A new Trainline GP9M

I picked up a Trainline GP9M at a show a few weekends ago. This makes #6 of this model. The model never was in production, so it represents a rebuilt version.

This model is Union Pacific 289. I had noted a broken handrail at the show, but it wasn't a big deal. I did not notice the handrail on the other side was missing and the front one was broken. Replacements would be needed, as this would keep even the most power-desperate railroad from using the locomotive.

The first step in converting any locomotive to DCC is to test the locomotive on DC. There was an unusual squealing from the trucks as the wheels turned. As the locomotive progressed around the test loop, it would derail at several points. These problems turned out to be caused by out of gauge wheels. I was really surprised the squeal was fixed by regauging the wheels.

To complete the DC tune up, I removed the wheels and cleaned them with a brass brush in a Dremel tool. This took just a few minutes, much less time than my first attempts at wheel cleaning.

Now that the locomotive runs well on DC, it's time to work on installing a DCC decoder.

Accurail Boxcars

I just completed a build of 6 Accurail boxcars. Overall, the process was not difficult and resulted in a nice looking box car when I was finished. There were a few steps along the way, however, that caused considerable difficulty. This post will cover the details of the build, both good and bad.

Building the cars is a process of installing the underframe parts and couplers, gluing the weight to the top of the underframe (to reduce warping), installing the trucks, and stuffing everything in to the shell.

Accurail uses a system of "half-moon" slots to indicate where the details are supposed to go. It's a simple process of cutting the part free from the sprue, installing the part, and adding a drop of CA from above to hold everything in place. Very nice.

The couplers are a different story. The included couplers were some form of dummy coupler, but the car easily accepted Kadee #5s with no modifications. However, Accurail uses a friction fit pin on the coupler box that makes it difficult to install the cover. If everything is not line up just so, the pin breaks and has to be replaced with a screw.

The trucks install with a similar friction pin. After promptly breaking the first two pins, I put the cars aside for a while until I could get screws to replace them. Fortunately a 2-56 is a direct replacement for those pins, and no drilling or tapping is necessary. (Unfortunately, the ones I got were flat head so the screwdriver kept on slipping. Next time it's Phillips or nothing.)

Before installing the wheelsets in the trucks, they were reamed out with a truck reaming tool. The tool only cost $10, but works wonders on trucks. The wheels roll much easier after a few seconds using the tool.

Once the undercarriage was assembled, it was time to install the shell. I had good success placing the undercarriage as high as possible against the most curved side of the shell and sliding a prybar^H^H^H^H^H^H^H screwdriver along so the shell covered the undercarriage. After a few seconds, everything would snap in to place and be solid.

With the build complete, it was time to make sure everything was up to standards. This involves weighing the car and adjusting coupler height.

It was wonderful to see that the cars weighed in at the NMRA RP weight. I usually wind up either trying to find something to add to the car or just dealing with the car being too light, so this was a welcome change.

Most of the cars required a washer to adjust one side or the other to fit the coupler height gauge. This requires removing the truck and installing the washer on the mounting screw. The pins included with the kit would have all broken by now if I hadn't used screws.

The Accurail boxcars produce good models, but beware the friction fit pins. Chances are good your car will need adjustment later and the pins will break (if they haven't already.) I was extremely pleased to see something meeting the NMRA weight RP, and am ready to see them running around the layout.

Update to version 0.8.1

Bug fixes:

Version 0.8.1
-----------
1. Removed ability to read goals and assists information from file.
2. Updated edit post code to include date of post.

Upgrade to version 0.8!

Change log

Version 0.8
-----------

1. "First" is now "previous" in the message navigation page.
2. "Return to Main Page" is displayed when in single-post (Direct Link) mode.
3. Temporary login update page directs the user to the login page if the log in fails.
4. An error message is presented when no posts are found.
5. New Feature! HTML in comments can be allowed.
6. New Feature! Anonymous Comments can be disabled.
7. New Feature! HTML on Main page can be disabled.
8. Option change page has been refactored.
9. Index page directs to install page if config file cannot be found.
10. Fixed error where a newly created database wouldn't be properly used by the software.
11. On successful install, the user is directed to log in and then taken to the admin page.
12. If the install fails, the install good message is not displayed.
13. If the config file is lost, running the install will create it again.
14. Fixed bug on initial post page where new category name wasn't detected.
15. Fixed case sensitive bug. ID was capitalized in the database, but not the code. (Honestly, what is it with these case-sensitive languages?)
16. Modified appearance of the main page so the post title is prominently displayed.
17. Refactored post comment page. Users shouldn't notice a thing. ;-)
18. The "View next 10 entries" message now indicates how many messages are left and does not display if on the last page.
19. Goals module displays list of goals with most recent first. (Admin only.)
20. Registration page displays error messages above form.
21. Comments display page now includes post comment form on page.
22. Subject field has been eliminated from comments. (They still exist in database, but that will be removed for version 1.0.)
23. Fixed error on post comment form where post text would be escaped, then when submitted properly escaped again.
24. New Feature! Admins can view comments posted by specific posters.
25. Post order on index page is based upon time rather than post ID.
26. Categories on post page are now displayed as a drop-down box, and now sorted by popularity. (Most common first, with single most common selected.)
27. New Feature! Future Posting!!!
28. Log in with disabled anonymous comments takes you back to the comment page.
29. Clarified that comments can still be edited or deleted by an administrator on the unregister page.
30. Fixed bug where < BR / > was displayed on blank registration form.
31. New Feature! Admins can now unregister users (given a specific username).

This post has been edited at 2010-07-02 04:35:00

Comment your test track

The test track is where new cars on the layout go first. They are upgraded to metal wheels, couplers checked for height and reliability, and weighed. Cars that don't meet the standards are left on the test track until repaired.

While this produces a fleet of good running cars for the layout, there's often several days or even a month or two between diagnosis of a problem and time to repair it. If only I had some way to remember what the problem was...

A time machine would certainly solve the problem, but I'd have to meet myself to ask what the original problem was. After meeting myself and asking me what the problem was, the past me might just fix the problem to avoid forgetting what it was in the first place. (Maybe I'll get lucky and need a part.) Fine, right? But if I fix the problem I won't need to travel back in time in the first place, so I won't fix the problem.

An index card with a description of the trouble (and maybe the results of a few common checks, like replacing Proto2000 gear axles) might just resolve the above paradox and solve the original problem of not knowing what was wrong. If the index card was stored with the car or placed nearby, the information I need would be there when I needed it.

Just like in programming, leaving yourself a note close to where you'll be working later can save you lots of time and aggravation. Plus, you won't have to mess with speeds above Warp 10, flux capacitors, or repair paradoxes.

American Flyer 303

I was given an old American Flyer set with old rusted track and the cars beat up quite a bit. The locomotive was an Atlantic type, number 303. It wouldn't run with the American Flyer transformer, and I thought something was wrong with the motor.

The motor was fine, but the reversing unit in the tender failed. It's a very common problem. There's a little tab which advances a drum to switch between the three modes: Forward, Neutral, and Reverse. The tab often gets stuck in the up position, and if the locomotive is in neutral mode the motor won't turn.

Once the reversing tab was moved back down, and the contacts on the reverser drum cleaned, the motor began to run. It had years of old and dried lubrication to defeat, but eventually did so. After a few minutes of running, it loosened up and started to run faster.

With a running motor, the next thing to do was check out the model. There's a smoke unit, and a chuff piston. The chuff piston pushes the smoke out of the smoke unit, synchronizing the exhaust to the motor speed. Cool!

An electromagnet is used in place of the traditional permanent magnet. This takes up a good amount of space, but in S scale they can afford it. The lack of a permanent magnet means that screws and screw drivers aren't attracted to the model when not running, so it's easier to work on.

The next step in restoring this locomotive would be to disassemble it again, clean all the old lube and apply new. It's probably going to be a shelf model, so I'll put that off for another day.

GP18 Decoder Installation

Click on the image for a larger view.

The HO scale Proto2000 GP18 was one of the earliest Proto2000 locomotives Life-Like made. They are not DCC ready out of the box, but not difficult to convert to DCC. The installation took a couple hours to complete. The high short nose provides plenty of space to cleanly mount a TCS T1 decoder.

This model is Norfolk and Western 2706. Norfolk and Western ran many of the early GP's long-nose forward, and the detail of the "F" near the platform on the long nose reflects this. There is no additional programing required to run long nose forward, the directional headlights will light properly in either direction.



This is the unmodified model.

The lights are 1.5V bulbs with a 6-diode voltage (and direction) regulator. Since the decoder provides light function, the diodes are not necessary and simply removed. The bulbs are under the electrical tape, and the 6-diode regulator is housed under the black plastic cover at the center of the frame.

A plastic plug holds the right hand rail power pickup wire in place. Attempting to remove this plug to use the decoder's red wire resulted in a damaged plug and no power pickup wire. A new connection was made by drilling and tapping for a #72 screw in the center of the weight. (Not shown.)



Opening the model involves removing the fuel tank shell and releasing the locking tabs that exist at the locations marked in green in the image above. My model had almost all the tabs broken off already, so gently holding the trucks and rocking the shell back and forth would release it.

Motor access requires removal of the weight. There are 4 screws (marked in yellow) holding the weight in place. Use caution when removing the weight, as the brass spring clip may come off with weight. This will cause the brush spring to be released and it may shoot off to some unknown location!

I like to run the decoder wires directly to the motor power clips. It is important to note that removal of these clips is necessary for soldering the motor leads without damaging the motor, and removal will release the springs and brushes. They will need to be set aside in a safe location and reinstalled after soldering new power leads.

It may be necessary to verify the motor is properly isolated. In my case, the Kapton tape on the bottom motor clip had begun to fail and needed to be replaced.



The bulbs were replaced with LEDs, using 820 ohm resistors. This image shows everything installed before the plastic cover is replaced and electrical tape is used to hold the LEDs in place.



Here is the finished installation. The plastic cover has been replaced and new electrical tape installed over the LEDs. The motor has been mounted in place with double stick tape.


This post has been edited at 2010-03-15 19:37:31

Remove single sheet of Homasote: 2 weeks

The club layout has been built with homasote screwed to plywood in places. While working on the grade reduction project, we needed to update some structure to hold new track. There was a little piece of the old structure tied to the homasote/plywood sandwich that needed to be removed.

Not wanting to waste a perfectly good sheet of homasote, we set about removing the screws from it. They had been hidden for years, so finding them was a trick. I got all the screws from my side, and took a look underneath. The screws penetrated the plywood and went all the way through! Now we could locate the screws!

That's the first time I've ever been happy the screw went through both materials.

The tip is grounded, forget working on live track

Tired of buying new soldering irons when the tip would go out in less than two years, I bought a really good quality one. The club layout had a bunch of feeders dropped for the track around the roundhouse, and it was time to hook them up.

Once the soldering iron was hot, I fluxed the joint and touched the tip to the live* bus wire. Sparks started to fly and I realized working on the bus wire with a grounded tip wasn't going to work. Good thing the club's iron isn't grounded.

*Yes, live. DCC track voltage is around 16V, with usually around .5-1A current draw. It's also very short circuit tolerant, so there's little risk working on live DCC track.

How deep is your roadbed? How long is your drill bit?

I had to drop a new feeder the other day, and came across a problem. The roadbed is a very deep 3 inches, and most drill bits that fit 16 gauge wire bottom out well before that. I'd either have to find a longer (and larger) drill bit, or find another method of making a hole.

The other method turned out to be a nail. With no hammer in sight, the nail had to be pushed through the homasote. The resulting hole was quite a distance away, leaving quite a bit of the feeder visible. The track now has power, but the wire is a little unsightly.

I should have gone for the larger drill bit. There's two things to learn from this story: Don't make your roadbed so deep a special drill bit is required to drop feeders and if you do make the roadbed that deep, you might as well go for the larger drill bit.