Servo Research

This week I completed a project that I probably started about 10 years ago.  This was when I made a couple of HO Scale 6.5m x 6.5m garages in styrene, but these models were fitted with a garage door that could open and close.  These were tilt-a-door models.  I included a servo at the back of a garage, hidden by a wall and then ran a cotton connection from the servo to the tilt-a-door mechanism.  So these models just sat on the shelf in the train shed for a very long time.  

So this week, I bit the bullet and started to get into working on servo control from an Arduino.  I first got a servo working directly from an Arduino UNO.  So I then connected the Arduino to my garage Arduino and after adjusting the throw of the hidden Arduino, I could get the tilt-a-door on the garage model going up and down.  Next step was to reconfigure the wiring to/from the Arduino so that I could install a push button inline so that would control the opening and closing of the garage door via a press of the push button.  So some quick changes to the Arduino code and it was working.  The code would wait until a press on a certain pin in the Arduino – in my case pin 8 and the door would open.  It would then wait again for another button press on the same pin and the garage door would close.  I took this working garage door opening model to yesterday's ‘Show and Tell’ at our local NMRA meeting.  I got a couple of nods from the audience.  I was happy.

However, during the week, the main task that I was trying to master was to incorporate servo control from the Arduino via a PCA9685 16 port expansion module.  I tried it about 50 times.  I could not work out why this was not working.  I was bamboozled.  So on Friday I rang up my Arduino mentor - Bob, and asked if he could bring a spare PCA9685 16 port expansion module from his collection to Saturday’s NMRA Meeting.  I thought that the only reason I could not get the servo to work via the expansion port, is that my expansion module must have been bung.  If I proved that by using Bob’s spare module that it worked, then I'd order a few more of the 16-port expansion modules online to replace my bad one and then have a spare nearby.  So on Saturday morning I was sitting in the audience of the NMRA meeting with my PCA9685 port expander sitting on my lap, and I just happened to read the bottom side of the PCA9685 expansion module.  It said that if using servos, then the power had to be supplied via the V+ pin.  Oh firetruck!  I had not powered the V+ pin on the PCA9685 expansion module.  Now I have 7 of these expansion modules working in the internal logic of my Signalman's Panel.  But these are used only to control LEDs.  However, all of these were powered by the Alternate Power plugs on the module.  This powered the V+ pin, but doesn’t use the V+ pin.  My testing with LEDs on the expansion module in question did not need the alternate power plugs used, but the LEDs still worked on the module.

The message above 'Use V+ for servo supply'.  It is quite clear.

So at half time in the Brisbane Lions and Hawthorn AFL game on TV later that evening, I reloaded the code into my Arduino to run the servo via the PCA9685 16 port expansion module, but this time with an extra cable from the 5V pin on the Arduino to the V+ pin on the module.  Guess what?  The bloody thing worked.  Damn Dork!  I did advise all the attendees at the NMRA meeting that I might have worked out why my Arduino testing was having an issue controlling the servo via the expansion port module.  I did get a few laughs.  I also got a few jokes throughout the meeting about powering the V+ pin.  We really need to separate Clinton, Greg and PK at future NMRA meetings.  Note to Div 1 Superintendent – separate those bozos!  When I told them that, they all laughed.  Given that one of them is a teacher you would think he would know better.

So now I just need to get some MERG servo components to fit underneath another test module and I will incorporate a servo that will actually change a point on this new test module, controlled and interlocked from my Signalman's Panel.  Hopefully my mate has placed the order from MERG in the UK.  I’d expect that in 2 or 3 weeks it might arrive.

I have an Operations Session scheduled for this coming Sunday, so this week I will be cleaning up the train shed and cleaning the track and testing everything in preparation.

Interlocked Points Control

So this week I went to Jaycar and picked up some 3 pin bi-colour (actually they are tri-colour if you power both sides at once) LEDs that I intended to replace my 2 pin – bi-colour LED’s that I already have installed on my Signalman’s Panel for the point indicators.  Jaycar had ones that were common cathode, which was what I wanted.  I set up a test rig on an Arduino UNO running through a 16-Port Port Extender (PCA9685).  I had it working as I thought it needed to be and I was happy.  Then I had another think.  So I then rewired it and tested it and I was even happier.  I halved the number of ports used on my PCA9685 by putting two of the 3 pin Bi-Colour LEDs back to back, instead of powering each pin separately.  The issue is that when I put two of these 3 pin LED back to back, the colours of the Red and Green LEDs are at different intensities.  So I did a bit of playing around and I needed about a 47 ohm resistor on the Red output and the intensity of both outputs was then the same.  So on Friday I started to install the new LED’s into the Signalman’s Panel and remove the old LEDs.  I did the first one and I could control the actuation of the LEDs by the panel mounted switch.  I then did the second one and the third switches and they would not change.  Then I realised that I had not uploaded the new version of code to make the points other than Point number 4 toggle correctly.  So I uploaded the new code to the Signalman’s panel.  Then I did a new test version of the LEDs as I had a crossover on the future layout which this time needed three LEDs to work together.  Two were for the parallel lines and the third was for the crossover toute.  The code allowed for the points to not be changed if the signals were not previously set to stop.  So now I have fully interlocked points on the future layout.

I had a small test jig for the 3 x bi-colour LED’s and the operation worked on the test jig.  It was plugged into the Signalman’s Panel and while I was testing other things, I heard a pop.  Oh No!  It seemed that I had blown up my test rig.  I have no idea what caused it.

So I worked out again how I would wire three bi-colour LEDs again and directly put them into the Signalman’s Panel and gave it a test and it worked.  I then did a full test of the Signalman’s control panel.  I set up various routes and now if you change a point and the signals are not already set to Red, then the points on the layout and the Signalman’s panel do not change.  The Signalman gets a message on the display either that the points was successfully changed, or that it was not changed, as he had not set the signals to Stop first.  So this is working really well.  I can set up all the routes and when I reset the signals, I can change the points back and I can get everything to go back to automatic mode.  I am very happy.

On Saturday it was the AGM of my Model Railway Club.  I nominated for a position on the Executive – General COMMITTEE MEMBER, for the first time in about 25 years.  I actually got selected.  There were 4 positions in this category and 6 applicants.  I was amazed that one member missed out for the same position.  The numbers of votes for each candidate was revealed this arvo.  63 people voted either in person or via proxy, and I got 62 votes.  Not bad I suppose.  But I wonder why PK did not for me?  While at the Club I received an order for a few styrene industrial bins.  I will make some at the upcoming RMCQ Exhibition in 4 weeks’ time.  I might as well make a few more – maybe 10 total.

Before lunch today I went down to the shed and started to do a bit of a clean-up.  I have an Ops Session in 2 weeks, so I need to move a lot of things from the floor and from the North Coast Control’s desk.  I also did a bit of work on a fettler’s shed that Paul provided me with last week.   I have another to clean-up and put together.  I had one at the Club yesterday and received a few requests from members for their own copy of the fettler’s shed.  I might have to ask Paul if he can print some more.  I have since found out that these models are Walker Models 3D prints and cost $15 to download.  So we will not be distributing copies to anyone.  They can get their own.  

Today after lunch I spend some time refining my presentation notes for the Armidale Convention.  They are basically complete now.  I might add an extra photo or two to the document, but that is it.  This coming week I will be working on actual servo motor control from my Arduino and my Signalman's Panel.

This Week's Review

This week I went on a bus, train and tram trip to Burleigh Heads on the Gold Coast for something to do.  Total trip $1.00 due to our $0.50 fares.  I chose Wednesday to go, as I thought I’d get the most mates available to come share the experience.  But as it turned out, it was just Dean and myself who attended after 16 people were asked.  They had varying excuses.  Some I can actually agree with.  Anyway, it was a great day.  I walked just under 10km that day, including around the hill at Burleigh.  Boy was there people around, on the train, on the tram and on the beach and stacks also walking around the National Park.

Burleigh Heads looking towards the hill.

Burleigh looking up towards North Burleigh and onwards to Surfers Paradise.

Saturday was a working bee day on the FreeMo modules.  We had about 6 of us there working on them.  Bit by bit we are getting them finished.  While taking a break yesterday I was shown how Paul controls his servo point motors on his layout.  I think I will use the same MERG servo mounts and then interface the servos with my Signalman’s Panel for control of the servos.  I’m thinking of also having up to 4 catch-points also controlled with servos.

MERG based Servo controls for point motors.

As today was Father’s Day, I got my combined Birthday and Father’s Day present (which I purchased myself) from the kids.  It was a Model Box BCC bus kit in Lego type blocks.  This is a fantastic model.  My only complaint is that my local bus service number and destination (the most populous Buzz service in Brisbane) or even the City Destination board, were not in the list of destinations available.  I might have to reach out to the guys and see if they can make up those destinations.  But apart from that my present is completed and I’m very happy with it.  Did I say it was a great little model?

The completed bus.

The rear.

During the week, I was confirmed as a presenter at this Year’s Armidale Convention.  My topic is Automated Signals for the Layout.  My presentation is completed, but I need to spend some time on the notes for the convention handout.  I know I started some somewhere, but I cannot find them.  I will keep looking.

More Work on the Signalman's Panel Occurs

This week I quickly ticked off the few tasks on the To Do list.  This included reversing the two wires for the Branch shunt ahead wires.  I also wired up the positive and negative supplies to the new bars inside the panel for the IR (Infra-Red) detectors connected via the rear of the panel.  I also fixed the Green aspect on signal D3, which had a wire come out of the socket inside the panel.  I did discover a new issue in that one of the I2C boards with all the white IR detector LEDs for the Dual Gauge (on a separate Expansion Port on the bus) was no longer connected to the I2C bus.  I traced this back to another plug coming adrift from my distribution bus for the 2 wires on the I2C bus.  Once I identified the cause, and rectified this, the panel all lit up correctly again.

The next job was to pre-install the 2 plugs into the Signalman's Panel distribution plugs, for each of the next 8 signals to be installed.  These are the Up and Down Dual gauge signals.  The 9^th Signal and hence pair of plugs are for the shunt ahead signal on one of the Up Signals DG2.  I then had to wait until my RJ45 CAT6e sockets come in at my local Jaycar store.

On Friday I went to Jaycar to look for a variable power supply, and on the off chance that my order for 10 x RJ45 Cat6e sockets had come in.  The great service from the staff there confirmed that the items had come in and the staff member went looking.  So I quickly connected all of the 7 wires in each of the 8 of the 10 x RJ45 sockets, and 2 wires for the 9th socket for the Shunt ahead signal.  However, this time I successfully installed the wires the correct way around.  I then threaded the wires in through the back of the Signalman's Panel and they are now ready to be connected up and then tested.

Saturday was spent at the Redlands Model Train Show and I worked on the NMRA stand with 3 other clinicians.  I must write myself a note, never leave the stand as those modelling mates (aka b@$tards) left at the stand, just print “for sale” signs with 75% off and leave them on my items that I am building.  During the day, I built one window, installed another, constructed and installed 5 doors, added a couple of roof sections and constructed a concrete slab and a veranda awning on one end of my Old Cassino station building.  This building is coming along, but I think I still have another exhibition left in it to be completed.  Maybe by the RMCQ Show at Brendale in October it might be completed.

So today I wired up the first 4 x RJ45 Plugs for the UP Dual Gauge signals.  They all seem to work.  I have yet to connect up the IR detectors.  So I guess that means that tomorrow is the day for this task and then wiring the DOWN Dual gauge signals will be wired in.  I did notice that the 'Shunt Ahead' signal on my test track is not working. I recently moved this track and maybe I snagged something.  I connected the troublesome shunt ahead signal to the UP shunt, DOWN Shunt, and the Branch Shunt and it does not light for any of these signal aspects.  So I must have snapped a wire somewhere.  So I will investigate that tomorrow.  So the panel will soon be almost complete for signalling.  Focus will turn to how I operate my point motors.

I am starting to think ahead as to how to control the 5 Servo motors that I will use to control the mainline point motors on the layout.  Once this method is worked out, I will look at how I control up to 4 catch points on the layout as well.  I am not sure if I will use more RJ45 cables, or some other method for these items.  I will also have to devise a small emergency panel to control just the points (and maybe the catch points) on the layout if I have a catastrophic failure of the Signalman’s Panel when operating the future layout, and I no longer have control of the points from that item.  All the points in the Sidings on the front of the layout will be manual.

Further Work on Signlman's Panel

This week I had a set list of tasks to complete on my Signalman’s panel.  Last week I punched in the first 6 x RJ45 sockets for the DOWN tracks.  I installed them into my 6 way socket holders at the back of the panel.  Next job was to test these.  As I wired these to the panel, I realised that I had wired all the wires to the rear of the Rj45 sockets back to front.  Whoops.  The wires that should have been on the left-hand side of the patch were on the right-hand side and vice versa.  So I had to remove the 6 x RJ45 Sockets from the rear of the Signalman’s Panel.  I then had to unpunch the wires in the sockets, and then re-punch them in the correct side.  Soon I was back to where I should have been when I started.   I then tested all the plugs with the test track and everything was working.  So today I installed the 4 x Branch Line RJ45 plugs and tested them.  By the way, these were also wired incorrectly last week.  The sockets have all signal aspects working, except for the Shunt Ahead Aspect on signal B1.  I’m now worried that I connected the two wires on this socket back to front.  The Shunt Ahead Aspect on the Signalman’s panel works correctly.  That should be an easy fix next week.  I have also set up a new Positive and Negative line inside the Signalman’s Panel for the 11 x IR detectors on the second slave Arduino Uno.  These positive and negative lines are not yet wired back to the Arduino.  But the first three IR detectors connected via the RJ45 plugs are wired to them.  These base power lines provide power for the 11 x IR detectors connected via the RJ45 plugs.  The temporary IR detectors that I use to test the panel are already wired and still working.  So I have not yet tested the 3 x IR detectors for the Branch Line.

I went to Jaycar this morning and ordered another 10 x RJ45 Sockets.  So when they rock up, I will install the 7 wires into each of these sockets (hopefully into the correct sides this time) and then put them into the rear of the Signalman’s Panel.  The wires then get fed through to the electronics inside the panel and everything gets terminated.  It then gets tested and hopefully it will all work first time, and this will close out the UP Dual Gauge signals and the DOWN Dual Gauge signals.

I did discover a logic bug when I get multiple consecutive IR sensors going off at the same time.  I know exactly where this occurs in the code and why it is occurring and how to fix it.  But I just need to find the time.

So while testing the new Branch Line connections I found that the Green aspect on DOWN Signal D3 is not working.  So I must have knocked something inside the Signalman’s Panel, as this was working last time I tested it.  But quite a lot of wire moves have occurred today.  I will investigate that failure to light that LED tomorrow.  Both the LED on the Signalman’s Panel and the signal on the test track when wired in, do not light.  So that gives me a good idea where to look.

I went to George’s shop during the week and replaced the bung IR reset detector on his level crossing display circuit.

This week I tried to fire up one of my home layout based level crossing detectors and it would not.  I never went back to it to determine where the power was not reaching.  But the Arduino has power but the 5 x IR detectors are not powering.  When I get some time next week I will fix this.  The other level crossing detector is working well.  I was planning on testing running the level crossing unit off track power.  So again next week I will rectified this.

Next weekend is the Redlands Club’s Hobby Show.  I will be there on the Saturday (at least) working on the NMRA stand.

I have now reached 1000 posts on my blog.

 

Working on the Test Track

This week I went to George’s Shop and did some repair work on the level crossing detector he has on his Shop shelf layout.  It is now working again after it was moved and some wires were then too short to reach the previous locations where the hole was in the baseboard for the IR detector to poke through the track.  But one IR detector had gone missing and another no longer worked.  I had replaced the missing detector and that exhausted my spare collection of detectors.  I ordered a few more on Sunday, and they arrived on Tuesday afternoon.  That will allow me to replace the no longer working ‘reset IR’ detector and throw that old bad one away.

On the home front, I got thinking that for two of my level crossing detectors currently run off  a USB power source.  On my recent home based Arduino projects, I have them running off my layout DCC track power, rectified the track power, filtering it and connecting it to my Arduinos.  So today I searched around and found two of my power supplies (rectifiers and filtering set ups), and soldered some wires to them ready to connect to DCC track power and then I found my plugs so I can connect the output power to the Arduino running my level crossings.  I will test this out next week to see if I can get enough power for teh Arduino and not take too much track power. 

I also did a bit of work on my Signalman’s Panel and test track.  I also traced back the non-working shunt ahead aspect of signal U2 on the test track.  It determined that the connections all worked at the signal.  They worked under the signal before the RJ45 patch point.  It all worked on the signalman’s panel.  Then I traced the wire from the rear of the signalman’s panel to the plug in the panel and it had separated.  So I resoldered it back on and the signal was back working again.

I decided to remove a few of the plugs that connected the 10 x IR detectors on the Signalman’s panel for the UP and DOWN signals.  I decided that I would solder these to strips of copper inside the Signalman’s panel and do away with the plugs.  These plugs gave less than satisfactory electrical connections when connected via the Arduino plugs into a distribution board.

On Friday I finally got around to laying track on my test track module.  But I did not check that the IR detectors would fit through the sleeper spacing before I packed it up and put it in the car for showing off at the Saturday NMRA meeting.  I was doing a presentation on weathering and talking about the Frosted Glass can of spray paint, but it appears that it is no longer available.  That stuff was great.  But we have it on good advice that a new product is on its way out of the same factory.  I also took my model of the future layout so attendees at the meeting could see what I was proposing to build.  I also took my test track and Signalman’s Panel and set them up on two blow-mould tables in a garage at the NMRA meeting.  However, even in the garage, the sunlight that came in was causing havoc with the IR detectors at one end of the Signalman’s Panel.  Eventually I got them working, by putting the garage door down.

While at the meeting I had a phone message from Jaycar that an order for some RJ45 sockets had come in.  So today I picked them up and I have also patched in the various wires to all the RJ plugs.  The first 6 plugs for the DOWN signals have been installed in the rear of the Signalman’s Panel and the wires fed through the rear of the Signalman’s Panel but the wires have not yet been terminated.  That is a task for the coming week.  I also punched the wires in the next terminating socket for the Branch line signals.   I will install them next week and terminate the wires inside the panel.

While at the NMRA meeting one signal decided to stop showing a red aspect.  The LED had come adrift from the common bus wire inside the third signal on the test track.  I am quite happy with the progress that is being made with the Signalman’s Panel.  

That was post number 999.

Testing the Test Track and Signalling Connections

Earlier this week I began wiring in the second cable connection from my signal U2, and the siding signal SS2.  I then did a bit of testing and some things worked and some did not.  It basically all came down to bad connections for my jumper cables in the Signalman’s Panel.  So these were adjusted, reseated and resoldered and most things are pretty good.  I then found that the two red signals on my extension U3 and U4 are not lighting.  So I will have to disconnect the signals and check the LEDs inside the signal.  I thought it is just a cold solder joint.  As suspected for U3 I had to resolder the red LED to the common in the test track.  For U4, I had to remove the red wire in my RJ45 socket and re-patch that wire.  All is now good. 

So at the moment everything seems to be working on the UP line testing, except that initially I had not tested the shunt ahead and routing aspects for signal U2.  So on Thursday I tested the shunt ahead aspect on signal U2 and it worked fine.  I then did some more testing and I was having an issue with the route indicators on signal U2 on the test board.  I jiggled a few wires around inside the box and then rebooted and all is now good.  However, I am seeing some strange aspects occurring when a random IR detector goes off out of sequence.  I will look into this and determine if I need to make the logic aware of this.  I am currently testing the Signalman’s panel and the test board out in the pergola and I am currently surrounded by lots of sunlight, so this is playing havoc with my IR detectors.  An example of this for example is if U3 and U4 are green and then U4’s IR detector goes off.  U3 stays green and the driver is then confronted with a red signal without first getting a yellow aspect.  On other signals this is handled as both the U1 and U2 signals have Hold switches on the panel so the signalman can set the signal to a Red Aspect to do something ahead, like set points etc.  However, no such logic appears for U3.  It will be easy to incorporate by I wonder if I really need to do this.  Maybe if someone is cleaning the track and sets off an IR detector?  But all the other signals can work themselves out of any predicament given the logic I have implemented on them.

I measured my test track and wouldn’t you know it, it seems about 1 inch to long to fit behind the back seat in my car.  That was assessed through guesstimation, as I have not physically tried it yet.  But on Friday afternoon when I did try and put it in the car, I could just get it to fit behind the rear seat headrests.  So I took my Signalman’s Panel and the test track to the Club on Saturday and after the monthly meeting I set it up on two blow mold tables.  I fired the system up and discussed how it worked for some interested passersby.  But at the end after all the opening and closing of the panel lid, one of the wires for a point control lost its wire.  So I could no longer route trains to the branch or the Down Main.  So on Sunday morning I fixed the broken wire and everything was working again.  I did notice that I needed to name the Points and the Catch Points on the signalman’s panel so when the messages come up on the panel, the signalman will be able to understand what is happening where.  So I have now printed off the new name tags, and fill affix them next week.  Another task I did today was to verify the resistance of the resistor on the Green aspect in Signal U1.  I removed it and measured it at 1600 Ohms.  It was supposed to be a 1K ohms resistor.  I don't know how I got that wrong.  So I replaced it with a 1K ohm resistor and now that signal is a bit brighter both on the Signalman's panel and on the test track.

I also took my level crossing test track to the Club and showed this off to the Club’s electrical guru following a discussion we had the previous month.  He wants to implement a couple of level crossing flashing lights in various locations on the HO Clubroom layout.  So he wants to borrow my Arduino code.  I sent him a copy this morning.

While at the Club yesterday, my mate Greg presented me with a couple of 3D items that I asked him to design.  That was a tradie's tool box to go into the rear of a 4 Wheel drive tray or the tray of a ute.  The second item was a surfboard.  I will paint them this week and see if I can place them on my layout.  If they come out good on the layout, I will place a bulk order for these and we will sell them at our Club's next Buy and Sell.  Another item that Greg 3D printed were the shunt ahead signal housing and the rabbit's ears (route indicator) for my U2 junction signal.  This afternoon I fed the LED's into their new homes.  But I must have knocked something as the Shunt ahead yellow aspect no longer lights on the signal.  But it does on the Signalman's panel, so it will be easy to trace back and fix the issue tomorrow morning.

I think I will take the Signalman's panel and the test track to the NMRA meeting next Saturday and show the guys there how the progress goes.  I am waiting on a bulk order at my local Jaycar which should turn up this week sometime, so I will be able to wire up the Down Signals and maybe the Branch Signals.

This actually is my post number 998 on the blog.  So in a couple of week's time I will hit the 1000 post mark.