Dingleberry, a N Gauge test track

[aleopardstail]

ok, the point controller is now correctly mounted below the board, and working fine, it got a slight adjustment with a five pin header added to talk to the servo board as thats got header pins so it was just the easiest way to do it.

 

tests the only two track blocks that are currently fitted for block detection with an early arduino prototype feedback board, and that all works nicely, really love the MERG sensors, have another seven to install (and build!) which is a total of 16 blocks wired up.. not quite half the layout but holding off ordering more until those are installed and working

 

more testing, I suspect the drive train on the class 47 is a bit gunked, measured as dragging about 700mA at one point, climbing a *very* slight gradient (the desk isn't totally level sort of gradient, not an intentional one) with four coaches behind it.. investigation ahead. the class 25 seems to be working mostly, with very occasional cut outs, I suspect a short somewhere, potentially trackside, likely just some debris.

 

more testing

 

and its not playing trains its serious testing.. aided by a cat

[aleopardstail]

little bit more prep work, the block detection input board.

 

 

the eBay bought 165 boards don't have suitable pull up resistors, hence the bit of stripboard with a bunch of 10k resistors on

 

 

another three to assemble, and one already installed, detection for 16 blocks, just under half what the layout needs.

 

also have pulled the trigger on ordering station building and platforms, platform lighting etc, leading to this little test for building lights.. 0805 SMD LED, with a couple of wires very carefully soldered to it. aim is to use as fluorescent tube lighting, there are warm white ones for incandescent bulbs - they are a bit larger as I also want to try making building mounted flood lights with the smaller ones

 

 

need more practice soldering small things

Edited March 8, 2023 by aleopardstail
comment on the 165 piggyback boards

[5BarVT]

I see you have the transistors inserted correctly the wrong way round in the DTC2s!!!  I have two of mine to alter once I’ve built the next two to replace them.

Paul.

[aleopardstail]

13 hours ago, 5BarVT said:

I see you have the transistors inserted correctly the wrong way round in the DTC2s!!!  I have two of mine to alter once I’ve built the next two to replace them.

Paul.

 

yes, read that bit of the instructions, a bit WTF? but having designed circuit boards it happens, at least you get the full bill of parts so can just order more of the transistors and cut them off to replace, hope to get the installed at the weekend, removing the current transducers from the earlier attempt at this, so all the dropper wires etc are in place, they just need the sticky pads and wiring up. the board to read them needs more work (like it needs the arduino side building) but hope to have the detectors in and tested

[aleopardstail]

Actual progress, as in something to do with the railway itself not stuff to eventually bolt to it..

 

A bit more ballasting along the mainline in front of the TMD and the two coach sidings, also the first bit of signal conduit (needs painting), have some relay boxes to go in with glue curing, they will be painted first. left the bits around the points and space for some ground signals, forgot to leave a bit for a signal for the second track from the left ahead of the crossover though, will have to be cut out later but thats a problem for another day.

 

 

also in sight, apart from the general detritus that any flat surface here attracts is one of the newer 0.2mm resolution portacabins, awaiting paint beyond just the grey primer

 

also have the other three occupancy detectors built, so one installed and now seven waiting to go in

 

really do need to get the point switch rails painted at some point too, plus pondering how to turn the ground to the left of the conduit into a car park, think its probably going to be filler of some sort painted. I've got the fencing for along that bit of the track as well

[aleopardstail]

Arduino board built, getting better only one mistake this time now corrected and installed. the 165 board also installed and its working well. JMRI sees it, sees 48 bits of input (the upper 8 bits are locked to zero in the arduino code for now but could be used later, C/MRI being multiples of 24 or 32 so may as well have it)

 

the one installed detection board hooked up, along with four more - one only connected to one circuit as I need to find the feed wire in the mess below

 

 

Think I've got it but will need a wire extension adding (not a huge issue). only 9 blocks (will be 10) connected but JMRI sees them and they work. three more boards to go in, need to move a few bits on the underside for this to make space and also need to make a few more 5v power distribution blocks to hook them up. should be in tomorrow.

 

so far have the two long sidings, and the ends live, and most of the track parallel to them. the remaining boards will complete both curves at one end. not planning to set a panel up until they are all in as the method of identifying which block is wired to which input involves driving a loco slowly and making notes.

[aleopardstail]

also a discovery.. and a "doh" moment..

 

flipping a railway board on its side to work on the wiring is a really good way to find which bits of ballast needed more glue, or even any glue

[aleopardstail]

Woo!

 

 

ok, not much to look at, but shows the part of the layout with working block detection, 16 blocks defined, turnout control works as well, still a few reliability issues to iron out:

 

- poor electrical switching by the turnouts, likely dirty blades and poor contact, cleaning job (I hope)

- poor running through the turnouts, fine in ahead, issue is usually when diverging, this could be down to the rolling stock objecting, cleaning may help I hope - extreme cases are usually the 47 or 25 hitting a turnout at lowish speed and either derailing (often) or outright tipping over

 

that said, goal of the weekend was to get the 16 blocks wired in and thats done, just another 17 to go.. in stages I want to finish the two outer loops and will then look at the station loop and sidings

- not all the track is cleaned up properly yet so some areas don't run very well

[aleopardstail]

slight progress, another four blocks wired up

 

 

more detectors on order, well the PCB since MERG have stuck a limit of two kits for these per order, ho hum, can source the bits myself so not a huge issue

 

have noted though that the sensitivity varies, guessing mix of component tolerance and likely wire length and maybe interference as some blocks will pick up a 10k resistor axle, some won't, some needing multiple axles to detect, e.g. the siding at the bottom has both test coaches, one with a single axle and one with a pair of axles. more testing needed

 

the derailment issues traced to about 50% the point servo actuator rods being fractionally too long - the cure for which is either removal & trimming or slightly shimming the servo mounts which is likely easier as thats basically slacken two screws and add a 1mm bit of plastic sheet. 

 

the rest are less clear but I suspect its old wheelsets and set track points as derailments are invariably around points, the class 25 being much better than the class 47 which is slightly older but also has longer bogies - lesson learned there, next time will be streamline points.

 

that said on the cleaned track both locos now run well, the two Lima class 86 less so and I suspect the two axle pick up isn't helping.

 

An eBay win class 101 should be arriving soon so will give that a whirl, plan is to add pickups on all eight axles and hard couple them with wiring between the vehicles and to eventually try adding some lights, but will get it running first

[5BarVT]

Interesting re the sensitivity of MERG DTC2 as 10k is what I want to detect.  I’ve only run mine with locos thus far so not a problem. They’re only for ‘add on’ sections when I have more than 8 so I’ll have to allocate them where they are not critical.

Paul.

[aleopardstail]

17 hours ago, 5BarVT said:

Interesting re the sensitivity of MERG DTC2 as 10k is what I want to detect.  I’ve only run mine with locos thus far so not a problem. They’re only for ‘add on’ sections when I have more than 8 so I’ll have to allocate them where they are not critical.

Paul.

 

I suspect they can be made, easily, more sensitive by adding another turn of wire through the coil - mine do detect 10k.. but its marginal and any track dirt or just iffy wiring pushing the resistance up spoils it, the next one I try will have another wire turn and if that works I will likely remove and adapt all the others the same (so far I've used the wire the kit comes with but since all the feed wires are single core I see no issues changing to use that round the CT). hopefully the parts required will arrive this week and can then experiment and report back

 

aim is to cover the layout, will be mostly locomotives though I do want to be able to automagically back stock into sidings so needs to be reliable

 

then I just need a way to prod the loco when it stalls, thinking of training a cat to help, what could possibly go wrong?

[aleopardstail]

Todays update, finally got around to adding the backscene, well the physical side of it

 

 

4mm plywood, with some stripwood bracing for which as usual too many holes were drilled, still its installed now and its staying installed (though may see longer screws added).

 

a plywood sky may be appropriate if you live in a boarded up house but around here we expect higher things

 

 

its possible a second coat of paint is needed..

 

hopefully this week will see the raised deck area in the far corner where the station building will live being added, a suitable electrical connection etc made to it and part of the blue overpainted with a dark grey colour for inside the tunnel area and on top, it will get a proper paved area added but grey paint will do initially

 

does look better though now the wall behind it isn't so visible, should also be slightly more stable when stood on the long side to access the underside

[aleopardstail]

more progress, not quite managed what I had hoped to this week as other distractions distracted me from this distraction

 

however, my eBay class 101 arrived, after validated it did actually run out came the soldering iron and a DCC chip

 

 

surprisingly simple DCC conversion as there is no split chassis or live chassis to worry about, just two motor contacts and two track pickup contacts.

 

and she does run... now...

 

whoever had this before noted a noisy runner and a bit stubborn. well for noise shes the quietest in my fleet, this is not saying much but old Poole era with metal drive train, whatever she runs. However to solve the "stubborn" that appeared to have dunked the poor beast in some thick black lubricant.

 

she ran far better with a strip down and clean, I suspect that gunk was causing the issues with poor pick up.

 

ok only two axle pickup so will never be great, plan is to add pickups to all eight axles which should solve that issue, this leaves just two small problems

 

1. the metal gears have enough lateral motion they can disengage leaving one axle drive, think a pair of small plastic shims may fix that, at least I hope it does

 

2. she doesn't like corners, the rivet atop the motor holding it to the bracket is stiff, resulting in derailments on 2nd radius curves, not yet tested on the 1st radius inner loop or over diverging points - this is why shes a bit naked in the picture - runs fine like this pickup issues aside, but even with just the unweighted chassis (whoever had this before "enhanced" the weight inside the bodywork with a ton of solder..) the rotation is far too stiff to take corners. this will take a bit more thinking. I've checked and its not wires fouling anything it is literally just the pivot is stiff.

 

all told though happy she does at least run after a bit of TLC and I suspect after a bit more will prove to be a decent runner. may see a bit of the interior detail removed though, have had to cut out a plastic peg to make space for the decoder as it is. I was pondering adding lights but not sure its worth the effort.

[aleopardstail]

update!

 

 

removed the screw in the yellow ellipse, which detaches the mounting bracket. The pivot here was utterly soaked in sticky black gunk, removed, cleaned with IPA, small amount of lube oil added (very small) and pivots freely.

 

reinstalled and she now runs, after wire brushing the wheels (and a bit more track cleaning) she runs, and runs round corners, and due to the added mass has enough inertia she actually manages a set track point with dead conductivity when moving at a decent speed, sliding over the dead bit to back to where the power is on.

 

figure a few more pickups (needed for the block detection stuff anyway) and she will run like a dream. 

 

The fleet now stands at:

 

Graham Farish

• Class 25, Poole Era, DCC fitting (Rails of Sheffield decoder) - "fun" install, not much space, does fit though
• Class 47, Poole Era, DCC Fitted (Rails of Sheffield decoder) - pretty easy install
• Class 101, two car, Poole Era, DCC Fitted (Rails of Sheffield decoder) - dead easy install

Lima

• 2x "Class 86ish", DCC Fitted (Zimo MX617F decoder)

I hope eventually to replace the Rails decoders with Zimo ones to have RailCom ability (The Rails ones are fine otherwise and for the money even better)

[aleopardstail]

More progress!

 

up to now driving trains has been via JMRI on a PC, using the on screen throttles. I have WiThrottle set up however my phone is ancient and not up to the job, even if the battery lasted, which is doesn't.

 

have been fiddling with an ESP32 and a modified version of the WiThrottle Arduino library, modifications that are not really used yet (have added turnout control handling), but the locomotive stuff worked.

 

so this evening the following crude prototype has been created

 

 

yes a bodge, and its got a few software bugs to iron through however when its behaving it allows selection of a locomotive: press "A", type the DCC address, press "#" and the address locks in, then the rotary encoder controls the speed and the two outermost yellow buttons control the direction (the middle pair will be set to +/- 10 on the speed at some point).

 

all a bit simplistic, the attached 128x128 OLED display (bought to be part of a scalextric start light gantry but repurposed as that project is going nowhere fast) is crisp and easy to read, will need to add an E-stop, and seriously re-write the code and likely a couple of the libraries used to streamline them to be just what is required but its not bad so far.

 

the code takes the roster list and the turnout list, currently doing nothing with them but the plan is when setting a locomotive to have the top half of the screen for typing an address and the bottom half with a scrolling list of the "known" locomotives to pick from. then with the turnouts allow picking from a list by scrolling through then just using the two outer yellow buttons to throw/clear. It not being practical to type the turnout addresses as they need a text prefix that keypad doesn't have (could bodge it though) but frankly the ID numbers mean a lot less than human readable text anyway.

 

not bad for a nights work on the hardware with some of the code work done last week.

 

when its working and cleaned up can post the code if anyone cares, the hardware is using PCF8574 I2C port expanders to stick the keypad and the buttons on an I2C interface for slightly less spaghetti

[aleopardstail]

development, there are some parts on order to tray and make this more usable as a proper circuit board..

 

 

the encoder controls the speed, the library has a nice acceleration function so has fine control without having to turn it several times for full speed.

 

yellow buttons manage reverse/forwards, and can also adjust the speed up and down (eventually will drive on screen menus as well). one of the white ones is codes as "set speed zero", the other is unused, the red is "all stop now" and the blue ones act as shift keys for the keypad so when running the 0-9 can activate function F0-F9, shift to F10-F19 and F20-F29 (well F28 as F29 doesn't exist)

 

a cludge but so far it works, will be swapping the keypad for two banks of 8 actual buttons which will simplify the code an allow pressing multiple buttons at once

 

now the wait for the bits to arrive on the China slow boat

[aleopardstail]

back to the actual layout for a bit, have now assembled and added the deck for the "town", which in practice will be the station building and likely a few low relief shops opposite and thats about it, little scenic break really

 

 

as annotated the town deck is removable by lifting vertically. the bridges will eventually gain metal girders which will be attached to the deck to make removal easier. once the glue is set the visible screws which hold support pillars on the underside of the deck will be removed, countersunk and refitted.

 

the idea is the deck mostly rests on the walls, which inturn are glued to small pillars visible behind them (the one of the far left is not yet glued as it needs a bit of reworking to position it properly. but the deck when removed then also has legs to stand on - this because its going to get some electronics stuff on the underside to control lighting for that module. the connections yet to be designed.

 

also shown are a few of the Metcalfe retaining walls which will be used to line the fixed walls, note "fixed" is an interesting concept given they are just glued to paper which itself is glued to XPS and then in turn glued to plywood.. so a tad fragile, will work on that a bit more.

 

obviously its removable to allow for maintenance, track cleaning, recovering stalled/derailed stock, entertaining the cats, removing cat hair but will also allow working on the thing as a separate module.

 

still, its progress :)

[aleopardstail]

Had the builders in!

the roofing contractor is due after the lighting technician has paid a visit.

 

the rest of the town is suffering from incomplete paperwork

 

found the GW shade "Agrax Earthshade" works quite nicely for the exposed edges of the brickwork, and diluted not bad for a bit of first pass weathering.

 

I was planning on adding SMD LEDs, however having now actually assembled the model given the internal floors and roof structures some 1.8mm THT LEDs are on order which can be easily mounted to the bits of greyboard and then wired up.

 

nice kits actually

[Lacathedrale]

Really interested in your custom throttle  - one thing I've felt is a bit lacking in commercial controls is the aesthetics and being able to mount in a mahogany cigar box with brass dials and ivory buttons - YES PLEASE!

[aleopardstail]

never thought on a wooden box with brass fittings.. would be possible, beyond my woodworking but possible. the software and electrickery stuff is actually quite simple, its basically an ESP32 module (off the shelf), a 1.5" OLED screen (also off the shelf, and there are many styles of screen, larger, smaller, whatever), a rotary encoder, which was on a shelf until it fell off. then some push buttons - I'm using port expanders to keep the wiring sane but no real need to...

 

Will expand on it more though as it gets more useful, hopefully the ESP32 with battery modules integrated will be here soon and the thing can be an actual wireless controller.

 

In the mean time while watching a mahoosive silver rocket just sitting there, I made a thing

 

 

needs sanding, and painting, and abutments etc making, but now have one of the plate girder bridge frames made. will let the glue fully cure then lightly sand it to get the front and rear work to be flat, add a further strip after to glue it to the removable deck properly and can then paint it up. Abutments will be made from the Metcalfe brick sheet and will likely be quite basic but will be nice to get them done.

[aleopardstail]

small update, made a start on building the bridge in, still have the other side and the top caps to add

 

 

a bit crude but to coin a phrase "it will do"

 

Edit: Boredom struck

 

 

forgot the Agrax Earthshade on the white edges prior to installation, that can be done tomorrow once the glue has cured.

 

not bad for something I've not tried in over a decade

Edited April 8, 2023 by aleopardstail
added second abutment

[aleopardstail]

Girder removed for painting, however the rest of the retaining walls round to where more custom parts are needed is in place now, also scratch built (well from brick card) a small building for the carriage cleaners to use - the roof will be added once the lighting is installed hopefully this week after the flipping LEDs arrive.

 

 

needs more work in this corner, it is however coming along

 

 

the little building, it is as you can see very basic but I'm happy how it turned out.

 

Pity Metcalfe don't do a generic pack of doors, windows, frames etc really

[aleopardstail]

A bit more progress today, the remaining DTC-2 block detection boards largely assembled, I still need to fit the CT coils, which need removing from some stripboard first and add the DCC terminals (on order)

 

MERG have revised the PCB which now appear to have the transistor outlines the correct way around

 

trying a different coil as the recommended one is not easy to get hold of in a cost effective way (e.g. Farnell want to charge an addition £16+VAT to bring them from a US warehouse, RS have them but want some £35 for 16 of them) - I have ZMCT103C type coils already so have bodged one board with them - can remove if its a problem, once the terminals arrive this one board gets installed as a test, the other seven will follow if it works. and indeed if it picks up my coach with a single 10k resistor accurately the existing ones will get additional turns of wire added

 

still no sign of the 1,8mm warm white LEDs for the building interiors though

Edited April 11, 2023 by aleopardstail

[aleopardstail]

tried one of the ZMCT103C coils, short answer: it doesn't work, well it will detect a 1k resistor or locomotive but not a 10k, nor a 5k. coil is far out of the required spec.. got some on order from the states, for somewhat less than a UK supplier, including duty & postage. should be here in a few days so will experiment further.

 

starting to take a serious look at some of the running issues though, specifically running through the set track points. I suspect back to back clearances may be a tad tight and for example my Class 47's tendency to derail every single time on one specific point in the trailing direction is I think its hitting and climbing on a check rail (turnout just off a curve so locomotive will be hard over one side and likely the inner wheel climbing the flange way) - batteries for the digital callipers on order, if not may take a knife and shave the check rail slightly to capture the wheel instead and see if that helps.

 

guessing all part of the fun with old rolling stock

[Grimly Feendish]

3 hours ago, aleopardstail said:

tried one of the ZMCT103C coils, short answer: it doesn't work, well it will detect a 1k resistor or locomotive but not a 10k, nor a 5k. coil is far out of the required spec.. got some on order from the states, for somewhat less than a UK supplier, including duty & postage. should be here in a few days so will experiment further.

 

 

Have you tried looping the through the coil two of three times? That should increase the ability to detect low currents.

 

John.