Toto's Traversers

There here - I ordered 4 of the wee fella's (I do have three traversers - so far).

With a screen size of approx 60 x 40 mm.

It's an arduino shield that will sit on top of either a Uno or Mega.

and notice the microSD card slot at the base, I have a few idea's for this.

Paul

 

Not now the bank accounts empty

But maybe by Friday ...........

Paul

 

Wow 4 years later .................

Life goes on and like most of the Bosses jobs list you can only punt them into the long grass for so long before it bites you in the ****.

My local club's new N gauge has had a few issues so will not be ready for the Clubs exhibition in October - so I've been asked to bring Another Bridge - oh

A few transport repairs carried out - stored vertically for 4 years in a spare room infront of a set of wardrobes was probably not the best idea. But then I only got planning permission for a couple of weeks .

Apart from it not working Toto's other main gripe was he didn't like using the DCC controller and or JMRI to control the traverser, he wanted more local control.

I tried to fix the existing program - coffee costs were mounting and I was getting knowhere.

Time to cut my losses - and start from scratch - looked at the sketch, and thought Forget that.

So I looked through the multiple versions to try and find one that I could follow - 4 years ago is pre-covid when I actually worked for a living and had some idea of what I was doing, definately older and not wiser.

Found one that I hope is similar to the version installed on the Mk2 traverser.

Stripped out all the DCC code - definately commited now - some would it was due years ago.

The intension was to make as few alteration to the board design as possible as that worked on the MK2 without issue.

Plan 1

Small form factor operating panel (may even be portable rather than fixed)
Oled display for showing status and track selection
Rotary encoder switch to choose between tracks, then press the knob to select the track.

The rotary switch has 5 connections - +ve, -ve, Switch, Data B (or DT) and Data A (or CLK).


Got the above to work, however was too sensitive, to move the track by one you needed to mid notch, and when added to the main program (sketch), it never picked up the changes.
Further research, recommended using 2 interupts, which on the uno / nano is ports 0, 1, 2 & 3 - Damn. Ports 0 & 1 are the serial connection, so already in use, and port 2 is connected to the Opto-isolator for the DCC decoding routines. If I was making this again then I may do it Rotary switch only, but I aint touching that side of the board as it works.

Plan 2

Small form factor operating panel (may even be portable rather than fixed)
Oled display for showing status and track selection
Push buttons to toggle between tracts and another to select the track to move to.

Now I have used 3 push buttons
1- Left hand Button - decrement track value
2 - Middle Button - Select track to move to
3 - Right hand Button - increment track value

As I only have 4 tracks and currently only one end of the traverser is in use, then a single button would have been sufficient - just three clicks away from any track. However I am wanting to keep the design flexible and easily transferable to a bigger project - you never know.

The OLED screen is an I2C device - just four wires +ve, -ve, SCL (Clock) & SDA (Data)

Ok first change, my homing pin was using D18 (A4) - it was convinient, but D18 is the I2C SDA pin, so a move to D14 (A0)
Final list is ..........................

D4 - Output Big Easy Driver (DR) Direction
D5 - Output Big Easy Driver (ST) Step
D6 - Output Big Easy Driver (EN) Enable
D7 - LED output
D8 - LED output
D14- Home Switch (A0)
D15- Push button - (A1)
D17- Push button Select (A2)
D16- Push button + (A3)
D18- SDA I2C for Oled Screen (A4)
D19- SCL I2C for Oled Screen (A5)

Now I would like to say it twas a doddle, but shear incompitance and ignorance prevailed

Everything was programmed in and not producing complie errors - eventually

But nowt appeared to happen.
I could see the inputs on screen, but no matter how many times I pressed the buttons the state didn't change.

The answer was only three lines lower, but I had added so many serial.Print statements to try and fault find the issue I couldn't see it.....

pinMode (pb1,INPUT);
pinMode (pb2,INPUT);
pinMode (pb3,INPUT);


pinMode(home_switch, INPUT_PULLUP);

I hadn't added the INPUT_PULLUP to the definition of the pushbuttons, just as an input. The effect is the pullup option puts 5V via a high resistor to the input, so it is held at 5V (or VCC which should be 5V or near as damit). The switch will ground it when pressed. I was getting spurious readings / values as the voltage fluctuated above 0.
The downside is the logic is then reversed - when the switch is open (not pressed) then the inut is HIGH (1), when the switch is pressed - closed then the input is LOW or 0.

So we know have

pinMode (pb1,INPUT_PULLUP);
pinMode (pb2,INPUT_PULLUP);
pinMode (pb3,INPUT_PULLUP);



pinMode(home_switch, INPUT_PULLUP);

Next up what do with the results

pb1State = digitalRead(pb1); // Reads the "current" state of the pb1 - Decrement counter
pb3State = digitalRead(pb3); // Reads the "current" state of the pb3 - Increment counter
pb2State = digitalRead(pb2); // Reads the "current" state of the pb2 Select

if (pb1State == 0 && pb3State==1){
counter=counter -1;
if (counter < 1) counter =4;{}
}
if (pb3State == 0){counter=counter+1;
if (counter > 4) counter =1;{}
}

oledDisplayCounterValue();
if (pb2State==0){
NewPosition = counter;
Serial.print("Move to Position: ");
Serial.println(NewPosition);
oledDisplayMoveToText();

Anything that starts with oled.........

Are seperate routines (something else I've needed to learn ) for displaying values on the oled display

The versions are still having effects on the program especially the enable/disable outputs,

pinMode(StepperEnable,OUTPUT); //Set pin D6 for output
digitalWrite(StepperEnable, HIGH); // Turn off stepper outputs until needed

Everytime I try and use this from the original program, I get to use it once as either HIGH (disable) or LOW (enable), but then I cant change the value so the stepper out is either permanently off or on.
There are newer commands

stepper1.disableOutputs();

but I can't get it to work either.

However I now have a working traverser

and the dongle looks like this

Just need design and print a cover, oh and make another for the other traverser.

Paul

 

To refine the ugly cable connection on the left, I have just ordered up these



6 x ethernet sockets, as I used an ethernet cable to create existing cable extension cable for the dongle it should do.
I will need to bell out the connectons to make sure the wire hookups align, but hey I have a crimping tool so can make up a custom cable if required.

Paul

 

That's a common problem - often exacerbated by Toto's curry selection.

The bit I like about the DCC control, is when there are only 2 of you operating, the fiddle yard(s) can be setup with 3 trains each, and one person can operate the layout on thier own.

The MK2 traverser, just powered it up and worked first time - I'm as shocked as anyone.

Paul

 

This old chestnut again

Mods carried out to the controller

But it needed some form of protection - another Fusion project I feel

A rough mockup of the circuit board was created - i'm sure I could have found the actual components - but that would have been too sensible.



The base was created



And without the circuit board



And finally the lid



The buttons are separate items and sit within a square recess on the underside of the lid, to stop them rotating





The button just rests on the push button



Just need to slice the models and do a test print

Paul

 

And sliced and ready to print



The round bit is a RJ45 (8P8C - ethernet) keystone faceplate which will be fitted to the baseboard side. Allowing the controller to be disconnected. For exhibition use I should make a second controller as a spare.

Will have to warm up the printers and resin, so maybe later.

Paul

 

3 hours print time - so fourish, hopefully Amazon will have delivered the RJ45 keystones by then so assembly and testing can take place.

And I can create some space - a 4ft 6in x 2ft lump on the workbench don't 'alf cramp cramp your style

Paul

 

Estimated print time of 2h 30m, actual time 3h 7m

Couple of issues - support failiures on the corner of the ethernet port and bottom right hand corner also on the top plate.

Cured and cleaned up. A taper reamer opented up the push button holes, and the rough edge of the pcb cleaned up. Now a snug fit

Assembled

The screw threads were modeled in fusion and printed without issue - worked without having to run a M3 tap through.

Cable even fits

Just need to get the drill out to recess the mounting plate for the RJ45 keystone connector, in the side of the traverser.

Due to the layout of the traverser support table I will need to trim the top 5mm off the mounting plate.

You will see later .......

Paul

 

Hi Jim

We get expansion and shrinkage on the same print

Holes tend to print undersize 0.6mm is probably the smallest vertical through hole I'd try and print, if you are slightly over-exposing the resin, then the hole will tend to fill up.
Horizontal holes can be printed smaller.

It's all about balancing the exposure - which is effected by temp, the resin likes to be between 25 & 30 C, below 20 and the viscosity really starts to change. Some of the newer high speed resins are very runny, so may be better suited to printing in colder climes.

Over exposing the print blooms beyond the edge so grows in size.
Under exposing the print gives more acurate prints but they're more likely to delaminate or just fall off.
Now we are talking 10ths of a second difference, throw in varying temps then most of us lean towards a little over exposure.

On the plus side, if making masters for casting, and you know the shrinkage values for the mould making and casting processes you can scale the print accordingly, far easier than having to sculpt a new master.

Regarding the threads created in fusion - be careful there are two variations



You need the Modeled selected otherwise its a non-printable texture.
As I was tapping into a blind hole I also selected Full length.

Paul

 

I normally would print a hex hole 0.2mm oversize and glue a 6BA/8BA or M3/M4 nut in place
But recently bought some of the brass inserts, unlike with ABS or PLA I can't heat them and push in, I'll be using super glue or epoxy to glue them in.

Paul