The Interior

As mentioned in my last real post, adding an interior to the lighted building was now a necessity. The building is in a very visible location on the layout, and looking into the office will be easy.

The hardest part about adding an interior is simply deciding what to put in it. Desk, file cabinets, pop machine, boxes, pallets, what kind of junk accumulates in an office of a warehouse? Rather than fully detailing the office, I added only four items: A table, a man with a briefcase, a woman, and white bear.

The floor was painted a grey color, with the same color used on the lower part of the wall as a chair rail. This may not be noticed, but it might help complete the scene.

Plastic glazing over the windows improves the look of the building while tending to obscure some of the details inside. This means that the items inside do not need to be A++ grade models, but only need to be C-grade models. The table, for example, leans a little and was quickly assembled, but I doubt anyone is going to notice.

Looking in the big window, you see the table, the lady looking at the direction of the bear in the corner, and the bear. The man was going to be visible through the glass on the door, but a small mishap with glue obscures him.

At night, with the room lights out a nice glow comes from the office windows. When you get close, the interior is visible and the effect is excellent.

Update to Version 1.0!

Version 1.0
-----------
1. Display of main posts are now CSS instead of tables.
2. Post a new message page UI redesigned.
3. Goals update page displays error messages correctly now.
4. Index page uses CSS completely
5. Added code to support version 0.8's lack of timezone identifier. PHP may still complain, but at least the code will run.
6. Updated SQL for version 1.0 updates; refactored update file to allow future updates to be easier.
7. The reset button on the post comment page has been removed.
8. Refactored category edit/delete page
9. Added comments_allowed field to install code
10. Refactored add admin page.
11. Updated, cleaned a bit, and refactored comment control pages.
12. Eliminated processing pages for forms.
13. Refactored unregistration page.
14. Removed API functions display_error_message and display_status_message
15. New API function: MessageBox(2-3)
16. The entire system has had a face lift. All pages now use CSS to enhance appearance. (The real reason to use CSS, eh?)
17. Broke up functions page (users shouldn't notice a thing).
18. Consolidated new post and edit post pages.
19. New Feature! HTML on individual main posts can be turned on and off.
20. Comments can be disabled on a per-post basis.
21. Commenting cut off date can be disabled on a per-post basis.
22. Colors and Options are now separate in the database.
23. The only reset buttons that remain are those on the options and color change pages.

Lighting an office

Adding light to a building demands a higher level of attention to the building. On the layout, the building will attract attention, especially if onlookers can see into the building. For this added detail, however, a certain price must be paid.

Light leaks through anywhere it can. The plastics used in many model kits tends to be translucent, especially if a light source is located near to the material. Gaps between corners and materials need to be sealed with a sufficiently light-proof coating to prevent light from getting out.

One method used to address the translucence issue was simply to paint the building, both inside and out. Several coats of paint will eventually build up and prevent light from passing through. A method I have not tried would be to line the building with an adhesive metal tape. If only doing one room, this might be an ideal solution.

Sealing gaps tended to be relatively easy because they were small. The building was intended to be lit from the beginning, so careful attention was paid to gaps to ensure they were as small as possible. In the case of this building, one gap was sealed by gluing a piece of styrene between the adjacent wall panels, while others were closed by using a gap-filling CA.

For the most recent building, only a small office in the much larger warehouse was being lit. The whole building was prepared to be lit, in case the office light leaked. It did. The office was lit with two small white LEDs through the ceiling.

When an LED lights, the entire thing glows and can put off considerable light. It may be necessary to mask the back and sides of the LED to reduce the amount of light escaping. One corner of the building leaked, but masking the LED reduced the excess light enough that no further work was necessary.

Now that the office has lights, it's quite easy to see that there's nothing in there. An interior seems to be a necessary next step....

...and that's why you need the marker

As my train was departing the yard, I was looking for some kind of marker (flag, push pin, etc) to put on the end of the train. I grabbed a push pin off the bulletin board, and was immediately asked about it by one of the club's young members. Explaining that the marker indicated the end of train, rather than using a caboose, the push pin fell off and I could tell it wasn't going to stay on without a lot of help.

Since this is a model railroad, I just decided to go on without it. After finishing my switching, I returned to the yard with the empties I had pulled out and as the train became visible there was a new car on the front.

That was the car I tried to put the marker on. Guess I needed it anyway...

Block Detectors

I've been using transformer style block detectors to signal my railroad. The design I'm using says it can detect current as low as 1 mA, but in practice I've found 4-6 mA to be required. Now, I don't doubt the documentation, it's just my results differ. The transformer style of block detector is sensitive to things such as wire routing and number of turns, so it's easy to see why the results differ.

Now Team Digital has a transformer style block detector that looks like it uses the same basic design as the one I'm using. (I have not actually examined it.) The manual states a typical detection current of approximately 3 mA at 5V, which is similar to what I measured with the above circuit.

Connecting the block detectors is a matter of taking 22 ga magnet wire and wrapping it around the transformer 4 1/2 times inline with the power lead. Originally, I had used solid 18 gauge wire and wrapped it around 1 1/2 times. Detection current was somewhere around 10-20 mA, which isn't bad until you realize that equipping every car on a layout to be detected would soon result in even a small layout running out of power.

The magnet wire reduced the current required to approximately 5 mA, which means that 100 cars would draw 1/2 A. The trick is to get the block detector as sensitive as possible (to reduce current draw), while keeping it from detecting other things, such as a hand across the rails.

Trip current has been measured by clipping a 25K variable resistor to the rails and adjusting it to the point where the signal changes from green to red. At that point, the resistance is measured and Ohm's Law is use to calculate the trip current. After adding the magnet wire, my values usually fell in the 2.20 kOhm range. The track closest to the block detector measured 3.20 kOhm, which is only 3 mA of current. Perhaps better results are received when the block detector is close to the track rather than located on a control panel away from the track.

These values are important for choosing resistors for rolling stock. If a car has more resistance than the value, it will fail to be detected. If it has too much less resistance than that value, it will draw power needlessly. Some modelers use a 10K resistor for their cars, and report success. Two resistor equipped axles per car results in a draw of approximately 2 mA. My testing indicates that 2 such cars would need to be in the block to be detected. Equpping all 4 axles with a 10K resistor would result in a car that draws approximately 5 mA and should be detected.

Adjusting the resistor size to fit the block detector can reduce the amount of work necessary to equip cars for detection. As in the above example, equipping cars with 10K resistors results in a 2.5K resistance per car, something that could be done with 5K resistors on only two axles.

Some question remains in my mind whether only the caboose or last car needs to be detected. If the middle of the train is not detected, the signal behind the caboose would still be correct (while the one ahead of it would not.) However, in areas where the signal is protecting a switch, detection in the middle of the train may be essential.

Results differ based on the installation and what size resistor a railcar has been equipped with. Installing the block detectors first and then adjusting the wheelsets to match up with the detector's needs seems to be the best way to go.

Corrupted database tables

Half of my database tables got corrupted and there was no way to save them. I lost the tables containing the users and comments, so to comment you'll have to reregister. All the comments are missing as well.

It was right... something else was wrong

I was running power to a spur track, and carefully made sure the color code was followed, track power pickups worked, and then hooked up the bus wire.

Nothing.

Checked the connections at the panel, the bus and feeder connections, and then grabbed my meter. Power is getting to the track, but nothing is happening still. Clean the track and the locomotive's light illuminates.

Just because something's new doesn't mean that's where the problem is.

Upgrade to version 0.9!

Verision 0.9
-----------
1. Dates on the goals entry page are checked for correct form.
2. Future dates are prevented on goals entry page, as well as dates from more than one month ago.
3. Count of comments is shown on edit/delete posts/comments page.
4. Default timezone is set and the admin can change it by typing in a new name.
5. HTML in comments status is displayed on comments page.
6. Comments can be disabled (per post) by the administrator.
7. Edit/Delete post code has been merged.
8. No link is displayed to admin page on admin page. (A page should not link to itself.)
9. When posts are deleted, the comments on the post are deleted as well.
10. The last admin cannot unregister himself.
11. NULL color values are not allowed in the database.
12. Admin options and color change pages are now separate.
13. New Feature! Users can edit their comments.
14. Deleted debugging code that could be a security hole.
15. Added database update page.
16. Page descriptions are given in the page title
17. Admin find comments by user now remembers the username entered.
18. The rules are no longer displayed again upon successful registration.
19. Admins can change the name of the user posting the comment.
20. New Feature! Commenting on posts can be disabled X amount of days after the post is made.

More wires than I thought

The club layout is undergoing a huge renovation project. We're almost done with the main wiring phase. One of my tasks was to design the electrical system for signals and power distribution. Where the old section of layout had 1, maybe 2, power blocks, the new has six.

The main lines wound up having 3 sets of wires feeding the various points around the layout. The first 2 sets would be 1 were it not for the booster being located right in the middle of the block. The 3rd set is to run power from around the duck-under bridge, so there's no need to run power across (underneath) the bridge.

Because of splits like this, it turns out that my panel was completely underdesigned to handle all the connections that were required. The panel was designed to handle 6 sets of bus wires but will probably wind up supporting 14.

Fortunately, the mounting board was quite large and will be able to support the extra bus connections. There was even an autoreverser added for a reversing loop that I missed in my plans.

The installation of the panel in the middle of the bus rather than at the end created twice the connections as normally needed. Special wiring considerations added yet more connections, so the power distribution panel unintentionally became more complex and much fuller. A seemingly larger than necessary mounting board is handling the changes well.

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.