Roswell Mill

[burgundy]

It has been another of those "two steps forward, one step back" kind of sessions.  Does everyone's modelling go like this?

As the previous photos show, the wiring was in place, but was to be tested when the mechanical linkages that actuate the turnouts were installed. Today's ambition was install the linkages and to check the wiring.

Slightly surprisingly, the mechanical linkages now seem to work. Some of the "fettling" (the posh word for "fiddle till it fits") has been a bit Heath Robinson and may need a little more tweaking. In one case, I have lengthened the throw of a crank to compensate for the long linkage to the further end of the loop (one switch operates both turnouts). On the three way I have adjusted the position of one set of diverging rails to match the throw of the 3 position switch. Crude - and a bit brutal to Tony's carefully constructed 3 way turnout - but, again, it seems to work.

This happy progress came to a grinding halt when it came to the wiring. There is a short circuit (at least one). The one that I have tracked down so far is in the little headshunt at the bottom of the photo. The loop itself is OK, but this wretched short section of track (which is physically gapped from the rest of the layout)  has a short circuit in it somewhere. I have disconnected the wiring, I have unsoldered two of the PCB sleepers and replaced them (I have done quite a lot of unsoldering today) and I have inspected the remaining 4 PCB sleepers from all possible angles to see if the gapping is bridged in some way. I reached the point where I will go back tomorrow to see if I have been missing something obvious before I start excavating the last 4 PCB sleepers (all the other ones are wood and that can't be conductive can it?).

Any inspired suggestions would be much appreciated!    

Best wishes

Eric 

[ullypug]

Eric

I feel your pain, having had to chase down pcb shorts on Weston. Every sleeper for Clevedon was tested with a meter before soldering. It's amazing how small a sliver of copper still conducts electricity! It's probably the sleepers and a burr at the side. Ask me how I know!

Couple of other observations:

Are the headshunt rails lower left resting on the sleeper for the stub point causing a short through the point? I don't see a gap in the second sleeper of that point, but it may be the image.

Have you checked the switches, connectors etc? I had a short on Cheddar which turned out to be an errant strand of wire bridging two pins in a d plug connector.

Are the frogs causing a short? It's where + and - meet!

Good luck.

[Gordon A]

Are your sleepers double sided copper clad and pinned?

If you have only gapped the top then a pin at each side could create a short through the non-gapped underneath.

 

What I do some times is put connect the a wire to each rail from a resistance soldering then press the button looking for a puff of smoke.

Also listen to the transformer. If its tone does not change after a second or two, then the short is quite a meaty one. 

 

Gordon A

[burgundy]

Andrew, Gordon

Thank you for the helpful suggestions.

Perversely, the section that has the short is the section at the bottom left hand corner of the previous photo, and, being a stub point, there is an air gap between the headshunt and the diverging rails leading to the frog. It is just the headshunt, which is all of 9" long with only half a dozen PCB sleepers, which is causing the problem.

The sleepers are glued rather than pinned, so there should not be an issue with the gap being bridged from the underside of the PCB.  

I spent half an hour this evening unsoldering one side of the track to check for the insulation on any of the sleepers being bridged and it still showed a short. By process of elimination, the only thing left is the tiebar and so, for the next stage, I may get round to some serious unpicking, to dismantle and replace it.

Oh bother.

Best wishes

Eric  

[burgundy]

I think I have spent nearly 6 hours, spread over the week, trying to trace this particular short circuit. 

I unpicked the headshunt completely. I checked each sleeper. I checked the tiebar. I then reassembled it all and the short was still there.

The next step was to disconnect all the wiring (tagging it all as I went along). Still no joy.

Before  

After

Eventually, it slowly dawned on me that Tony had soldered thin brass strips under the stub switch in order to link the two ends of the turnout, which otherwise have no physical connection. Soldering the underside of the PCB sleepers to these strips potentially creates a bridge between the + and - sides, unless the underside of the sleeper is also gapped. Any connection between the top and bottom surface of the PCB would then create a short between the rails. My dodge, to secure the stub ends vertically by holding them down with a short length of brass tube, creates exactly the link between the top and bottom of the PCB.

This evening's job was to extract the two sleepers under the ends of the "switch" rails, gap them underneath and reinsert them. Hey presto and it seems to work. The other two turnouts seem fine, so I can only assume that they have been gapped.   

So finally, a happy ending. 

Best wishes 

Eric 

[ullypug]

Glad you found it.

[Crichel Down]

Eric

I feel your pain, having had to chase down pcb shorts on Weston. Every sleeper for Clevedon was tested with a meter before soldering. It's amazing how small a sliver of copper still conducts electricity! It's probably the sleepers and a burr at the side. Ask me how I know!

Couple of other observations:

Are the headshunt rails lower left resting on the sleeper for the stub point causing a short through the point? I don't see a gap in the second sleeper of that point, but it may be the image.

Have you checked the switches, connectors etc? I had a short on Cheddar which turned out to be an errant strand of wire bridging two pins in a d plug connector.

Are the frogs causing a short? It's where + and - meet!

Good luck.

 

I had a similar problem on "Crichel Down" way back in 1982, when it was being 'converted to P4' at the old MRC show in Central Hall, Westminster (all done under the merciless gaze of Joe Public!)

 

All the track had been laid using copper-clad sleepers, being careful to saw a gap just inside the running rail on both sides of the track, and yet a mysterious short circuit appeared, and was impossible to trace.  Eventually, a very small piece of metal swarf was found bridging one of these gaps in the copper clad.  Having cleared it out, the sleepers were promptly painted, taking care to brush paint generously into the insulating gaps in the copper-clad sleepers.  After that, there was no repetition of this problem.

 

As the French say,   Plus ça change, plus c’est le même chose!

[technohand]

Me'a culpa should have gapped the undersides of the sleepers when I soldered the reinforcing strips underneath. My only excuse is that no one had ever (as far as I know) built stub points in 3.5mm/1ft scale before. In my defence no shorts were evident when I handed the points over.

Glad its all turned out OK and hope Eric is still talking to me!

 

Tony

[burgundy]

It is a long time since the last progress report, as things have continued at rather less than a breakneck pace.  With the track complete and electrics working as intended, I have been trying to figure out what the surroundings should look like. 

The television screen frame for the front has been cut out, painted and clamped loosely in place. Some heavy lining paper has been run around the sides and back – although there are some wrinkles and I am tempted to do a second layer to leave fewer bumps to conceal. And that gets us to what the scenery around the track should look like.

I feel at a disadvantage with this layout, as I have fewer references from which to create a plausible, fictitious location. Firstly, it is in the US – and the daughter who lived out there and started me down this slippery slope is now back in the UK. Secondly, I suspect that the landscape, that I photographed when I visited, has changed quite considerably since the 1860s.  Nowadays, it is pretty heavily developed – but also still very wooded. Then it had only really been settled for about 40 years and I suspect that much of the land would have been cleared. In part this would have been driven by agriculture but also by the voracious appetite of log burning locos; anywhere near a railroad would have been stripped pretty quickly.

The raison d’être for the fictitious railroad is a real mill, which was constructed in the valley of a tributary of the Chattahoochee river and was waterpowered.  Surviving buildings are brick and there are references to brick built workers’ houses from an early stage of development.  The mill was incorporated in 1839 and so would have been a little over 20 years old by the time that the civil war broke out. It was important as a supplier of Confederate uniforms and therefore became a target for the Union.

My approach for trying to create a plausible scene (and my excuse for hastening slowly) has been to play around with some simple card-built structures. Ironically, Abraham Lincoln’s family homes are well documented and have card kits that are free to download. They are essentially log cabins and I have made the assumption (possibly rashly) that there was not a great deal of diversity of vernacular architecture in the US at this time and that a log cabin is pretty much a log cabin. There is also a wooden barn of German origin – but, again, as a simple wooden structure, how different can it be?

For brick buildings, I have tried two options.

One is a single storey building originating in Germany - currently placed at the rear of the scene. Although described as HO, it looks better when enlarged by 10% and I wonder whether it was not originally TT. It came as part of a glassworks but it has a nice atmosphere of a scruffy industrial building. The sharply pitched roof does not look right for an American building and may need to be removed if it gets used.

The other is a Scalescenes factory, which I downloaded some while ago for an entry to a competition entry by NWSEMGSAG* for the EMGS. Taking the downloads, I reduced the scale to 7/8ths (coincidentally 87%) to get something that was better proportioned for HO. I have also done a cut and shut on it, so that it looks less monumental in rural Georgia. One section is two storey and the other three storey, with a step of one bay between the two. The kit features some concrete lintels, which I will have to replace, as I don’t think concrete was in use as a structural material in the 1860s. I should be happy to be told that I am wrong.

There will be an unloading area and possibly stacks of baled raw cotton on the far side of the tracks. On the near side, there will be a log pile and a water tower to service the locos using the branch.

Currently, the buildings are simply plonked on the baseboard to see how they look, and I have not yet tried to create any contours, which might make the scene more interesting.

Any suggestions or inspiration would be very welcome!

Best wishes

Eric

*North West Somerset EM Gauge Society Area Group (actually covering most of the County that Used to Be Avon).

Edited July 28, 2019 by burgundy

[burgundy]

Cross posting of the message above to a US group that discusses American Civil War Railroads resulted in someone providing a link to the image below. 

This has been very helpful in visualising how the scene might look. The factory and its outbuildings actually sit in quite a narrow river valley, with the buildings overshadowed by the river bank. Presumably, once the dam had been built a short distance upstream, the location was judged to be reasonably safe from flood or washout. 

My version is not going to be anything like a close replica of this scene, but there will be a brick built mill with timber outbuildings and the critical addition of a railroad - which has been carelessly omitted from the prototype. 

Best wishes 

Eric 

[burgundy]

Another small step.

The main structure for the scene is now largely complete. The “mill” building is built from a Scalescenes factory design that I downloaded some years ago for another project. This time, it was printed to HO scale (by reducing to 87% on the printer) and then subjected to some rather brutal cutting and shutting to give it a more interesting profile in the space available. It is also considerably more modest in size than the image above suggests - but otherwise it would have completely overpowered a cameo!

It has a pitched roof, since it stands away at an angle from the back scene. I had been anguishing about what sort of roof a building like this might have and had assumed that the choice would be metal sheet, tar paper or shingles. However, I was advised by a contact on the American Civil War e-group that there was actually a slate quarry not far from Roswell, near Marietta. Since slates are more familiar, I went for that option. I have some grey craft paper, which I ran through the printer to print out a blank spreadsheet - with the cells set to the right size for the bit of the slate that shows. You then cut the “slates” off two rows at a time and stick them to the roof so that one strip lays across the top row of the strip below. The rows need to put the “gaps” between slates alternately to reflect the way in which slates are actually laid. It is a boring job, but it does give you some texture to the roof of the building, which tends to be the bit that you end up looking at!

The interior of the building was made up from some scenes printed off from old images of cotton mills. You do not see much through the windows, looking at an oblique angle, but it does give some impression of an interior.

Best wishes 

Eric 

[technohand]

Hi Eric. for some reason I've not been informed of updates. It's all looking good.

Tony

[burgundy]

A little undergrowth beginning to appear in the background to give an impression of progress.....

Best wishes 

Eric 

[burgundy]

The grass seems to have been growing quickly over the last week, as I now have all the area behind the tracks covered with nondescript greenery. 

What has not gone so well is that the background, which is wall paper lining paper, did not take kindly to getting a bit damp and has developed a bit of a ripple and a small crease, which is highlighted by the room lighting. Any suggestions for removing or concealing the crease would be gratefully received. 

There also needs to be something to fill the gap between the scrubby hillock and the timber cabin where the road reaches the backscene. The black and white illustration in the post of 13 August shows a much deeper river valley, so possibly more scrubby bushes? I assume that all decent sized trees would have been cut down in the preceding couple of decades for construction timber or for fuel and so all that is left is new growth.

The broad gauge track shows up quite well and I will now have to get down to "ballasting" - although I don't think US railroads of this vintage used much in the way of ballast. I shall also need to find a way to blend in the timber and copperclad sleepers  (and I know that they should probably be split log, half round but that is not going to happen!)

Best wishes 

Eric 

[burgundy]

Another small step forward. Seeing the shops full of Christmas lights, I decided to do something about the built in lighting for the cameo.

The LED strips that I had used originally were brilliant white, which gave a rather washed out "wet Sunday in Hastings" atmosphere.  Buying a set of warm white LEDs gives a much sunnier impression. This is how it looks with all the room lights turned out (hence the black surround, which is actually a sort of Confederate grey).

Best wishes 

Eric 

[burgundy]

With Seaford now completed, work has resumed on Roswell Mill.

The mind was focused by an imminent tour of inspection, which was scheduled for the March meeting of NWSEMGSAG (North West Somerset EM Gauge Society Area Group). However, the group decided to cancel the meeting as we collectively, and somewhat ahead of official guidance, felt that it was an undue risk. Nonetheless, it has prompted a burst of activity at Roswell Mill.

Much of the scenery has progressed although there is still a large blank space at the front left hand corner. Currently I can envisage a large pile of loco fuel (logs) there, but there probably needs to be something else to conceal the hole in the sky. Any suggestions gratefully received.

The blank space at the front edge of the layout, next to the factory, also needs something to fill in the gap.

There is a rather annoying crinkle in the backscene (just behind the largish tree), which I need to remove or conceal. Any suggestions welcome. I could simply add another tree, but I suspect that the landscape around any railroad yard was pretty barren and any substantial trees would long since have been turned into loco fuel.

“The General”, aka “Roswell”, has had more work done to get it running – around the pick ups, the front bogie and a semi-permanent coupling to the tender.

I had intended to take some wagons to Spain, to build in the sunshine, but that plan has changed! They are making steady progress but have presented the odd challenge.

The main question that I am trying to resolve is what couplings to use. American modellers tend to resort immediately to Kadees, but, in 1860, those had not yet been invented. The US standard at the time was the link and pin – which does not lend itself to automatic coupling! Given the current lack of access to most mail order, I am strongly tempted to use AJ couplings as I have the necessary stock of Ernie Ball guitar string. Couplings would need to be attached to the bogies and I suspect that the only viable solution would be to use a hinged version of the AJ. Alternative suggestions would be welcome! Whether this will work with rolling stock that does not have buffers, is an interesting question.

And quite how I shall put a front coupling onto a loco with a cow catcher is another good question.

Best wishes

Eric

[Corbs]

I always look forward to updates from Roswell Mill, Eric

[burgundy]

One step forward, one step back.

Bringing together all the bits that have been built seperately and hoping that they will work is the most frustrating bit of building a layout. I did and they didn't.

First, it became even more obvious that the flat wagons are going to need a bit more weight. This is no great revelation, as they are made from wood and the bogies from either wood or 3D print. Imagine a wagon made from balsa wood, which will blow away in a stiff breeze – but not necessarily along the track. The first one built has now been “adjusted” by the addition of lead shot in the underframes and some bits of lead in the bogies. I have experimented using floor polish (Klear) as a way to glue the shot into the frames and, at first sight, it seems to work. However, I suspect that I will find a trail of lead shot around the layout in due course. Fortunately, it is probably about the right size for HO musket balls, which will make for an interesting talking point.

Secondly, having cleaned up the track and checked that the loco works happily upside down in a cradle, I have been trying to work out why it will not work right way up on the track. Immediate discovery is that there is a short somewhere – I am not sure whether it is still (see earlier in this thread) or again. I am now doing what I should have done at the outset and started drawing out the wiring diagram. The good news is that, by looking back at some of the earlier photos, I can see exactly where some of the better concealed wires actually run. I will need to sit down with a damp towel and try to work out whether it is a random short or something that I have designed into the wiring.

Back to spaghetti junction.

Best wishes

Eric

[burgundy]

An unproductive week, with more than its fair share of “oh bother” moments.

Is there an award for the ability to create new and improbable wiring faults? Something like an NMRA “Master wiring cocker upper”? There are examples earlier on this thread and also on Vintners' Yard.

I have progressively tested and disconnected the wiring on the system so that I am now convinced that the fault is in the 3 way stub point. The stupidest thing that I have done was to not produce a drawing of the wiring before I covered it all over. I am now on first name terms with every inch of it, having had to rediscover it the hard way. The smartest thing that I have done was to connect all the wiring from underneath the switches to the chocolate block terminals; they may not be perfectly accessible, but they are easy to test and to disconnect.

I have had at least 4 “eureka” moments when I thought that I had found the solution. One involved a short circuit in one of the switch positions on the 3 way switch – which involved digging out the switch from its panel. I am still not sure whether it was actually a problem but I have made sure that it will not be in the future!

My current best guess is that the problem is a failed tie bar. All the wiring is currently disconnected and the fault only appears when the tie bar is in “push” mode. My theory is that the two ends of the tie bar are making contact inside the insulation when the point is pushed away. If anyone has any better suggestions, please let me know! Similarly, can anyone think of a way to repair it?

Plan A is that I have ordered a packet of new tie bars and I shall have to figure out how to install one of them. I have about a week to wait for them, so there is time for thought.

Plan B is to take a break and do some more testing, when I have forgotten about it all, to make sure that there is no other possibility. It is easy to get fixed on “the only possible explanation”; this is the third or maybe fourth that I have reached!

Plan C is to reconsider whether the old fashioned EMGS solution, of a dummy sleeper with loose fitting rivets, might be feasible in the space ( a bit like the way that Technohand originally built them!).

Getting this far has taken a good few hours, spaced out over several days. Fortunately, there has been plenty to do in the garden!

In the meantime, I may have a go at some of the rolling stock and find a way to install the AJ couplings.

Best wishes

Eric

Edited April 3, 2020 by burgundy
spelling

[Oldddudders]

No risk a piece of lead shot has fallen into a point frog or similar? 

[burgundy]

3 hours ago, Oldddudders said:

No risk a piece of lead shot has fallen into a point frog or similar? 

Ian 

Thank you. That is a good thought. 

It is not the source of the problem this time, but it is certainly a possibility that I will keep in mind when I go back to working on the freight wagons. 

Best wishes 

Eric 

[Corbs]

Hello Eric, I've had some success using high viscosity superglue/cyanoacrylate to glue liquid gravity in, building it up in layers. Only things to watch out for are risk of it 'bleeding' through gaps in the model and the potential for it causing misting on surfaces it reacts with.

Edited April 3, 2020 by Corbs

[burgundy]

I have parked the great short circuit mystery for the time being, pending the arrival of a packet of replacement tie bars. I did some experiments using a sleeper with swivelling rivets but concluded that it was not really feasible, given the extent to which the turnouts have been embedded in place.

In the meantime, I have gone back to work on some of the goods vehicles. Working in wood is a bit of a novelty and makes for an interesting change. I can see why it might be difficult to produce a flat car in etched brass or a bogie box car in white metal.

The kit of parts for a flat car.

Box car or house car - wooden with 3D printed bogies.

There are two questions that I am trying to resolve.

How high should the platforms of the flat cars be above rail height? The kits contain a wooden bolster to which the bogie (truck) is attached and this seems to push the platform up to an unrealistic height. Without the bolster, the platform is about 4 feet above the rails, which sounds like a reasonable height to allow loading from ground level. Including the bolster adds about another 9 inches and creates a rather top heavy effect. On the other hand leaving the bolster out causes the bogies to conflict with the tie bars as they strech back from the cross beams, so there is a trade off between appearance and running qualities. It would be nice to know what the prototype did in order to inform a decision on the best compromise.

From left to right - box car with additional bolster, flat car with no bolster, tender of loco.

Even worse when you compare the boxcar directly with the tender!

The second question relates to couplings. This was discussed in a prevous post but remains unresolved.

Best wishes

Eric

Edited April 10, 2020 by burgundy

[burgundy]

With work on the electrical side stalled, I await the arrival of a couple of replacement tie bars. Railway modelling is supposed to provide a relaxation in times like lockdown: even on a simple system, chasing short circuits has rather the opposite effect, resulting in carpet chewing frustration! While waiting for the postman to arrive, I have cotinued to work on rolling stock.

The available stock consists of 5 cars; an ALKEM flat car, two BTS box cars and two BTS flat cars. A couple are US Military RR, which have evidently been hijacked in time and space to run on the Western & Atlantic RR, and two are Georgia RR.

On a US based e list on US Civil War railroads, I asked a question about through running of freight stock among different companies. The answer seemed to be that it was generally unusual, although possible where particular lines had established a relationship and trusted each other to send the stock back! Evidently the idea of a Railway Clearing House had yet to arrive. My conclusion was that the appearance of GRR cars on the W&A would be plausible and so they will be lettered GRR.

When you look at a map of the Southern States, for example on Dave Bright's Confederate Railroads site,

http://www.csa-railroads.com/

it is clear that the network was actually pretty thin, with none of the interconnectivity that developed quite quickly in the UK and also probably in New England.

The other point that I concluded from the Confederate Railroads site was that I had a disproportionate number of flat cars. Looking at the various railroads in Georgia, box cars seem to outnumber flatcars by 2 or more to 1. I have therefore used the components in the Alkem kit to convert that into a gondola (which is why the side planks need an extra coat of paint). This is an interesting contrast to the UK, where the vast majority of stock consisted of open wagons for either merchandise or minerals – all 4 wheeled of course!

Having decided to have some vehicles as W&A and others Georgia RR, I have tried to use different colours to provide some degree of differentiation. The choice is limited to what I have in the cupboard and to those which have not gone solid! Shopping for more is not an option at the moment. Slate blue/grey seems to be the description for both companies and so the GRR vehicles have been painted dark blue (in the photos it has come out almost black!) and the W&A have got dark grey. I have asked on the US list for comments, as the lettering has not yet been applied and both can still be painted over. If the consensus is that both should be the same, that can easily be done! Roofs of both are “black” and can be toned down with powders or chalk; should black also be used for the footwalk along the roof? Out of curiosity, what would slate blue pigment have been made from? I cannot think of any UK company that used any shade of blue on its goods stock. Grey (all 50 shades), iron oxide and brown are all common on goods wagons and, as far as I know, blue only very rarely appeared on locos and passenger stock (GER, Caley, S&DJR?).

I have also been struggling with the height of vehicle platforms. The photos attached show the first time that I have had all the cars mounted on their bogies and standing on track. The wheels are NWSL P87 profile, gauged to 16.5mm back to back. This fits the track gauge of 17.5mm, representing southern 5 foot gauge.

One vehicle uses wooden bogies from the ALKEM kit; the other four are on 3D printed bogies. I have not used the thicker bolsters in the BTS kits in order to reduce the ride height, but this leaves the wooden trucks standing rather higher than the others. I shall have to see what can be done. It also means that there has been come conflict between the 3D bogies and bits of the vehicle frames. I am also uncomfortable with the way in which the axles seem to rest against the central longitudinal member of the 3D printed bogies, rather than running on bearings inside the axleboxes. Do NWSL wheel sets need metal pinpoint bearings in the bogies which would be the normal way to mount wheels in the UK?

Best wishes

Eric

[burgundy]

You know things are going astray when you deviate from your deviations. With the tie bars still somewhere in the post(?), I have been working on the freight vehicles and making quite good progress. I have been picking away at things that are not photogenic - like adding weight, getting the ride height more or less aligned and doing some painting. However, while the paints were out, I had a look at the figures that I have collected for this model.

Inevitably, there are some of Andrew Stadden's Victorian labourers in HO. With the addition of broad brimmed hats to keep off the scorching Georgia sun, they fit in pretty well. The less usual figures, however, are from Knuckleduster and are two sets of wild west figures. Upstate Georgia at this time was still not long settled and so they seem to fit in quite well – although I did not find a use for the cowboy shooting up the town or the lady who looks “no better than she should be”.

As an aside, at the beginning of the war, 44% of the population of Georgia was Afro-American, so my five out of fifteen is a slight under representation.

Best wishes

Eric

Edited April 24, 2020 by burgundy