Turnout Control Progress

I have mentioned previously that getting hand laid points connected up and suitably under control has been a stumbling block in past efforts. The achievable scope of Comstock Road (4 or 5 turnouts total) makes the mental size of the task easier to contemplate. I have begun the new year as I mean to go on, by tackling the mentally hard things and have made further progress.

First up is the connecting rod from switch stand location to throw bar. Increasingly prototypical possibilities have occupied my imagination but when I found myself contemplating scratchbuilding scale clevis’, I realized that I was making things harder than they should be, certainly for a first attempt. I resolved to make something out of the piano wire on hand.

I needed an eye or loop in the wire to connect to the vertical shaft comping up from beneath the layout. (I am going for a rotational motion like a switch stand rather than the model railroady back and forth in a big hole. I fashioned a simple jig consisting of a piece of scrap plywood with a nail driven in and cut off, and adjacent to a piano wire sized hole. A right angle bend near the end of the wire goes into the hole and the wire is wrapped around the nail to form the eye. I got the idea for this jig from the Animated Scale Models Handbook.

Here is the jig.bentwirejig

And here is the result trimmed up.eyeinwire

I have got the vertical brass tube and wire combo installed and connected to the throwbar. (We pause while I dash downstairs to take a photo of the installation which I apparently forgot to do. Lack of photos is usually a good sign since it indicates that I have got a head of steam up.) Here is a shot of the connecting rod installation. Bending the crank in the end of the vertical wire was a challenge and I will consider better alternatives such as soldering on a separate piece of brass bar. It does work and will be concealed by the switch stand. The other reason for a separate bar would be to allow the vertical wire to continue up through the stand so the target can rotate.connectingrod.jpg

Finally, we get to installing the servo, Tam Valley Octopus servo driver and associated electrical bits. I have got as far as fashioning a bracket for the servo using a section of 1/2″ aluminum channel from the big box store. I picked this idea up somewhere in the model railway reaches of the internet and it works a treat. The servo is just a friction fit in the channel after a slight pinch with a pair of pliers.bracketmk1

Weekend Reading: Animated Scale Models Handbook

animatedscalemodels.jpgWhile I was in Kingston, ON on Saturday, I was able to take a bit of time to visit the Kingston Railfair train show. As is my habit, I browsed through the used books because one can never have too many books! I was lucky enough to come across Animated Scale Models Handbook by Adolph F. Frank. It was inexpensively priced so I took it into custody out of curiosity.

I am pleased to say that I am not disappointed in my latest acquisition. I have not finished reading it yet I soon will. Published in 1981, Animated Scale Models describes methods, materials and mechanism for animation predating the advent of inexpensive microcontrollers, stepper motors and servos. While some of what is described has been superceded, much of the wisdom of creating mechanisms from simple materials still looks useful.

Materials and tools described are only the ordinary sort that can be easily obtained. While the tool list certainly does not include a lather, I am also certainly going to find things to do with mine in this book. And save money either way. One can buy pulleys from a hobby robotics supplier, for instance, but the cost can add up in a hurry. Using the techniques in this book, one can readily build inexpensive alternatives that are exactly what is needed.

The book itself is soft bound and printed on non-glossy paper. It is well illustrated with plenty of clear drawings but despite the blurb on the back cover, no photographs other than the one on the cover. No pretty pictures here, just the stuff you actually need. Chapters include basic components, speed reduction mechanisms, mechanical movements,  and various example projects including a ferris wheel, a factory with a bicycle assembly line and more prosaic things like a small house with a man swinging a hammer to repair the roof and grandma rocking her rocking chair on the porch.

I think I can safely assert that any model railroader could find something useful in the Animated Scale Models Handbook. I look forward to employing some of these techniques to liven up Comstock Road. While apparently no longer in print, the online book retailers seem to have multiple used copies available at very attractive prices if you are interested.

Jeweler Saw Practice

Having just finished the electronic part of installing a DCC decoder in Comstock Road’s lone Atlas O SW-8, I set out to finish the job by making sure everything was correctly insulated, taped down and able to fit inside the diecast metal shell. Perceptive readers might wonder why I mention the shell material. That is part of the fun!

Getting the wires all tucked in was not too much of a challenge but I then discovered that my estimation of clearance between speaker top and shell was, er, optimistic. The issue was the speaker mounting lugs cast into the inside of the shell. They are almost a fit for the chosen TCS speaker but only almost. I have also elected to not attach the speaker to the shell to give more clearance for a future detailed grill.

Here is what the inside looked like when I started.sw8mountlugs

Since the lugs had to go anyway, I resolved to “daylight” the opening back to something akin to the prototype rectangle. This would give plenty of room for the speaker to shoot upwards.

I got out my trusty jeweler’s saw and my excessive supply of #0 blades. Back in the day, I bought a gross of this blade size in a fit of enthusiasm and only afterwards discovered that #0 is too coarse for .015″ material. I have despaired of every finding a use for all those blades but no longer!

Even the relatively heavy #0 blades are fragile and do not tolerate careless use. You can turn corners in a cut but you have to carefully saw in place while rotating cautiously. If you push to hard you can jam the blade and snap it. Letting the work twist on the blade while repositioning can break a blade… You get the idea. Fortunately, I have about 12 dozen blades for just such an emergency. Or I did. I am now better at sawing and have less blades.expendovblades

Eventually, I got the opening cut out to my satisfaction. Here is the view from above with the shell on. (Apologies for the bad focus.)speakerhole

And here is the top view with the stock screens and grills back in place. All being black, the speaker is not visible unless you get up close and look straight.sw8postsurgery

DCC Install Complete-ish

Here is the full installation less wire tidying and any lighting connections. I don’t plan to hook up any of the lights because stripping the shell is in the near future so I don’t see the point.

This shot is very similar to the previous post except for the addition of the TCS speaker and LokSound PowerPack. Getting those soldered on was not any more difficult than the basic motor and pickup connections but my haste and ignorance did result in a bit of comedy.

AtlasODCCInstall

Firstly, I have hooked up the ESU CabControl system but I haven’t read anything but the Quick Start card nor have I operated with the system elsewhere often enough to know it. I can just about remember that the first three function keys are light, bell and horn. So, first heart stopping moment was when I applied track power after hooking up the speaker and got… nothing, nichts, nada. (no smoke, either). But then I tested the controls and the loco moved so no fried decoder. Eventually, I figured out that I have to F8 to “start” the loco and the room filled with the glorious sound of a 567 winding up. After noodling back and forth enjoying the sound, I was off to apply the keep alive capacitor.

The connections for the keep alive are on the side of the bottom board so I removed the decoder proper to get clear access to the relevant pads. Soldering the three connections was easy and I excitedly put the loco back on the track for the big test of the full system. And nothing at all. Not even movement. A bit of checking of documentation (does it need to charge or something?), connections (did I short something this late in the game?) and head scratching later I realized that I had a very good view of the soldered connections. Almost like the decoder wasn’t blocking the view. After plugging the decoder back into the board, all was well and the family was summoned for the big demonstration run.

I then did a bit of switching of my test tank car back and forth between two approach tracks and the traverser. I now have strong motivation to get the wiring sorted so more extensive running can be done. And also so I can move the DCC base unit off of its precarious perch on top of the layout.

In at the DCC Deep End

My big recent hobby related purchase was a DCC system. Since the incumbent control solution is an MRC Tech II 2500, moving to an ESU CabControl system skips about 30 years of progress. With the cart firmly in front of the horse, I set off to develop a decoder solution for Comstock Road’s singular non-DCC ready Atlas O SW-8. Which will also be my very first decoder installation.

After some research, I elected to keep giving ESU Loksound my business and went with the following:

  • LokSound L Select decoder (with recently available EMD 567CR sound file for correct 8-cylinder 567 sound!)
  • TSU large oval speaker that almost exactly fits the opening under the top grill.
  • LokSound PowerPack Maxi

One of the big advantages of standard gauge O scale is that there is buckets of room inside a diesel shell for you to put a decoder. No milling frames or faffing about trying to isolate the motor from the frame.

After playing hide and seek to figure out where I hid my sheet of .040″ styrene, I was off. Credit for the general approach goes to a post by “Bob, Curator of the A&O Historical Society” on the O Gauge Railroading forum.

Over the course of Sunday, I got the loco apart, the platform made, platform and decoder fixed in place and the leads for the track pickups and motor soldered. This is the minimum to actually run the loco so of course, I did.

Not pretty but there was a soothing lack of magic smoke containment failure and the DCC age dawned on Comstock Road as the shell-less chassis trundled back and forth on a traverser track. Sound, electrical resilience and wire management pending.

MarksFirstDecoder

DIY Jumper Wires

Anyone who has ever built a prototype circuit using a breadboard has encountered jumper wires. These wires have connectors on the ends which can fit into the holes on the board as well as the appropriate connectors on circuit boards. There are also male and female connectors (a jumper wire has a 1 wire version of these on each end). These connectors are usually referred to as type JR. You can purchase some premade but if you want anything odd like a 10 wire connector splitting to 6 and 4 connectors, you need to make your own. And of course, there is a crimper for that.

My plan for hooking up the control panel for Comstock Road’s traverser control panel involves the aforementioned split arrangement. This became the pretext to purchase yet another crimping tool along with the appropriate connector kits. And learning began…

I discovered that getting a good crimp was a bit of a challenge and it took a couple of YouTube video watchings and 3 out of 4 pins on the connector botched before I got it sorted. You cannot buy the pins separately so off I went to get another connector. Fortunately, the 4-wire connector is less than a dollar so the tuition is not high. If you get one of these crimpers, buy extra connectors for practice.

Here is the tool and the 4 and 10 pin connectors in progress.jrcrimp

LCD Standoff

One of the challenges I face in constructing the control system for the traverser automation is that I am using prototyping components. These are standalone boards used to mock up a system but not usually used in a “product”. Typically, one designs a custom printed circuit board that has all the components on it and does not require various boards wired together. Since I am only ever going to need one of these, I am just building it out of the prototyping components.

Which leads to the challenge of attaching the LCD board to the panel. I cut a suitable sized opening but I still needed to fix the board in place. I settled on using 3/8″ standoffs. These have threaded holes in each end so I could use the holes already provided on each board corner and the ones I drilled in the panel.

The fit in the opening isn’t perfect but about as good as I can get without some sort of bezel to cover the gap. I also had to cut down 8 screws to get since 3/8″ is less than 2×3/8″ and that is the shortest machine screws I could find.

lcd_standoffs

And here is the current panel front.

traverser_panel_front

Next step is some sort of wire harness for a detachable connection from the panel to the Arduino which will have to be a bit of distance away. Jumper wires and a breadboard are not a great permanent solution. 😛

A Couple Buttons Short of a Panel…

is not what I want on my tombstone. It does describe my progress this weekend. I planned, cut out, drilled and primed a control panel for the traverser. This seems early since I don’t have the prototype control system and firmware done but the quantity fo buttons and all flapping about on jumper wires would get a bit out of hand. By the time I tied it all down to a plank, I might as well just build the panel. The current plan is:

  • One button for each of the five traverser positions. Moving to a position will be just a single button press.
  • To make non-routine actions like testing and programming user friendly, I decided to splurge on a 4×20 LCD screen. (4 lines x 20 characters per line).
  • A sixth button for non-routine actions like panic stop and invoking a menu system.

Here is the panel with 4 out of sixth buttons installed and the LCD friction fitted in place. I need to pick up some standoffs to attach the LCD to the panel. I am mildly irritated to admit that I also need to pick up two more black pushbuttons. I am sure I have more but I couldn’t turn them up. And they come in pairs so there is at least one adrift somewhere.

panelpartial

The LCD is backlit so it is easy to read when powered. I have also got as far as a basic “Hello, World” hookup of LCD board to Arduino board via I2C. The coding is, as usual for Arduino, dead easy. The minor challenge was in figuring out which wire went to which pin. I will need to find an app for creating circuit schematics so I can document the hookup.

LCDHello

The lit LCD confuses my camera phone since this shot was taken in a lit room.

0.35mm Off the Middle

The lead screw kit I recently acquired for the traverser included an 8mmx6mm flexible coupler for joining 8mm lead screw to a stepper motor. I had hoped that the specs meant 6.35″ aka 1/4″ but no such luck. My choices where either to go and buy an 8×6.35 coupler from my local robotics store (yeah, being able to say that is kind of cheating) or drill out the 6mm hole to take a 1/4″ shaft. The couplers are less than $10CDN so it isn’t that big a deal but I decided that drilling with my lathe was a capability I needed to exercise.

Of course, I didn’t have an appropriate set of drill bits so out I went to the local big box store. $50 later, I was ready to save $10. Payback on the lathe is going to take a long time at this rate. 🙂 Seriously, though, I think a boring capability (for bigger holes) is the only thing I still need to acquire. From hereon in, it should be all down hill.

The actual setup and operation was refreshingly simple, I just removed the set screws from the coupler, chucked it in the self-centering three jaw chuck, chucked the bit in the Jacobs chuck in the tail stock, adjusted for distance and drilled away. It took no more than five minutes.

If this was all I ever needed to do, the lathe would obviously be overkill. On the other hand, having one available makes this sort of thing dead easy and it gives me satisfaction to get useful things done with it.

Here is the setup with the work just done:upsized

And here is the coupler installed on the stepper shaft.quarterinched

Going with a Lead Screw

leadscrewdelivered.jpgAfter much pondering (variations on that phrase get used a lot around here), I have decided to go with the simplest (relatively) solution for propelling the traverser, a single lead screw. It was always one of the options and after chatting with Trevor Marshall and Ryan Mendell at lunch, it was decided. Ryan deals professionally with assorted ad hoc projects and he tells me that lots of people try belt drives and fail to consider the tension required to get them to work properly. Since I have no desire to see if the traverser baseboard will bend under sustained tension, lead screws are the best alternative.

I conflate ball screw with lead screw in my thinking but I understand a lead screw is better for my applications since there is not requirement to handle heavy loads or avoid backlash of a couple of thousandths of an inch. My concern with this option was that aligning the screw, table and baseboard would be tricky and prone to binding. Ryan suggested attaching the nut (brass bit in the picture) so that it floats a bit. I thought about it and I have a plan to do that.

So, the alternatives considered:

  • single lead screw
  • single belt drive: two pulleys doing the same job as the lead screw but avoiding fine alignment challenges
  • two sychronized steppers with two lead screws or two belts attached to each end of the table.
  • one belt attached to each end of the table belts driven by pulleys mounted on a common shaft attached to a single stepper.
  • one belt routed so that it was attached at each end of the table. This is a simplified version of how printers are designed.

As you can see, I have been looking at all kinds of fun approaches. I still like the single last one but, besides the tensioning challenge, it represents a substantial increase in required hardware with an estimated 5 idler pulleys in addition to the driver pulley as well as roughly 8 feet of belt. None of that is especially expensive on a per unit basis but the total adds up. At that point, I am would be investing in additional entertainment value.

I ordered an 8mm lead screw, nut, flexible coupler and bearing combo of the required length (sadly not available locally as far as I could discover) and it arrived yesterday. Actual installation work will wait on vacation and the right mood.