Mim

Got a question on rail chamfering. For things like turntables, sector plates and traversers, where moving sets of rails meet fixed sets, does it help to chamfer the insides of the rail ends to compensate for slight misalignment. If so, how much chamfer should be put in? Half the width of the rail top, full width? How far back should the chamfer start (angle)? I suppose the same question applies to baseboard joints too. Thanks, Mim

Page 19 of the 2018 yearbook for the website user name and password. Mim

Indeed. 179 solder used to be an alloy of Tin, lead with 2% silver. It is a eutectic alloy, so all melts at 179C with no pasty range of temperatures. A lead free version made up of tin, indium, bismuth and what have you would almost certainly not be eutectic and will behave differently when soldering with some pasty region of mixed liquid and solid metal. Either they have changed the alloy to a lead free version, in which case you'll need to experiment to see how it performs when soldering compared with the old version, or they have been rather naughty and mis-labelled it as Lead free when it is not, or it no longer melts at 179C. No problem with still selling lead based solders, just so long as it isn't used for commercial electronics (with a range of exceptions). Jen

I have the same chassis kit on my shopping list. My understanding is that you need six frame bushes 3-113 for the wheels to go through and four drive bushes 3-112 for the idle gear spindle that does the two stage gear reduction and two support each side of the motor worm gear. The association wheels have brass axles, so go in the pbhosphor bronze frame bushes. The idle gear will need a split spindle from 1.5mm silver steel rod 3-110 and run in brass drive bushes. Te idea is that you don't have the same material for both bush and spindle ideally. Brass ones are cheap to make p-bronze are expensive, so are only used on the brass wheel axles where they are best suited. If you are only building one chassis, then I suppose you could use one 10 pack of P-bronze bearings for everything, but for more chassis it works out cheaper to use different materials. Notbuilt one yet, so needs a sanity check from those that have! Mim

Hello Richard, The solder pastes have a limited shelf life before their dispensing properties go off. The refrigeration bit is a recommendation for industrial scale electronics soldering, where consistent and repeatable dispensing is essential (my day job at various times!). The high solids content and particle size make syringe clogging a common risk, hence the shelf life and refrigeration restrictions. Turning what you have got in to a paint by adding more solvent, water, isopropyl alcohol, or whatever would make it usable again as Caley Jim suggests. These days, water is probably a good solvent to do this, but try with a small quantity first. If not, then an organic one like IPA. The solder paste consists of fine solder particles, solvents to make it flow and liquid flux to deoxidise the metal and solder surfaces when the joint is made. If the paste was made for electronics the fluxes are not that active compared to the liquid fluxes model makers use. Particularly the pastes sold as not requiring cleaning of flux residues after use. The solder particles will grow thicker oxide layers as time goes on and the flux may go off, but adding liquid flux to the surface being joined if required and more solvent to make it a paint should extend its life for model making for many years. Mim

Some years ago I was stood on the platform at Long Buckby station waiting for the stopping train to arrive. Every few minutes it seemed a Pendolino would sweep by going one way or the other. The station is on a curve and since Pendolinos all look alike to me it wasn't hard to imagine I was on Richard Bransons's roundy round train set. Rather like a 1:1 version of this layout. Sometimes a model can look like the real thing. Sometimes the real thing can seem like a model. Mim

How about the fictional Northamptonshire village of Tilting Over? Mim

I am planning to use Neodymium magnets, swung through an angle under the baseboard to operate DG couplings. Is it worth having half the magnets North up and half South up to minimise the operating wires on the couplings getting magnetised? I also will have at least one location where I plan to have two magnets being swung by one servo to allow uncoupling on closely spaced double track. The magnets will be 21.5mm centre to centre here (4mm diameter by 3mm thick magnets). Again, is it better to have the same magnet pole up, or opposite poles, or does it make no difference? I am thinking of the interaction in the magnetic field between the two magnets. Am I overthinking again?! Wasn't sure where to put this, but a lot of 2mmfs modellers use DG's and some of them use permanent magnets for operation, so I thought this would be the most likely place. Mim

I've only just found this topic and will be going through it and shamelessly copying ideas. Have learnt something already; the existence of the "Code" button in the forum editor. Sure there is more to come. I have been working on an Arduino and stepper motor controlled fiddle yard traverser, which so far seems to be successful. Here is a link to the topic that describes it. The next Arduino project for the layout will be servo operated points and uncoupling magnets. Mim

An update. Slow but steady progress here with the electrics currently being worked on, if you'll excuse the pun. I have put in a set of five copper tape bus bars from one end of the layout to the other. The five rails are track front, track rear, ground, 5V and 12V. 5V is to power servos. 12V is to power Arduino PLC's, lighting LED's, controllers etc. Connections to, from and between the two boards are by five pin XLR connectors. I originally bought 5 pin DIN connectors, but decided to change to the more robust and better designed XLR's. The idea is that with at least two XLR sockets on each board, then any board can be powered on its own by plugging 5V and 12V in to one socket and controlled by plugging a controller in to another. A plug to plug jumper connector is used to connect the fiddle yard board to the layout board. Any socket can be used for any purpose. The modular design of the electrics is also in line with the aim of making the fiddle yard reusable for future layouts. I'll be using DCC, with a Sprog 2 unit. This uses 12V, so the Sprog is attached to an XLR plug to both draw 12V power and supply DCC to the layout bus. Other DCC units could be used if required, just by fitting an XLR plug to the output. The layout could also be run on old fashioned DC control by plugging a suitable DC controller in to any XLR socket. The layout would be operated as one engine in steam under DC control. The fiddle yard traverser has isolator switches for the four tracks. Thus, any other DC locos, or DCC fitted engines could be isolated. DCC will only be used for train control. Turnout and uncoupling magnet operation will be by switches on each board, connected to an Arduino that will move servos. Each board is then independent of the other and no external control panel is used. This is in the spirit of this part of the Cromford and High Peak Railway, where points were thrown with local levers and there were no signal boxes, or ground frames. The track power to the traverser table is via an 8 way ribbon cable. This is clamped to the underside of the table and to the base of the fiddle yard board by curved hardwood clamps and M3 nylon bolts, with a length of cable to flex gently in between as the table moves. Four of the wires are connected direct to the track rear bus. The other four are connected to the track front bus, via four SPST switches to isolate traverser tracks. The copper bus bars on part of the main layout board, with connections to XLR sockets on the front and back of the board. The fiddle yard board with the copper bus bars and the ribbon cable connection to the traverser table being put together. The cable is clamped at each end of the flexing section and tails have been pushed through the table ready for connection to the track. The other end has since been connected to the track bus and the isolation switches. Mim

You mean the films about the illegal unaccompanied child migrant from darkest Peru? Not seen either of them yet I'm afraid. Should be sent back to where he came from without any marmalade sandwiches. Mim, who loved the books when she was little.

Thanks Andy,Will look in to these. Probably going to stick with the stepstick for now, if you'll pardon the expression. I've fitted the support bearing on the other end of the leadscrew and it has damped out the worst of the vibration. It is never going to be super quiet as the thin plywood baseboard is acting like a sounding box, but I am ok with it. Mim

This one caused me some head scratching recently trying to make sense of a LNWR drawing and the quoted wheel diameter. There is an explanation on the LNWR Society web site. Basically the wheel size quoted for LNWR engines doesn't match any dimension on the actual wheel! Sort of, well a bit if you squint. Mim

The homing procedure after power up is a bit noisy. Comes from the way the stepper is triggered during that stage. It isn't intrusive once it has homed and is using the accelstepper algorithms. The cheap phone microphone makes it sound worse than it is. Mim

I've now got the fiddle yard traverser moving automagically. I found the 28BYJ-48 stepper just didn't have enough torque to move the table reliably without missing steps. Tried changing from the rack and pinion drive to a 2mm pitch timing belt, but had the same problem. There was nothing for it but to go to a bigger stepper motor. This is a NEMA 17 with 1.8 degrees per full step. The ULN2003 driver board wasn't powerful enough for this, so I tried the H-bridge board again. This wasn't successful. The motor was missing steps and reversing itself randomly. After a lot more reading up I bought an A4988 Stepstick board to drive the motor and this works well with the new stepper motor. Another problem was the timing belt. Each step of the NEMA 17 motor would move the table about 0.15mm using the belt and pulleys. I wasn't happy with this large a movement considering the precision required for 2mm track and wheels, so a leadscrew and follower nut were purchased. This is 8mm pitch and gives 0.04mm of table movement per full step. Some new brackets were fabricated and the whole thing assembled. The motor is mounted in a bracket at one side of the traverser. The lead screw is attached via a flexible coupling and runs through the follower nut, which is attached by another bracket to the table. I am still waiting on a pillow block bearing to support the far end of the leadscrew. For the moment, this is in a plain aluminium hole. There are five push buttons for the five possible positions of the four tracks on the table to access the double track of the layout. A single press of the button will send the table to the appropriate position. To prevent any inaccuracy from backlash in the system the table does a little joggle at the end so it always comes to a stop from the same direction. There are a pair of microswitches to stop the table over running and getting damaged. One of these is used as a home zero position for the stepper motor. I still need to check the accuracy and repeatability of the table, but so far it looks OK. There are options of sub-stepping the motor with the driver board, but for the moment I am using full steps. The idea of automating this has lead me in to a bit of a diversion from actually producing the model. I now know a bit more about stepper motors and their control than I did a month or two ago. Trouble is I am now thinking "Home made CNC milling machine. Shouldn't be too hard!". Only two and a half years to finish this layout. Don't get side tracked. Below is a video of the traverser in operation. From power on it finds its home position by operating, then releasing one of the microswitches. From there each button press sends it to a new position, with a little joggle at the end so it always stops after approaching from the same direction. Nice smooth acceleration and deceleration, which should keep 2mm trains on the track. I've also included the sketch as Nick requested. This is extensively commented, so I've got a chance to work out what I did later. Mim Edited by Mim, who has just discovered the "code" button in the editor. Edited again 4/1/18 with the most recent version. /*Traverser stepper motor control for Middleton Top fiddle yard. Written by Madam Mim November 2017. Modified December 2017 to make the code more compact and add a single uncoupling magnet servo operation. Feel free to copy, plagiarise, steal, modify, or ignore. Especially, feel free to improve my kludgy coding. Works with an Arduino Uno, an A4988 Stepstick steppper driver and a NEMA 17 stepper motor. Can be modified for other Arduino PLC's, stepper drivers and motors. Uses the AccelStepper library to give smooth acceleration and decelleration, which must be installed. See http://www.airspayce.com/mikem/arduino/AccelStepper/ The uncoupling magnet servo uses the standard IDE servo library to drive an SG90 servo, powered from an independent 5V supply. No sub-stepping is used here, but it could be enabled if required. A single press of any of the five buttons will send the traverser to the appropriate track position. Pressing another button while in transit will change the destination. There are limit microswitch inputs each end to prevent damage. Hitting a limit switch will stop the motor. One of the limit switches is also used as a home sensor. The destination buttons, uncoupler switch and limit switches are grounded to operate. Arduino internal pull-up resistors are used to prevent spurious triggering from noise. Home sensing and all positions are finally approached from the positive steps direction. This is an anti-backlash measure to improve accuracy and repeatability of positioning. The sketch includes the operation of a single uncoupling magnet servo. This can be removed if not required. */ #include <AccelStepper.h> #include <Servo.h> //Set motor interface type to Driver (1) to suit the A4988 Stepstick which uses two pins, step & direction. byte stepStick = 1; // pin outs to the stepper driver. byte stepPin = 2; // The pin to send the step signal to the A4988 stepstick byte directionPin = 3; // The pin to send the direction signal to the stepstick //Output pin for uncoupler magnet servo. byte magPin = 4; // Define stepper1. AccelStepper stepper1(stepStick, stepPin, directionPin); //Define uncoupler magnet servo Servo magServo; //constants to hold input pin assignments. byte limitforward = 7; //forward and reverse limit microswitches. byte limitbackward = 6; byte button[] = {8, 9, 10, 11, 12}; // Traverser go to position buttons. byte magSwitch = 15; //Analogue pin A1 used as a digital input for uncoupler switch. /*constants to hold steps for each position. Adjust to tune traverser. 1mm = 25 steps (200 steps per revolution. 8mm pitch leadscrew and NEMA 17 1.8 deg/step motor).*/ const int backlash = 50; //steps to take during anti-backlash manouver. const int pitch = 563; // Pitch in steps between tracks. const int homed = 273; // distance from home to first position. //define positions to go to for each track. adjust +/-0 for fine tuning. const int pos[] = {homed + 4 * pitch + 0, homed + 3 * pitch + 0, homed + 2 * pitch + 0, homed + pitch + 0, homed}; //Constants to hold uncoupler magnet raised and lowered positions byte magLowered = 50; byte magRaised = 150; //Constant to hold uncoupler magnet servo speed delay. byte magSpeed = 15; // ms/deg. //variables to store current table and servo position and remember previous position. //Used for anti-backlash and change detection. int targetpos = 0; //initially zero for after the traverser is homed in setup. int oldpos = 0; byte magPos; void setup() { // put your setup code here, to run once: //Stepper1 speeds, accelerations. Adjust to prevent stock being rattled around. stepper1.setMaxSpeed(1000.0); stepper1.setAcceleration(100.0); stepper1.setSpeed(100); //define buttons as inputs with internal pullup resistors used. for (byte n = 0; n < 5; n++) { pinMode(button[n], INPUT_PULLUP); } //define microwswitches and uncoupler switch as inputs with internal pullup resistors used. pinMode(limitforward, INPUT_PULLUP); pinMode(limitbackward, INPUT_PULLUP); pinMode(magSwitch, INPUT_PULLUP); //home uncoupler magnet. magServo.attach(magPin); delay(100); //Delay for servo to sort itself out after attachment. magServo.write(magLowered); magPos = magLowered; //magPos set so changes can be acted on. //home motor //Travel backwards till you trigger the home microswitch. //Works by setting stepPin high, then low with delays. A bit noisy when running! while (digitalRead(limitbackward)) { digitalWrite(directionPin, LOW); digitalWrite(stepPin, HIGH); delay(5); digitalWrite(stepPin, LOW); delay(5); } //Back away from the microswitch slowly till it goes off. //As above, but longer delays to slow it down. while (!digitalRead(limitbackward)) { digitalWrite(directionPin, HIGH); digitalWrite(stepPin, HIGH); delay(10); digitalWrite(stepPin, LOW); delay(10); } //makes the home position zero steps. All subsequent movements are referenced from here. stepper1.setCurrentPosition(0); } void loop() { // put your main code here, to run repeatedly: //overtravel protection. Stops while either limit switch is activated. while (digitalRead(limitforward) && digitalRead(limitbackward)) { //Check each button in turn. //If they have been pressed, then move to the new position. for (byte n = 0; n < 5; n++) { if (digitalRead(button[n]) == LOW) { stepper1.moveTo(pos[n]); targetpos = pos[n]; //used to decide when to trigger anti-backlash moves. } } stepper1.run(); //move the stepper to the position. Non blocking. Program is looping round. /*Anti backlash protection. //After moving to the new position, step back, then forward. Final position always approached from the same direction.*/ if ((stepper1.distanceToGo() == 0) && (targetpos != oldpos)) { stepper1.move(-backlash); //relative move backwards. stepper1.runToPosition(); //runToPosition blocks everything else until complete. stepper1.move(backlash); //relative move forwards. stepper1.runToPosition(); oldpos = targetpos; //once complete, oldpos is set to targetpos to break out of loop. } //Check the uncoupler magnet switch. If it has changed, then move the magnet servo. if ((digitalRead(magSwitch) == LOW) && (magPos == magLowered)) { for (int m = magLowered; m <= magRaised; m += 1) { magServo.write(m); delay(magSpeed); } magPos = magRaised; } if ((digitalRead(magSwitch) == HIGH) && (magPos == magRaised)) { for (int m = magRaised; m >= magLowered; m -= 1) { magServo.write(m); delay(magSpeed); } magPos = magLowered; } } }

Welcome Benn. If you have just joined then how about making your first project the right size for the Diamond Jubilee Layout Challenge? That is what I am doing. 600mm x 240mm of scenic area and at least one turnout. Just a thought. You can download pdf templates of all the Peco point ranges direct from their web site. From there you could do a comparison with various 2mmfs switch lengths and crossing angles by importing them in to Templot and overlaying and tweeking a templot template to best fit. https://www.peco-uk.com/page.asp?id=pointplans Alternatively, the Association sell paper templates for a variety of geometry points to do an eyeball comparison with Peco points, or paper templates. Mim Not done it, but if someone wants to have a go and report back...

The size of cows has increased dramatically in recent decades. If your layout is set in anything but the recent past, then having cows that look small to our eyes is probably a good thing. See some of the pictures from Exeter cattle market in the 1940's. They look tiny compared with the ones you see now. Mim (who has the perfect signature for this subject!)

Finding a microswitch that will operate with the low weight of 2mm stock will be key. You can get some very light force ones. See. I reckon the incline rolling stock will have to be well weighted anyway to roll reliably on each end of the cable. A non contact alternative would be reed switches on the sleepers and a rare earth magnet under one of the vehicles. Or an optical beam from an IR LED to a detector across the two tracks, broken by a vehicle. The IR beam needs more electronics to be able to switch a motor, but a light force switch, or reed switch might need a bit anyway to switch and reverse the current of a motor. I had a chance to think about this stuff myself. One of my ideas for the DJLC was the catch pit of doom on the Cromford incline of the Cromford and High Peak. Two turnouts and an opportunity to divert wagons to their destruction when challenged if they work! It would have needed some shrinking to fit in the DJLC area limit, but was doable. Mim

Well the ULN2003 board arrived this morning. Plugged it in, uploaded a suitable sketch to the Arduino and switched it on. It works!!! Lots of blinking LED's (pretty! ) on the board to show which wires are being energised. Very smooth acceleration and deceleration from the AccelStepper library. Now need to finish the wiring for the push button inputs, wait on some suitable screws arriving to mount the microswitches and fine tune the programming. Mim

Beautiful. One thing it needs is a protective iron strip on the corner of the bridge on the tow path side. This protects the masonry from being damaged by the tow ropes of horse drawn boats. The ropes were made of cotton, but even so the iron gets worn away over the years as you can see on the grooves in this picture. Mim