Applications

In the last few posts, the various uses and usage of a test light was discussed. To wrap up the series, we'll cover a couple uses that might not be obvious.

When installing block detection, it is sometimes useful to have a way to trip the detector without messing with a train. As long as the test light draws more current than the block detector needs to trip, it will trip the detector. Simply place the test light leads across the rails so the light lights.

Need to temporarily connect two rails? (i.e. To test if making a connection will fix the problem.) Simply connect the test light straight through just like it was a wire. The light may glow a little bit, but power will pass through. As more current is drawn, the light will glow brighter. (This is simply connecting a light bulb in series.)

There are likely many more non-obvious uses for the test light, but these are two I've found most useful. Additional uses can be suggested in the comments.

Reversing section diagnosis with a test light

Previously, the basic operation of a test light was discussed. In this post, the use of a test light will be employed to diagnose issues with a reversing section.

DCC reversing sections are placed between two sections of fixed-phase track. The phases are opposed, so a short would occur without the autoreverser. The autoreverser works by sensing a short and flipping phases to correct the problem. As long as this works, operators don't notice a thing. When this fails, trains will often work correctly until the crossing between the opposed fixed-phase track and the reversing section.

The test light can be used to indicate which phase the rails are connected to at the moment. Connect the test light to one rail in the fixed section and one rail in the reversing section. If the light lights, you know the rail is the opposite phase of the one connected to the test light. If the light doesn't light, it's probably a good idea to check the other rail just to make sure the connection is good.

To test the reverser, a momentary short must be created. If the test light you're using has a long probe, it can be used to bridge the rails to create a short. At this point, the reverser should flip polarity (check by repeating the test above.) Sometimes the reverser flips multiple times, or faster reversers (such as the PSX-AR) may see the short for too long and not change polarity. If this happens, another way to trip the reverser is to use the test light to bridge an insulating gap. A test light with a long probe can be used to bridge the gap.

If the reverser does not trip, start looking for the trouble in the end that's working fine. Most likely, something is closing or bridging the gap.

If needed, power can be strategically removed to help isolate the trouble area. By disconnecting power to the fixed section, then looking for current to flow from the reversing section to the fixed section, the rail that contains the short can be located. For example, if the test light is connected to Rail A and Rail B in the reversing section it should light. Keeping one end of the test light in the reversing section, move the other out into the fixed section. If the light lights, that rail is receiving power from somewhere.

Once the operation of a test light is understood, it can be a very useful tool in troubleshooting. Troubleshooting can be described as the process of eliminating what works properly in order to reveal what does not. The test light is an excellent tool for doing just that.

Checking phasing with the test light

One of the most important requirements of a multi-block layout is that the phasing (or polarity) is correct. Trains will short at block boundaries if this is not correct. In a multiple booster configuration, it is essential that each booster's output properly matches up with the adjacent blocks. Use of a test light can ensure this is the case.

The test light leads do not have to be connected to one booster for the light to work. This means the light will work across booster boundaries, which will be the basis of checking for proper phasing.

To check for proper phasing, connect the light across the rails in the same block. The light should light. Next, move one connection over to the other side of the gap to the same rail. Again, the light should light. If the light does not light, move the connection over to the other rail. If the light lights, that means the phasing is incorrect and the wires will need to be swapped.

Test Light

One of the most useful diagnostic tools is a simple test light. The test light I use most often has an alligator clip on one end and a light bulb and long sharp metal probe on the other. It looks similar to the ones here: http://www.harborfreight.com/3-piece-circuit-tester-set-94130.html (It likely came out of that set.)

The test light lights when a circuit has been completed. That usually occurs when one side of the light is connected to Rail A and the other to Rail B. The test light requires voltage to be present for operation, but with DCC this is no problem. With DC, power to the block must be on and the throttle set high enough to turn on the light. If either side is not connected to the rail or connected to rails of the same polarity, the light will not turn on.

Locomotives run under the same conditions that make a test light light. Just as a locomotive will move with the rails at any polarity, the test light will light with the rails at any polarity. Conversely, the test light will not light under conditions where the locomotive will not run, so it's safe to assume that if the light is lit a locomotive will run. (A DCC locomotive at speed step 0 is still technically running. The decoder is receiving power and interpreting commands.)

The test light works according to this truth table: (This doesn't represent all the options, just the important ones.)

Probe A | Probe B       | Result

--------------------------------

Rail A  | Not connected | Dark

Rail A  | Rail B        | Lit

Rail A  | Rail A        | Dark

Note where the probe lights when connected to different rails, but does not light when connected to the same rail.

Armed with this information, consider the following scenario: Trains can run on a section of straight track until they reach a specific point, then stop dead. How do you diagnose where this happens with a test light, and determine what rail the issue is in?

Insulated Rail Joiner

On a piece of track rarely used, there's an insulated rail joiner. It does not look like its needed any longer, perhaps it was back in the DC days. No reason to tear up existing track to get to an insulated joiner that's not bothering anything, in fact, I often advocate installing more insulated joiners than you need and wiring around them.

But here's the thing: the rail joiner has melted. It looks like someone tried to attach a feeder to the rails at that point and solder to a plastic rail joiner!

It's really not that bad

I subscribe to the theory that every rail should be soldered to something--either another rail or a feeder. Others say a feeder every 10' is good enough, while others report having gotten away with feeding an entire layout with one pair of feeders.

The game changes as soon as you paint the rails or glue ballast. The paint or glue works its way in between the rail and joiner, and turns an electrically conductive connection into an insulating one. This might not be a 100% insulating connection, but it often is enough to stop trains.

Now there's an insulating connection and the ballast has been laid or the rails painted. Fixing the connection requires either scraping some paint off the rails or drilling through ballast. The easiest repair is to drop a feeder at that point, so now there's a hole for the feeder and a bright shiny spot on the rails. Might as well do it before you start.

Compared to repairs and time that must be spent doing diagnosis, dropping a feeder to every rail is really not that bad.

Twisting bus wires

Some areas of the layout have multiple wires running through them that look exactly the same as other wires but are for different purposes. Multiple main buses coming from a circuit panel would be such an example. If the wires run through holes with one wire per hole, keeping track of which wire belongs to what can be difficult.

Twisting the bus wires together provides not only an electrical benefit (as discussed on the Wiring for DCC website), but a physical benefit as well. Twisting bus wires together forms a cable, so any time you need to drop feeders to that bus you know that you've got the correct pair of wires.

This can make up for some deficiencies in a color code. An easy color code to use is color/black for track power, but in a complicated area multiple buses may have feeders dropped to them. If the wires are separate, it soon becomes easy to connect the track to the wrong black wire. Twisted, it's easier to keep track of.

There are some downsides to twisting, though. Twisted wires can make it harder to solder feeders to the bus, and some block detectors will detect false occupancy if twisted bus wires are used downstream from the detector.

Designing a panel

In a previous post, I mentioned a new panel will have more on it than I originally intended. Recently, I had the opportunity to design another panel and took a different route.

Panels are usually made up of components such as circuit breakers, autoreversers, distribution blocks, and internal wiring. The whole point of a panel is to route power from the distribution blocks via internal wiring to the circuit breakers and autoreversers. A good one will be easy to follow, while a bad one is just some stuff with wiring.

Input and outputs usually work best if they occur near (but not at) the edges of the panel. In some cases, it's necessary to run an input or output through the panel, but generally that's where the distribution wiring should be. If the panel will be installed in the middle of a section, be sure to leave space for multiple wires to be leaving at that point.

Lay out the distribution blocks first. This will give you a good idea where the power will come from to go to the components. With something like a PSX circuit breaker, the distribution block is the wires that connect via the Power Link terminals. Otherwise, a distribution block could be a terminal strip with wires that join multiple terminal lugs.

The circuit breakers and autoreverser can be laid out next. Consider where the input and outputs will be coming from. If the input is coming from the right, put the input connection on the right. Likewise, consider where the outputs will be. Laying out components then becomes the process of matching inputs and outputs.

Be sure to leave space around the components around for the wires. Wires need a certain amount of space to bend and flex to make a connection, so it's a good idea to leave a couple inches open around connectors. It may also be a good idea to cut the bus wires a little long in case the ends need to be cut and restripped.

With the components laid out and secured, it's time to handle the distribution wiring. Route each wire so that it is easy to follow. In the case where multiple wires travel together, try to avoid crossing wires. Take the time and use the extra wire needed to make this neat. Wires can be secured using staples.

If necessary, larger cable staples can be used to provide a path for bus wires to travel through the board. These need not be driven tight, they're just guides to keep the bus wiring separate from distribution wiring.

With the panel now designed and built, it's time to install. Good luck!

Quick Tip: Big jumps

Some DCC CVs have a valid range of 0-255. To determine the proper setting, increase the value by 20 or 32 instead of 1 or 2. When you get close, then change the value by smaller steps.

Hexadecimal Programming

While programs such as JMRI have abstracted many programming details away, there
are several places where it may be useful to look at the CVs directly and in
hex. Trying to use "normal numbers" or base-10 to program CVs may make things
harder than they have to be.

Some CVs such as CV29 can be thought of as a series of 8 on/off switches.
Programming these CVs in hex can be much easier than using decimal. If you want
to turn off bit 3, using decimal you have to subtract 8 or using hex all you
have to do is set the bit to zero and look up the new value.

8-bit hexadecimal can be thought of as two groups for 4 bits. The low nibble
contains the first 4 bits (0-3) and the high nibble contains the next 4 bits
(4-7). The two groups are side-by-side to complete the byte. 0100 0110 is
0x46.

Memorize this table:

Nibble Hex
0000 0
0001 1
0010 2
0011 3
0100 4
0101 5
0110 6
0111 7
1000 8
1001 9
1010 A
1011 B
1100 C
1101 D
1110 E
1111 F

The low bit (0 & 4) is the rightmost bit.

To change a bit, leave the others alone and change the appropriate value. Look
up the new result in the table.

To use Hexadecimal mode with the DT40x-series throttle during programming, set
the left-hand value to something other than an address and press the right knob
once. (It does not appear address programming can be done in hex.)