A Ground Frame for a Plank!

[SHMD]

Now I had better decide on where I am going to place the ground frame on the wooden base.

(Both left and right, and front and back. Where is the front anyway??? I must decide on where the 4ft is going to be?)

 

 

Another decision is - which way up should I solder the PCB to the toggle switches?

I can solder the toggle switches, and other components, to either side but the 10way IDC connector must go on the silk screen side of the PCB. I also want the IDC connector to be accessible so that I can pull the ribbon cable out.

 

 

Getting closer...

 

 

Kev.

 

[LNERGE]

Quite inspirational. I wonder if just using the quadrant plate and microswitches without the lever extensions would pass as a miniature power lever frame like a Westinghouse L?

 

Edit to add..

 

I posted a link but it didn't seem to work so substituted a photo..

 

[SHMD]

Lots more progress today.

 

I've been so busy that the misses not only cooked lunch but brought it to my work bench!

Sometimes I really really REALLY love my wife.

 

 

Kev.

[SHMD]

Quite inspirational. I wonder if just using the quadrant plate and microswitches without the lever extensions would pass as a miniature power lever frame like a Westinghouse L?

 

Hi LNERGE,

 

The link didn't work (for me) but I googled the frame from your description.

There appears to be quite some travel on those levers.

 

The trouble with common electronic toggle switches is their small throw angle. About 15 degrees.

The actuators (toggles) can come in a variety of (short-ish) lengths and styles though.

 

 

Kev.

[SHMD]

For mounting the 'Quadrant plate' onto the wooden base the first thing I needed to do is mark out and drill the base plate.

This I will use as a template for all the holes required in mounting the frame with no additional marking out required and the holes are guaranteed to line up! It doesn't mean that it will be symmetrical though. (Just keep on marking 'F' and 'B' on everything and it will all work out!)

 

With the base plate drilled, the frame case was clamped to it and the 6 off 2mm holes drilled through.

 

Now a plank was removed to allow the frame to be added to the base.

 

I find a razor-saw very useful...

 

Next, the base plate was used to drill the holes in the wooden base.

After each hole was drilled an M2x8mm screw was dropped in to keep the alignment true.

 

The base plate, and levers, were a then transferred to underneath the wooden platform ready for the Quadrant plate to be added and the whole assembly bolted together.

(The slotted screw heads will be underneath the platform out of sigh.)

 

 

Kev.

 

[SHMD]

Here is the now mounted lever frame.

 

 

 

 

 

Kev.

 

[SHMD]

Next on the list was the lever's PCB - the TIB1.

 

I like clear schematics - even when there are so few components.

 

..and a few minutes later the finished board.

(Although is still does need to be soldered to the toggle switches.)

 

(Don't tell anyone but I only use 60/40 full fat solder!)

 

With the TIB mounted underneath, the whole platform will have to be raised. I am also thinking of adding a step too.

 

 

 

Kev.

(I might have to write a test routine this week!)

 

[SHMD]

I like it when you can combine your hobby with what you do at work.   (Ok only when some free moments became available...)

It just so happens that I play work with Scientific Instruments for a living, and there I was able to investigate the 'brass' failure earlier in the thread.

 

First, I prepared a sample.

The copper 'stub' contains 3 samples.

1, is a gold on silicon 'control' sample - on the left. (ok, it was just there and I left it!)

2, is the suspect 'brass' sample - the large piece on the right ("Br"), and

3, is just another 'random' piece of brass ("Br3") - at the bottom nearest the camera!

 

As XPS is a surface analysis technique I prepared the sample by removing 0.5mm of the sample with a file.

It may look clean but ant the molecular level it is truly contaminated. (Mainly with carbon but also with oxygen - especially with such reactive materials!)

The first two 'wide scans' (to 'see' what's there) are 1/BLACK - the initial scan, and 2/BLUE - the scan after cleaning the surface with an Ion Beam of argon atoms (singly charged and accelerated to 4.0keV)

After cleaning the true molecular make-up of the surface is revealed.

(The carbon peak @285eV and the Oxygen peaks @532eV are greatly reduced by 'etching' the sample, to allow the surface constituents to come to the for.)

 

Here I use the software to help identify the 'many constituents of the sample.

(Note that I have ignored the argon peak - and didn't even notice the Phosphorus peaks!)

 

This is the Spectrum of the failed sample. I have identified the regions of interest so that the computer can quantify the elements identified.

 

 

Our Computer said :-

 

                                     Quantification Report    "Br"                             
      /C=/data/Kev/101115.dset                                      Mon Nov 16 13:28:04 2015
 
                                                                                             
      State #0 : Etch Time     0.00 seconds                                                  
                                                                                             
      Peak                Type  Position   FWHM   Raw Area    RSF    Atomic   Atomic   Mass  
                                 BE (eV)   (eV)   (cps eV)            Mass    Conc %  Conc % 
                                                                                             
      Zn 2p               Reg   1021.000   2.932  1052959.1  5.589   65.387   21.18   19.94  
      Cu 2p               Reg    932.000   2.875  3278628.0  5.321   63.549   71.27   65.20  
      Sn 3d               Reg    494.000   3.450   328709.9  7.875  118.744    6.02   10.30  
      Pb 4f               Reg    138.000   6.128    81616.7  8.329  207.206    1.53    4.56  
                                                                                             
                                                                                             
         
                                     Quantification Report    "Br3"                             
      /C=/data/Kev/101115.dset                                      Mon Nov 16 13:33:38 2015
 
                                                                                             
      State #0 : Etch Time     0.00 seconds                                                  
                                                                                             
      Peak                Type  Position   FWHM   Raw Area    RSF    Atomic   Atomic   Mass  
                                 BE (eV)   (eV)   (cps eV)            Mass    Conc %  Conc % 
                                                                                             
      Zn 2p               Reg   1021.000   2.947   842186.6  5.589   65.387   19.00   18.78  
      Cu 2p               Reg    932.000   2.849  3155738.2  5.321   63.549   76.93   73.91  
      Sn 3d               Reg    494.000   2.957   197963.4  7.875  118.744    4.07    7.31  
                                                                                             
                                                                                         

Now, I am not paid for my analysis ability so these figure s are 'rough' but you can get an Idea that my 'brass' wasn't!

(There was plenty of oxygen still in the sample which I ignored.)

(There was a huge massive gigantic amount of carbon still in the sample - which I ignored!. This carbon did NOT reduce by anywhere near the same amount as the oxygen did during the atomic ion bombardment of the surface.)

 

 

Phosphor Bronze!

 

 

Kev.

 

 

[SHMD]

The first thing I did tonight was to replace the ACB1 prototype board with a production model.

(The LED has changed, I bent the programmer port pins so that they can be accessed whilst in the cabinet, and there is now a heat sink on the 5volt regulator!)

 

And the I wired up the whole thing!

I just plugged in the two ribbon cables.

(Which I managed to get wrong initially but the design/components are robust enough to survive!)

 

I then wrote (modified an existing) program (very simple) and applied power (via the programmer).

 

It all works! (Once I swapped the ribbon cables.)

I pull the lever, any lever, and the LED changes colour.  

 

OK, this 'quick' test program does not 'sense' the mid positions yet nor does it check and transmit the codes necessary to operate the points and isolation sections yet - but it does prove that all the electronics/wiring/build, and therefore the concept, are sound.   :)

 

It might not be obvious from this photo but the LEDs are actually way too bright!

 

 

Kev.

 

[SHMD]

OK, a couple of evenings of some fleeting work produced a fully working test program where the LEDs indicate green for home, nothing for mid-positions, and red for the off positions.

A simple program that allows testing and fault finding.

 

Unfortunately, I forgot which bit was which...

 

I had swapped both PCBs which still meant that bit 3 still mapped to bit 3 and bit4 and 5 and so on BUT I could not get the nearest LED and LEVER to work. Short of getting a meter on the job, I had tried everything. Until, 2 days later, I realised bit 0 was on the right and not on the left!

Now all working. Whew!

 

The test program does allow you to evaluate the operation of the levers though and they need work.

I either have to 'centre up' the toggle levers in the barrel connector (with a shim to centralise them) or I have to lengthen the slots that the levers move to and fro in.

I'll probably do both but I must be careful that the levers do 'hit home'. I might even have to add a 'C' shaped channel to the base that only allows the barrel connectors to align one way in operation.

 

I'll run some tests but as it stands, operating the frame is a joy. 

 

 

Kev.

 

 

 

[Barnaby]

very GOOD Kev and just in time for Christmas.

[SHMD]

Now the bit I have been dreading - Painting!

 

Some "Scotch Brite" soon made light work of cleaning up the outsides of the quadrant plate.

 

I then use Halfords Black Etch Primer to paint the metal work.

(I didn't watch the paint dry - honest!)

 

After the paint (had dried) I re-assembled the ground frame.

..and then tested it again.

 

Amazingly, the paint thickness was enough, (combined with the rotating and offset barrel connectors), to make the Home and Off positions become unreliable with some of the levers.

Now I really must centre-up the toggle-levers in the barrel connectors. (.and probably extent the slots too.)

 

 

Kev.

 

[SHMD]

I am now giving the Lever's colours some thought.

The first three are easy - glossy black for points, but the last two are more difficult.

 

Levers 4 and 5 will operate isolations sections on my Test Plank.

So what colours should I paint them?

Lever 4 will (in the Home position) allow movement onto the main straight.

Lever 4 (in the mid position) will isolate both the siding and the main straight's head shunt.

Lever 4 (in the Off position will allow movement onto the loop from the siding.

Lever 5 (in the Home position) will allow movement from the main onto either the loop or the straight.

Lever 5 (in the Off position) will isolate the main's head shunt.

 

Maybe I should treat Lever's 4 and 5 as if they are operating ground signals?

 

 

Kev.

[SHMD]

Here's what I am thinking...

 

 

Kev.

 

[SHMD]

More detail..

 

 

 

Hmm, but now I've drawn it I can see what would be 'logical' from where the Ground Frame would be.

 

Soooo.. I think:-

 

Lever 1 (BLACK) The point entering the loop from the Main,

Lever 2 (RED) The ground signal to allow movement onto the straight, or loop, (In reality the Isolated section.)

Lever 3 (BLACK) The head shunt point,

Lever 4 (BLACK) The siding point,

Lever 5 (RED) The 2 ground signals at the head shunt/siding end. (In reality the 2 isolating sections there.)

 

I'll draw it up.

 

 

Kev.

 

[SHMD]

Latest lever assignments.. ..but I'm still thinking.

 

 

 

(I don't really want to leave Lever5 in the mid position when the frame is locked out.)

 

 

Kev.

 

[RailWest]

I'm puzzled why you do not put the crossover points on the same lever

And in what way will be the GF be 'locked out' ?

[SHMD]

I'm puzzled why you do not put the crossover points on the same lever

 

 

OK, the crossover is now on the same lever.

 

 

 

And in what way will be the GF be 'locked out' ?

 

I always thought that a GF needed to be released from a Signalbox before any levers could be pulled.

 

 

Putting the crossover on one lever now frees up a lever so that each isolating section (and ground signal) can have its own lever. No need for any mid-positions now.

 

Here's the duly amended diagram.

Thanks Railwest.

 

 

Kev.

[RailWest]

>>>I always thought that a GF needed to be released from a Signalbox before any levers could be pulled......
 

Not necessarily, it depends upon the circumstances.

[SHMD]

One of the problems with this design is the fact that there is nothing to stop the levers from spinning the toggles around and around.

Because the barrel connectors have straight internal bores and the switches levers are conical then when clamping the barrel connector onto the toggle there is an angular offset from the toggle's centre line.

 

This doesn't matter for one lever but once there are 2, or more levers, then when the levers spin these offset angles to 'point' the levers in all directions!

You could continuously keep turning them back so that they all align again.

I had been doing this but quickly got fed up with it!

 

The solution was to make a 'comb' that stops the barrel connectors, and thus the toggles and levers, from spinning.

Again, this was made from a piece of that scrap 'brass'.

 

The comb must be pretty square, parallel and fit inside the Quadrant Plate.

 

Some close ups of the comb doing its job.

 

Note that there are three pillars that will hold the comb at the right height and orientation.

(The Pillars are not yet screwed in place.)

 

 

..and now the Quadrant Plate has some real 'workings' inside!

 

 

Kev.

 

[SHMD]

The next job was to drill the 3 holes for the tri kinematic mount of the Comb plate. 

 

Always use 3 mounting points, (instead of 4), because with 3 you are guaranteed to not have any wobble.

(If you are worried that this is not very professional then call it something posh!)

The S&T lads have been busy, not only in setting out their tools for the job in hand, but also in collecting the eggs for a very fresh omelette.

(This view even shows that the barrel connectors are all 'pointing' in slightly different directions. Imagine where the lever handles will be pointing after amplifying this error!)

 

A close up of the frame giving a glimpse of some of the inner workings.

 

The next 2 views show the 'troublesome levers' being kept well in line.

 

Even trying to 'move/twist/abuse' the levers, they now always return to the correct positions.

 

 

Anyone contemplating using barrel connectors to extend toggle switches really must incorporate a 'Comb' in the design. It is that important.

 

 

Kev.

(Note to self. Eggs roll off work benches! Doh!)

 

[TimC]

Kev

 

Just found this - absolutely first class!

[SHMD]

Now for a review of what's happened and, more importantly, what needs to be done.

 

 

Job 1

Thoroughly test the current build for repeatability and reliability.

 

Job 2

Finish the 'test' wiring so that power and coms are working.

(The 'test' wiring would be the 'finished' wiring but a bit longer to allow the final lengths to be determined 'on the job'!)

I would probably have a 'bare bones' program running here so that control of the 'Plank' would be possible.

(What better way to test everything!)

 

Job 3

The metal work is done but the levers need to be painted.

1. 
   

   Black {loop entry/exit};

   
   
   
2. 
   

   Red {loop entry signal};

   
   
   
3. 
   

   Black {cross over};

   
   
   
4. 
   

   Red {loop headshunt};

   
   
   
5. 
   

   Red {siding ground signal}.

   
   
   

Do I leave the handles 'brass' coloured or should I paint them a bare weathered iron colour?

Or should I paint the whole lever the assigned function/colour?

 

Job 4

The Platform needs to be raised and a step adding to the front, but this really needs to be done with the following thought in mind:-

 

The Instrument and the Electrical cabinet need to be fixed (somehow) to the platform.

Do I fix them with stripwoord or do I mount the lot on a baseboard and create a little diorama?

 

 

Job 5

All the wood needs to be painted/coloured/weathered.

This is the real worrying part for me. (Me and paint usually results in scrapped work.)

 

Job 6

The software needs to be finished for the ACB1 controller.

This is easy.

I already have the base routines and coms protocols (proven). I would just need to change the I/O assignments.

 

Some questions remain though:-

 

Do I add software interlocking?

ie A point can not be changed for a road already set/in use. A signal can not be pulled off if the point is set against it. Etc.

How do I 'feedback' this to the user, flashing LEDs in the Instrument panel?

 

Another question, (for the future perhaps), would be to only send signals to the Point and Signal Controller as a valid lever is pulled OR do I have the Point and Signal Controller 'polling' the ACB1 for valid lever positions? Both have advantages and both are interesting!

 

 

So, still lots to do and plenty of choices to come.

 

 

Kev.

 

[SHMD]

Well, the S&T lads have had the Christmas off but materials were dropped off ready for the PW gang to do work their magic...

 

 

 

 

 

 

 

 

I think I will have to make another quick sketch...

 

 

Kev.

[SHMD]

OK, the image below shows the major components and their location in the diorama.

 

(The intended size is 1ft by 2ft6 - as can be seen from the steel rules originating in the left hand front corner of the scene.)

 

 

Painting rails...

 

(Undercoat in this case - Halfords Bodyshop Grey Primer - placed on top of two (snapped) Ikea coffee stirrers.)

 

 

5" gauge says it all...

 

(Approximately 1/12th scale.)

 

 

A couple more general views...

 

 

(The eagle eyed amongst you may have noticed that the "Hershey's" bar has finally been removed from the scene - and very nice it was too!)

 

 

Kev.