references 12V, and you don't have a source of 12V power handy and want to power it from the DCC track power? A suitable change in resistor value for the LEDs should allow that, if that's what you mean.

Sticks of that diameter, yes. But I've seen people build this kind of thing out of 2x2. If anyone wants to go that way, it's a good idea to make it independently supported.

Although if you make it solid, they'll lean on it. And if it's cantilevered out from the layout, that might tip the layout over.

I went to Expo EM recently, and there were no barriers at that show. Initially I found it a refreshing change -- you could really look things on layouts in detail, and not be kept at arms length away the whole time. But then I saw someone literally put their head in a layout. Like above the baseboard, below the lighting pelmet. I later also tried to film some trains with my phone, and was forever struggling to keep other viewers out of shot when watching the popular layouts - it would have been much easier with barriers. That wasn't even a 'second row' problem, it's the front row causing the issue. I'm now on 'team barrier' - it's fairer for everyone if that barrier exists. ... That brings up a good point. There should probably be something larger, possibly round or softish on the ends of those sticks to negate the 'pointy stick' effect, should someone accidentally fall onto one.

While everything you said is true, there are a *lot* of gates in the real world that are built and hung like that. e.g.: http://www.lancashiregates.co.uk/images/Galleries/5&6_bar/Irthing 5 Bar Iroko Gate.jpg

I think the issue is with the contrast between the lighting on the layout and the sky. The clear sky indicates there should be direct sunlight, but the model doesn't reflect that. An overcast sky would suit that image better. A photo with brighter lighting would suit a blue sky. The photos of D209 work much better with the blue sky -- there's defined shadows and light on the side of the locos.

Yes, they absolutely would know about Genesee & Wyoming. But they wouldn't know that one of those shortlines it operates is called Freightliner. Ah, now Freightliner *is* known about down there, because they got *actual* freighliner, not just G&W: