Rudy's Model Railway: a Folded Dogbone, HO, 3x3m

[RudyB]

While I was soldering the 52 DCC rail droppers to the track (a video and some info available in this blog post) I encountered two problems.

The first was a stupid mistake I made. I had an electrical short. Inspection of the track lead to the 3 North junctions ... I had forgotten to use isolator joiners in the middle of that junction street! Bummer ... now I needed to remove some track and correct that error ... only to find out that I was one isolator joiner short! That will need to wait until later this week.

The second issue is more severe. A test drive with a loco over some junction streets (that had all isolator joiners in place ) lead to some DCC interrupt problems with a few of the junctions. I use the switches straight out of the box, so without frog polarity switching.  It was not a cleaning issue with the fork contact point.

I have the idea this may be a mechanical issue with the Peco switch design. See this blog post for the details.

I'm curious if there are any Peco users that have encountered similar issues?

 

[RudyB]

A full track test drive.

 

The one missing isolator joiner has been acquired and has been placed and the final 10 DCC rail droppers have been soldered, completing the track. Time for a full track test run!

 

At the East junction street of Station South some loco light flickering is clearly visible. It is due to the bad point contact the Peco junctions make, which is a mechanical issue (not pollution). They are electrofrog junctions, but they are not (yet) wired that way. Mounting the servo drive helps, because that exercises an extra force on the point. Time will tell if this is a reliable solution.

 

When driving becomes frustrating because of stutter or halts at the junctions, I'll be forced to add the frog switching. Not a pleasant project, but not an impossible one since the track is only fixed to the board with some metal flower binding wire.

 

Next sub-project: mount 19 of the 23 servo drives (4 are already done).

 

Link to the test drive video on the blog.

 

 

 

..

[RudyB]

LiPo and DMM

Yesterday I didn’t feel like crawling under the table, where a lot of work awaits with mounting the servos. There still are some other half finished projects to work on ... coach lighting being one of them.

The wires were still sticking out, and also I did not like the use of 2 AAA batteries, since I would have to open the wagon to replace them when empty.

For the wires a couple of male and female standard gauge PCB pins came to rescue, they're used as a mini-connector that hangs vertically in between the wagons.

For the batteries I found a few 3.7V LiPo accu's. They are a replacement part for a quadcopter. These batteries can stay inside the wagon while charging them via the wires that will stick out somewhere.

While browsing for the batteries I came across a digital multimeter for just 10,-. Although I already have a DMM, this one had some functions that my other has not. I could not resist and ordered one. Can this be any good quality? I'll have to wait and see.

More info and some images on the blog.

[RudyB]

Corrosion and ATF.

 

Today I again made a few test drives over some junction streets. With ultra low speeds there's a lot of light flickering and loco stutter, or even complete stops. Not good.

 

Using a standard eraser gum I polished 4 junctions and just some 20 cm of track. The image shows the result ... terrible amount of corrosion and/or black dust came off  just these 4 junctions! (The visible top of the gum represents just 2 junctions, the other 2 had as many dirt at the invisible bottom side of the gum).

 

Apparently this is the amount of corrosion and/or black dirt that gathers while it lies in stock at the supplier, for an unknown period. Maybe when trains run around, corrosion build up is less because it is polished off by the wheels, and possibly also a polishing wagon? Anyhow ... the cleaning helped a lot. Alas will have to be repeated often I bet, after periods of no use.

 

I have read quite a bit about a 'wonder potion' called ATF (Automatic Transmission Fluid). Everyone who used it, is very positive. There also is a lot of scepticism, but interesting enough that comes only from people who never used it themselves.

 

I think when I am at the point that the servos are mounted and trains can really drive around ... I first will polish all track thoroughly and then apply some ATF. I'll post here about the results. Does anyone here have good or bad experiences with ATF? Or with other cleaning methods?

 

[RudyB]

Coach lighting: the LiPo batteries are built in

Past week the LiPo batteries plus a charger came in from Aliexpress. They were meant to be built in the coaches to power the led interior led light of 2 trains with each 3 coaches. Via a small connector the batteries can be charged from the outside.

More images and info is availableon the blog.

[RudyB]

Past weekend a lot of wiring work has been taken place, under the table.

 

A 5V 0.6A supply is used to power the Arduino's. Wires are spread to both the West and the East side of the table. A separate 5V 2A supply is used for the servo's. They are all over the place, the wiring to power them is a 3 node star 'network'. Both power supplies are old adapters I had in a drawer (with still another 12 or so, gathered over years of gadget collecting).

 

Originally I planned to have only one optocoupler circuit to feed DCC to the Arduino's (see the software page for details), but I could not get that to work. After a lot of testing and measuring and trying, it seemed the long GND wires, coming from the central node where the GND of both power supply's are connected, was too sensitive for interference.

 

The solution was to place the optocoupler circuit close to the Arduino. Which means I'll need 2 of those circuits. Luckily I initially bought 2 of the fast 7m136 OC's, no need for shopping. But also now, although it seems to work fine, it still is all quite sensitive. When I so much as touch a GND wire with my finger, the servo moves. I think I need to connect my 0V GND to a mains ground terminal ... but that is not available in the sleeping room where the lay out is built.

 

The video on the blog shows a bit of the current situation under the table, the optocoupler circuit and the Arduino, movement of the first actual layout-mounted servo, and the control of the servo via the ECoS and via PC (using the Dutch program Koploper).

 

[RudyB]

Servo Trouble

Past weekend I continued working on connecting 23 servo's to the 3 Arduino's that act as DCC servo decoders. Alas I ran into big trouble. I had connected and tested 3 servo's and up to that point all went well. But as soon as I connected the 4th servo, the first 3 started to aggressively jitter. So much even that the spring wires through the turnout holes came lose!

Not good! What was the causing this? Power supply too small? Electrical interference? Something weird with the Arduino? After a lot of puzzling, measuring and exchanging servo's it became clear that there were combinations of servo's that did not show this problem, and there were combinations of servo's that acted weird already as soon as only 2 were connected.

Although logically and rationally I could not understand it, slowly it became more and more apparent ... there are servo's that 'like each other', they work well together, but there are also ones that 'hate each other', they give aggressive jitter when attached to one and the same Arduino.

Closer investigation of the servo's showed that although they are the same brand and type ... they are different! See the image. When these 2 are attached to one Arduino, the one on the left starts to jitter and distort. I do not understand why. I also do not have a solution. Tried adding a capacitors to suppress distortion, added a resistor to the output to give it a less disturbance prone load ... nothing helped.

The only thing that seems to help is to use only one type of servo's on one Arduino. Alas that means I have to remove and then remount and retune 7 servos. Work I was not particularly waiting for. All and all a few days of problem seeking and a fair amount of frustration all this makes for a lost week.

[RudyB]

It works fine now.

The solution to exchange servo's and have only one type connected to one Arduino seems to work.

 

More info on this blog post.

[RudyB]

We're driving!

Finally all 23 turnout servo's are installed, both mechanically and electrically.

A video is available on the blog. It shows three trains simultaneously driving around the full layout.

All is still manually controlled. The ECoS screen is on the iPad, via a VNC client (Mocha VNC Lite), enabling 'remote control'. To safely drive multiple trains, automation is really needed. Manual control really asks for great concentration to take switch turnouts in time (or an unyielding short will be the result) and to halt trains in time at terminal tracks. Man ... it is hard work ... I was already sweating after three rounds.

With the servo's there have been 2 hurdles on the way.

First it seemed there were 2 types of servo's and when connected to one Arduino they act weird. In order to take care I had only one type of servo connected to one Arduino, 7 servo's needed to be exchanged. That meant 14 times disassemble and assemble and tune again. See this post for the details.

Then it turned out that the servo's make a little jump at power-up. All 23 of them jump together, which made that my 2A power supply would not start up. Although 2A would have been enough for normal operations, a heavier supply was needed. Luckily I could use the ATX power supply out of an old PC. See this post for the details.

Now all is well. All 23 turnouts work without any issues. It is fun to see that a fully functional DCC servo control can be made for less than €2,50 per turnout. Let's hope they continue to function for quite some time. Let's hope it turns out to be reliable over time.


 

[RudyB]

Three video's on working with routes with the ECoS are available on the blog.

 

The first video handles the definition of accessories and how we can use links to them on different control panels to switch that same accessory.

 

The second video handles how to make a drawing of the layout and to switch accessories from it.

 

The third video shows how to create and to use Routes.

 

 

 

 

[RudyB]

Switching routes in on the ECoS to manually operate a layout in essence works fine. There is a drawback however. Having just one ECoS screen available requires constant changes between the turnout / route switch panel and the train throttle panel.

 

It would help a lot if there would be a second screen. That is possible, using a PC with a Train Control software program. I plan on truing iTrain, JMRI, Rocrail and Koploper. iTrain seemed to have the most user friendly and intuitive layout editor, so that is the first one to give a try.

 

Two video's are available on the blog.

 

The first one shows the iTrain layout and routes editor. The second one puts it in practice driving 3 trains around.

 

[RudyB]

JMRI soute switching

 

After iTrain, now also JMRI has been put to the test for manual control of a layout using route switching.

 

Personally I like the iTrain user interface to create layouts better, it is much simpler an intuitive. On the other hand, JMRI has some nice possibilities for '######' your panel (you can use different icons / images, you can even use your own) and it has some more switching configuration possibilities. Drawback is that there is a bit more of a learning curve, but in the end it is doable.

 

I thought that the iTrain feature to lock all turnouts that are used in an active route would be nice to prevent errors. But in the end it proved cumbersome. When you need to be quick and switch different routes because a train is getting near, first having to deselect the old route(s) take away valuable time. With JMRI every route or junction can immediately be switched.

 

2 video's are available on the blog. One is about the JMRI layout editor and the creation of routes, the other shows 4 trains driving on the layout with JMRI route switching.

 

 

[RudyB]

Fun with Traincontroller

 

Past weekend I downloaded the Traincontroller demo and started playing with it. Immediately I got quite caught by it. ECoS, Koploper, iTrain and JMRI ... I tried them all for switch panel and routes ... well ... I can safely say: nothing beats TC! (OK ... that is a personal opinion ).

 

The 'switchboard editor' has a friendly user interface, the whole layout was entered in 'no time'. It was actually fun to work with. Yes ... to do more advanced things there is a learning curve and there is a 300 pages manual plus extensive Help files. But really ... anyone who has worked with similar software before can start up just like that.

 

But then the fun really starts. There are so many possible ways to control trains, from fully manual to fully automated and many variants in between ... and everything can be combined seamlessly. This makes playing with your layout so much fun! Really ... I am enthusiastic!

 

There is this one thing though that I am less enthusiastic about ... that is the price tag. TC Gold is out of my league for my tiny layout. Even TC Silver is. Bronze ... maybe, but then, once you've played with Gold it may be hard to deal with the Bronze constraints.

 

 

Anyhow ... on the blog there's a video that gives a short overview of how much fun TC is to play with.

 

 

 

[RudyB]

Video's on working with Traincontroller

 

After several days of playing with Traincontroller I can only say is that I'm having a lot of fun with the software. So much fun actually, that I plan to control my layout with it.

 

I though it might be an idea to do a series of video's on "Working with Traincontroller' while going through the process.

 

This is the first video in the series, it's about installing the software and then making the connection to the Command Station.

 

For ease of finding the video's I started a separate thread in the Special Interests - Computer Control section.

 

 

Link to Traincontroller video 01.

 

 

 

[RudyB]

Arduino S88 and first sensors mounted on the track

Progress on the layout hardware has been slow for a while ... I got a bit distracted with the Traincontroller video's.

This week however, the S88 bus wiring has been done and the first S88 Arduino is mounted under the table. The first 2 reed switches are fitted between the rails ... it is time for a first test drive!

More images and a video showing the S88 Arduino at work are available on the blog.

[RudyB]

Test Run With Traincontroller ‘Auto Train by Drag and Drop’

Besides working on Traincontroller videos (which is fun to do) the highest priority with the layout right now is the installation of the train sensors.

Well ... the past few days 60 holes have been drilled and today 30 reed switches have had their lead wires soldered on. That means all the preparations are done ... what is left now is to install the reeds and make the connections to the two S88 Arduino's.

A few sensors have been mounted and connected already, amongst which 2 sensors that are used for the speed profile measuring track. The speed profiles of all 4 engines have been measured. There will be a Traincontroller video on that later.

Once TC knows the engine speeds per DCC speed step, the brake distance accuracy, over different trains and with different block entry speeds, turned out to be within a +/- 1 cm (!). That is more than good enough to me. Based on these findings I decided to work with just one sensor per block.

Of course I could not resist to already have a little test run with automatic traffic   The video shows how that went, using 'Auto Train by Drag and Drop'. It also shows how accurate TC can work with just one sensor per block.

Link to the first automatic traffic testrun and accuracy video.

[RudyB]

All Sensors Mounted – Automatic Driving!

It has been a while since there was any progress worth mentioning. First there was a short holiday, then there was a period with temperatures too high to comfortably do any work under the table. And it was under table where all the final work was ... wiring 29 reed switches to the S88 Arduino's.

Past week it cooled down a bit and I crawled under the table to do the wiring and the necessary soldering. A new milestone is reached: all connections are done! Should be able to drive around fully automated now. Play time!

The video shows all trains drive happily around, all on a series of Schedules with successors.

Link to video RMR042: All Sensors Mounted – Automatic Driving!

[Snofi]

This is wonderful stuff. Like you, i'm really impressed with the possibilities computer control offers - bringing a new dimension to the idea of "watching trains run round" but with the added bonus of them doing something more realistic than simply cruising at whatever speed. I'm miles behind you in terms of progress - still struggling with the basics of layout constrction and getting too carried away with minor details rather than pushing forward with the overall plan...

[RudyB]

Tuning Traincontroller for Improved Throughput

Of course now that all the sensors are connected I can't get enough of playing with the trains.

Yesterday I tried to see if the 'throughput' could be improved some, such that trains follow up on each other a bit sooner. Not that such is anything near prototypical, rather the opposite, it was just an experiment.

There are a few settings in the Schedule Rules that can be changed to speed up the start of  trains that stand still because track ahead is reserved.

1. By default there is a 'Start Delay' after a green signal. This setting can be used to make things look more realistic when you have physical signals on your layout. On green, it takes a few seconds for the driver to react. This time is set to 2s by default. Since I do not have physical signals yet, I made this time 0. Trains start up 2s faster now than they originally did.

2. There are several choices in the Rules for the 'Release of Blocks and Routes'. By default this is set to release 'At Stop Marker'. With long blocks this means that the train must first reach the end of the block (where the stop marker usually is) before the junction street behind is released. I changed this to release 'Upon Complete Entry'. Now the junction street is released the moment the train has entered the block, well before it reaches the end.

This only works by the way if you specify train lengths. With that I smuggled a bit, filled in values 20 cm shorter than they actually are. That speeds up things even a bit more. I could also have reduced the scheduled stop times (now 10 seconds) but did not do that, it's just about increased throughput when trains are 'on the way'.

The video shows some 5 minutes of train traffic. With music. The effect of 'release upon complete entry' can clearly be seen at the end of the video, where the 2nd cargo train starts up well before the first reached its stop marker. And what to say on that stop position accuracy?


Link to video RMR043 - Tuning for Higher Throughput



.
 

[gwrcollet]

Hi Rudy

I have just started watching your video's Thanks for you for taking the time to

help all of us to get started with RR &CO, Some of your video's are hard to

find after no 16, have you done one about linking sound to RR &CO/

so they can sound horns or whistled on startup? If not are  you thinking

of doing one?

Many Thanks

gwrcollet

[RudyB]

Sugar Beets and Tail Lights

 

Although all my wagons came from trade shows and were overaged 'el cheapo' models, they still were 'new in box'. I'm far from a 'weathering artist', yet I do like the plastic look newness to be off. Time to tinker a bit with some cars.

 

A tank car got a few streaks of eye shadow. Just a minute of work and still a reasonable effect (OK ... looking from a distance that is ).

 

Two open cars got a load of 'sugar beets'. A grape stone is a reasonable HO representation of a suger beet. On the the week market for a kilo of grapes and eat!

 

Four closed wagons got a tail light. I was inspired by a fellow hobbyist ((see this Dutch forum post) and tried to make something similar. Well ... it was an attempt. Even though the size is not right and it looks not very neet and tidy, I still like the atmosphere when it drives around with the light on.

 

The result of this all can be seen on this blog post.

 

 

[RudyB]

Coach lighting with a DCDC converter

 

The 6 coaches that I have did get LED lighting built in early this year, but that was done, somewhat as a gadget, with Christmas lights. I still liked to change that into proper LED strips powered by a DCDC converter. The converter has an 'enable' input, it should be possible to switch the light on/off via an AUX output of the DCC decoder in the loco. That will be phase 2.

 

Expenses for three coaches: less than €2,50.

 

There are a few disadvantages still. The converter is too large to be completely hidden, it shows through the coach windows. Then, even though the converter is regulated and has two capacitors on board, there still is too much light flicker.

 

More details on the 'ins and outs' and some images are available on the blog:

 

Coach Lighting Version 2.

 

 

[RudyB]

Coach Lighting Version 3 is a Keeper!

I was not quite satisfied with the light flicker from the coach lighting version 2. A large capacitor could help some, but I had my doubts. It was more due to bad pickups. More theory on that on the blog.

The plan was to run a wire from the loco to the coach, to switch the light via pin 5 ('enable') of the DCDC converter. Well ... when there is already 1 wire, why not 2? Take the blue '+' power wire too and feed the LED's from the loc!?

The result is good ... very good even, I have no light flicker anymore, and there even is no capacitor in use. I'm content.

This is a link to the blog with more info, schematics, and a video.

Result:
+ Simplest schematics imaginable.
+ Price for 3 coaches: $1,- (3 feet of LED strip plus 1 resistor).
+ Lights can be switched on/off via a loco function.
+ No light flicker. No capacitor needed (on my layout, with this loco that is, may be situation dependent).
+ Nothing in sight inside the coaches.
+ Light output can be tuned via a decoder CV value (may depend on decoder type in use).
+ Light can have a subtle fade in/out effect, also via a CV value (may depend on decoder type in use).
– 2 wires run between the loco and the coach, with a connector in sight. Same connectors are used between the coaches.



 

[RudyB]

SMD LED lighting for loco

 

Past week a first step was made to get some LED lighting on the two Roco cargo loco''s I have. They have only one headlight, a light bulb (well, actually 2, one for each direction). On the buffers there's the 'suggestion' of lights. The red tail lights are just paint. The white headlight is brought there via a plastic light transmitter, but in daylight it can not be seen and in the dark it is only very faint.

 

100 (5 colors, 20pcs each) SMD LED's size 0603 were ordered at Aliexpress for just €1,10, free shipping.

 

They're just grains of sand. But they needed to be small in order to fit inside the lamp holders on the loco. With thanks to fellow club member Willy, who helped and who had all the right tools and materials like a mini soldering station and lacquered wire, the first 5 LED's were placed. It was a first test to see if it is doable and if it looks OK.

 

The result can be read on the blog, as well as a few images with the new lights on.

 

Link to the blog.

 

 

[RudyB]

Rail side signals at €0,05.

Adding signals to the layout would make it a bit more lively. I need 24.

I thought the commercially available signals a bit too pricey at around €25,-. 3D printed DIY variants are available but still too expensive to My liking. I went on the look out if it would be possible to make reasonably looking signals at a very low price using mostly scrap materials. And it was. An old umbrella being the main ingredient!

Actually the only 'out of pocket' costs were the SMD LED's and the resistors ... which added up to a whopping €0,05!

Admittedly the signals do not look as good as the ones that can be purchased and they're also not looking prototypical, yet ... they are good enough for my purposes. And at that price I can afford 24.

 

The result, with some more pictures and a description, can be fond on the blog via this link