aleopardstail

ok, the track is fine electrically, dropped the layout into DC mode and broke ok the Poole era GF Class 47, it runs just fine, added the peco wagons I have behind.. and apart from looking actually quite nice now the track is less toy like it runs well - I have noted some vibration in a few places - places the Lima Class 86 struggle to even move. closer inspection shows the inside edge of the railhead has paint, as expected but isn't all that smooth. can also hear the logo flanges hitting the sleepers in a few places but with the six axle pick up the loco keeps going. it does however object to a few of the points, where I suspect a bit too much paint in the flanges around the check rails and crossing check rails. pushing the Lima Mk1 stock around and the track doesn't feel smooth, the peco wagons, which are a lot more modern and have smaller flanges run without any issues - suspect the cure is going to be to remove the pain off the inside lip of the top of the rails, probably not a massive job and to clean the turnout flangeways the same - this a short term fix, longer term is basically get some newer stock or replace the wheelsets with something a bit more modern for the rolling stock. may need someone with a lathe to adjust the locomotive wheel flanges though.. hmm.. still its quite nice to watch the Class 47 circulating.. may consider more seriously one of the Class 04 shunters I've been pondering a while and get something less than 30 years old to have a run with.

supply pickup today, still working on the poor quality running issue - the rail head even after cleaning and using the track rubber doesn't feel completely smooth, still rough to the touch so I suspect it needs a bit more work. anyway at my local gaming shop today while picking up another couple of bronze medals for gaming prowess, I'd not mind but its only a two player game *cries*.. they do scenic stuff so now have a bag of Jarvis extra fine ballast to experiment with, also picked up some plastic strip that is a small I beam but looks closer to bullhead rail to my eye - its not perfectly in scale, however the plan is to experiment with a bit of that and some small surface mount red LEDs I have and try to make a red warning light for some buffer stops with the LED and some enamelled copper wire that is pretty much invisible, glue it together, add a bit of paint and install with a resistor below the board. I did think about using a PWM LED controller board allowing variable brightness but to be honest probably just pick a resistor and connect to the 5v power bus I have and be done with it, always on when the layout is powered up. interested to see how well it turns out

Hmm cleaned on the right, as it was before on the left. have checked the track and all but one section is live - one wire adrift which has been reconnected. neither loco will run correctly, both better on straights than corners - to be certain it wasn't perfect prior to painting but its a lot worse now have gone round the track with the PECO track rubber and with cotton buds & IPA, track is essentially clean, only have to touch it with the multimeter and it picks up the power fine.. an ponderance for sure, not got much more time to play tonight but I think the next step is back to DC mode and one of the DC diesels to see if they are any better or any worse - given they have more pickups hopefully better

testing... well the track has power, however neither class 86 enjoys it - both with run but only on limited areas. I suspect whats happening is the rail head is clear but the locos actually seem to pick up from the flanges, the actual bit of the wheel that touches the trach is clean, but yellowed which I suspect means further more in depth cleaning is needed, slightly disappointing but then they are very old models - which usually needed a bit of encouragement to get going anyway. however they do look a lot better on the track now, cleaning time

decent weather, and a delivery of paint.. what is a humble leopard to do except demonstrate that yes a 4'x2'6" plywood frame layout is technically portable but its not something I want to shift that often.. Dingleberry lugged downstairs, past the cat minefield out to the garden, where a coat of army painter "uniform grey", "leather brown" and "black" primers were used, followed by railmatch sleeper grime to produce this quite like the colour, the lighter brown is the leather brown primer. the darker area isn't just shadow this is the bit that will eventually be a tunnel so black seemed suitable. could likely replace the sleeper grime with a paint from my gaming range but to be honest as a one off cost the spray, which is also an enamel, is worth it. have the acrylics to manually do the bits around points (taped off in the above) - that will wait for the temperature to drop a bit anyway back inside, the buffers got a bit of paint basically drybrushed the beige brown, the flat earth then the excellent riza rust. no idea what a "riza" is meant to be but apparently they go rusty. the red and white buffer beams work, this will get a brown and black wash once the ballasting is done but for now it works nice an atmospheric as well shows the bits the tape covered, have printed more servo holders to sort of the turnout entry to the depot area, however the back wall now fits in nicely. next two jobs, paint the white paper bits with probably "flat brown" which is pretty similar and good enough, then the bits of the turnouts not sprayed with the brush on acrylics

the depot crossing is now physically complete, will be painted when the track is not massively straight, once the ballast and a few weeds are about should be fine though also painted up the retaining wall needs a hint of green and some matt varnish adding but I'm quite pleased with that, once its fully integrated I think that will look quite good. need to think of what to do with the triangle of land in front of it, thinking a portacabin or two on some hard standing with various plants growing around, and maybe a bit of graffiti on the wall

beginnings of the depot access road over the depot tracks basically 3d printed in fills, the ends have a slight slope and a check rail effect that I hope will work when painted. also the first actually usable 3d printed walls completed and stuck together, though not yet stuck in place, again prior to painting bit to the left will be inside a tunnel the weird bit next to it is the abutment for the plate girder to sit against, all this hopefully painted up pretty soon to stop it looking quite so "train set", even though thats pretty much exactly what it is

next few sections printing, machine says about 20 odd hours to go.. ouch experimented with painting a length of track, my first thought was "spray it all dull flat brown, add a bit of drybrushing and a wash", the brown sprey I have is too light though. second thought was brush application of vallejo model air "rust" (airbrush off line currently), that sort of worked, but took three coats and even then doesn't adhere very well, looks ok with a black wash over but still not happy. have ordered some railmatch sleeper grime, figure may as well take advantage of the hot weather we have coming and get this outside, a grey primer added then get the railmatch over it, once cured can bring back inside for drybrushing, washes and similar.. in preparation all the buffer stops are now fully glued into place, holes drilled for adding some lights to them, will stick slivers of blu tac around the point contact areas and generally remove the rubbish. expect this to pretty much transform what the layout looks like

test blue wall now painted, this has actually turned out pretty nicely looks better than earlier versions, maybe needs a bit more layering with the brown for a more gradual fade, eventually will have a bit of greenery in a few places. design revised to have a "foot" which will make installing easier as well as I hope printing better, for these that will go against a backwall may also print the narrow pathway as part of the wall, probably easier then a separate strip (goal being to have a couple of pedestrians maybe) painting is basically a base, bit of a drybrush in a few lighter shades, then a brown and green wash will have another go at printing some tomorrow, also thinking I can probably get the area this will go painted up as its a pretty basic area in a corner

(lack of) Engineering Blue Brick... (no accurate measurements were harmed during this process)

more progress, first test section of the second version of the wall printed, sadly a slight misprint makes it unusable (lifted from the bed at one end so its not square - may add a small base raft for stability anyway which should help with this. Metcalfe original in the middle, the first iteration on the right, looks ok but not happy with the archways and how they printed. 2nd version to the left looks a lot better, got the height better too - the buttresses slope back at 3 degrees (zero slope back on the 1st version) and look a lot better for it. flatter arches came out better too still have the horizontal texture, not getting away from that but once painted can hardly see it, probably need a top cap detail on the buttresses. needless to say the blue will get painted over. still learning how to use Blender & the printer but coming out ok

cheers, thats very effective, especially given what it is. I tried a brickish red and it looked rubbish, may have to try that amazing range of kits they make

redesigned the wall section slightly, now have more flatter arches instead of the full roman arch which I think will look a bit better, also tweaked slightly to print better. to the rear the test article had a small ridge to glue the road deck to, revised so thats a bit larger and better supported to print better - likewise the cap stones now are in a few layers to print better. devised the bit for a tunnel entrance which will be printed over the next few days before moving onwards the bit over the tunnel is a placeholder to get the spacing and alignment correct, will be replaced by a plate girder section eventually, the walls including the bits for that to be mounted on

looks good from where I'm sitting, nice use of space, shows what N gauge can do given a reasonable length too. Q: what paints have you added to the Metcalfe buildings to hide the corners?

In the middle of everything else I thought I'd have a go at a bit of scenery, for a change. I want some retaining walls, the Metcalfe ones look nice enough but need quite a bit of work to hide all the exposed edges, they also would require quite a bit of work around some of the shapes needed. I considered my 3d printer, now this can't really do 2mm scale brickwork well, it can sort of do it but takes a lot of work at the modelling stage, but what about concrete? well concrete has been used for walls, especially in the 1970s and 1980s, and while it may not look amazing it should at least be possible. A bit of work in Blender, then a few hours printing and a few more painting and this is the current state of it only the right half has had the weathering started, definitely needs more work here, likely some drybrushing to lighten the centre of the flat areas and likely a second application of the base in a few areas (its not quite this bas to the human eye), overall its a start though - well this section is a tester, may make a few changes to it but as a general concept open to thoughts and comments. colour is vallejo model air "concrete" and the lighter base is "cement grey" basically just to break it up, weathering is sepia shade and a bit of model air "dark green"

yes the design is clever too, I did have my own, firstly starting using infrared sensors, easy to make, derived from stuff I made for a scalextric layout, though here just checking something is there and not trying to decode the IR light the cars have. That worked, mostly, but the black underframes needed a bit of white paint to work reliably, and was only a "point" position. went over to current transducers, initially using analogue to digital converters and an arduino - good as you get a current reading and wasn't hard to tell locomotives from coaches - that worked but was a bit fiddly to set up. ended up with the MERG ones which are clever, using a 555 (well two as a 556) timer and having the DCC pulses reset it is very clever, does mean you get a slight delay on exiting, which is no bad thing and is remarkably good at rejecting noise in the signal. the only difficult bit of assembly, and very much "for a given value of difficult" is fitting the transformer loop wire into the board, have made two slight changes - firstly the LED is replaced by a 2 pin header so the LED can be on flying leads to be where I can see it - though won't bother for future ones now I know they work so well and replacing the four pin connector with a small four pin screw terminal block to avoid faffing with cable crimps. can see at some point investigating more of their DCC system

slightly more hardware development, the board has the first twin channel MERG occupancy sensor board installed, this has been basically running on its own with just the onboard LEDs as indicators - have now hooked this to a breadboard based processor - an Arduino Nano & a 74HC165 static shift register to read the inputs, currently just one but they daisy chain easily. this is driving presently a paid of LEDs just to check its reading fine, wasn't expecting issues and its working perfectly, the initial pass will be 16 blocks and updating the software to manage that won't be difficult, I have some RS485 boards here as well so adding this to the C/MRI input is not thought likely to be hard. will leave as breadboard initially but likely move to a stripboard design with the eventual five shift registers on board along with the Nano and interface board. amazing how simple the block detection stuff actually is, the MERG board is really quite clever

more testing completed, and yet more track cleaning and loco wheel cleaning, waiting for the Zimo 617F decoders to be back in stock at sensible prices and will get the Cl25 & Cl47 converted. Also have some more of the MERG block detection boards on order, not enough for the full layout (some 33 blocks, excluding the TMD which will not be wired with sensors as there isn't much point), but enough to get around half done - the outer loop and sidings then a bit of the inner loops. working also on the point servo shims which will complete that side of things nicely

a pair of slightly elderly class 86 by Lima, each with three similar vintage Mk1 coaches, also by Lima, having a bit of a run out initial track testing showed that by eck there is some track cleaning needed, not helped by the twin axle pickups of the locomotives. Also one or two turnouts do indeed need a shim adding as the actuator arm is slightly too long and catches on the subframe - e.g. the depot entry point is fine in straight running but blocks on the diverging route into the tmd itself. that apart it does appear to actually work which is good and once the locos have run for a bit they are a lot better at stopping & starting without the divine finger of "nudge" helping. and also very nice to sit back and control the lot with the mouse given the two DC only (currently) locomotives do run a bit better I need to get them chipped at some point, could also do with a slightly more modern locomotive I suspect, but need one that can cope with 1st radius curves and settrack points

FIXED! problem was classic PEBKAC. turns out the code I wrote was cleverer than me, it knew -5 was smaller than +15, and I somehow missed that and was telling it to start at +15 and to increase until it reached -5... and vice versa, the code seeing thats gibberish sensibly decided to sit there and go "nope". in effect an ID10T user configuration mistake. now have all 16 turnouts working, and driven via JMRI, though not yet with an actual panel but can control directly from the control table, and each is connected to a pair of virtual sensors managed by the driver software so it can report when the turnouts are usable or when they are moving, and all that works nicely. onwards with testing!

this is... odd... I've gone into the servo driver code, and basically nothing in there that differentiates between the servos, all 16 use the same code, the only difference is which of the control registers get written to.. yet still one behaves differently, will have to see if I have a spare driver board somewhere and do a bit of bench testing, I do hope thats not the problem as replacing the driver board will be an embuggerance as its got a big sticky pad holding it in place currently

done some investigation. the "non functioning" servo is nicely locked solid as if its on power, power off and can be moved manually and returns to the locked position when powered on - swapping wires round between servos indicates the servo is working fine, thus its a control issue - can't blame anyone else as I wrote the code on the little Pi Pico that is driving that. digging into that is a tad more work as it means rearranging the desk to get at my Pi 400 development machine and then hooking it all up, if I had to guess I have at some point used #13 for device testing and its got a hard coded test output I've used to auto centre servos at some point (the one on the board returns to the centred position). will likely have a look later as need to rearrange the desk for tomorrow anyway (work stuffs). the joys of debugging hardware & software systems. Still on the plus side I have a Pi Pico talking to the PC over C/MRI which I'm quite happy with, really do need to get that actually installed properly out of the way

all motors now connected, one point (#13, go figure) is not responding, past performance indicates its probably wired in backwards, will investigate further. otherwise all working and now configured, bit of fine turning when the lot is tested and a few actuator arms to adjust in length slightly. frustratingly the hole in the tie bar is close enough to the turnout that its quite possible for the actuator to catch passing side frames if not set just right - a bit easier to adjust with the newer servo mounts as not too hard to stick a 0.5mm or 1mm "shim" in the right place to slightly adjust it. overall happy so far, will get point #13 working and then spend a few days testing, which is serious stuff and in no way shape or form is like playing with toys, it just happens that to the untrained observer it looks exactly the same

last five servo mounts printed, hope to install tomorrow and promised the extensions to wire them due Wednesday... job after that is to shift the controller from breadboard to stripboard and get it physically installed and the job after that is to investigate how hard it will be to power the crossing rails with dedicated feeds from the stock rails of the remaining points (so I'm never depending on switch rail - stock rail contact for anything electrical)

the driver software is (currently) set up to drive 16 servos and simulate 32 switches back into JMRI, it wouldn't be all that hard to replace some or all of that with actual switches (reading them is just a few 74hc165 static shift registers and not a lot of code afterall. have noticed one or two going "heeeellllllooooo Mr Endstop" so far have this as a basic 3d printed mount, no switches though likely wouldn't be overly difficult to adapt to have a pair of limit switches low down and swap the control horn for one that sticks down as well as up. will have to have an experiment, the ideal is something monitoring the actual point blade position but the servo horn position is probably next best, pair of small micro switches and a small circuit board to hold some screw terminals and some pull up/down resistors for position.. pondering...