DCS strength diminishing; do TIU's produce weaker signals as the units age?

Bear, if a layout has worked fine and then performance starts to deteriorate, I doubt that it would be the wiring scheme.

I had a similar problem many years ago.  I had 2 original-issue Rev G's.  Signal started to deteriorate.  Thinking the TIU was aging, I bought a then newly-issued Rev H.  No difference.  After much searching, I found that a feed into my ground common buss had come loose.  Resoldered it, and problem has never returned.

So I recommend carefully going over all track wiring tp check that no connections have failed.

It could also be that the passage of trains over the layout has loosen wire connections or track connections, or that corrosion has formed on track pins/fishplates/trackjoiners.  I would pursue your "obvious solution."

You mention having frequent power drops to the track.  The experts say there should be only one drop per block.  Is the center rail "broken" between drops?

If you have run trains in a high current condition the internal inductors that carry this current can overheat and melt the insulation causing a change in inductance.  If that occurs you can loose DCS signal strength.  In severe cases (I have seen 4 TIU) run in over current conditions, the inductor melts sufficiently to kill the DCS signal in that channel.

But I agree with RJR to look at track connections first, and any additions of devices and trains or lighted passenger cars that may be killing DCS signal.  G

I hadn't heard that before, GGG, but it would seem to be a good argument for replacing the 20-amp internal TIU fuses with 10-amp or, better still since the TIU case is plastic, leaving them 20 amp and using external 10-amp breakers or fuses.

Hello RJR,

Thank you for your detailed response.  Your answers prompted me to think about how I wired my layout.  For example, if I understood your explanations correctly, I should not have more than one power feed per block.  As you can see from my description above, I have more than one power feed to each of my blocks.

In addition, your comments inspired me to re-read my copy of the DCS O Gauge Companion.  In particular, I reviewed pages 54-66 (Planning for a New DCS Layout) and 151-154 (low signal strength problem solving).

Unfortunately, as pointed out above, I discovered that I need to reduce my positive feeds to only one per block.  Until I re-read the Companion, I had no idea that using more than one positive feed per block may confuse the electronics w/in DCS locomotives.  Furthermore, I should change from old style bus wires to home run or star style wiring. 

Also, I apparently violated the no-more-than 11 or 12 track sections per block.  The instructions indicate that, by soldering connections between tracks, I can reduce the number of sections. I use Atlas stainless steel track for my loops and nickle silver for my yards.  I'm unsure whether I'll be able to solder jumper wires to Atlas rails.  Fortunately, I have some extra Atlas curves with which I can experiment.  If I were using Gargraves, Ross, or tinplate track, soldering would be easy.

As I was reading the layout planning section, I learned that I shouldn't have stacked three of my five AIU's one atop of another.  Uh oh!  My AIU's have no heat escape vents.  As a result, I assumed that AIU's don't produce heat.  I don't have room to unstack them.  Additionally, feeder wires are cut to lengths appropriate for the distances between the units..

Over the years, I've had no AIU failures; nevertheless, I'm going to do a touch test to see if the cases get warm as I operate my accessories and switches. Maybe I can figure something out to unstack the AIU's.  I wish that I had built some sort of rack to hold the AIU's.

FWIW, I wired my layout for DCS when the system first hit the market. At that point, there wasn't much guidance.  I was unaware that traditional bus wire methods were inappropriate.  I used multiple connections per block to make sure that there were no power drops.

In any event, I am going to study my current set up to see how I can move wires to accommodate a home run type plan.  In the meantime, I can still run my DCS locomotives, but with some occasional glitches, e.g., lost engines, difficulties establishing and breaking down lash ups, etc.

Thank you to GGG as well.  When I run passenger trains, I have observed no signal degradation; instead, I have observed some minor signal improvement.

No problem soldering to Atlas solid rails, we do it all the time.  For jumpering track sections, I recommend a small chunk of wire, don't just try to jump the gap between the two track segments.

Bear, since you say your layout used to work well, I would not re-wire.  Something must have changed.  When you say passenger trains get a better signal, that raises a flag for me because passenger cars have lights.  Have you tried applying light bulbs to the track?

I note you spoke of stacking the AIUs.  Do not stack TIUs or place anything on top of them.  That will degraqde remote-TIU comms.

AIUs are simple relay stacks, so I would not expect overheating to be a problem.  You could use short dowels to move them apart (I think AIUs have holes like TIUs do.)

FYI, my layout was built long before DCS.  I used 14-gauge wire, with 70 blocks.  The blocks were rather short to permit operation of multiple trains on each "loop."  Transformers fed to a control panel with 6 sections, one for each transformer handle.  In each section the wire went from toggle switch to toggle switch, a switch for each block.  Ground is via a 12-gauge common loop buss that goes all around the layout.  When DCS came out, I plopped in the TIU between transformer and control panel, and, with some light bulb tweaking, have been running ever since.  Have had no problems although recently I started getting double toots, which had not been a problem; this indicates somewhere something has changed.

Reducing to one center rail feed per block is easy.  Simply use a moto-tool with the fiberglass cutting disc to sever the rail.  On my layout, I did the equivalent by severing the metal strip under Gargraves switches that connects the 3 center rails.  Every one of my 45 switches (save one) marks the end of 3 blocks.

RJR, yes, I am aware that AIU's relays; however, I didn't know that stacking them can interfere with the TIU's signal output.  I'm pressed for vertical and horizontal space with the AIU's.  I'm going to study the situation.  Usually, let problems percolate through my brain.  If I don't pressure myself.  Solutions seem to present themselves.

I have questions about blocks.  I am going to ask the question through an example:

Let's say that I want to power on loop of approximately 35 track sections using only throttle A on my Z400.

For DCS signal strength purposes, It's OK to power the entire loop only with throttle A so long as I create four or five blocks within that loop.  Is that correct?  If so, I would only have to insert a few more insulator pins into my problem and move positive wires to the middle of each bock.  Is that correct?  Of course, with only

In a couple of weeks, I'll be replacing several Atlas switches with Gargraves switches at key locations.  During the replacement process, I'll have to dismantle some track.  That will be a convenient time to establish blocks w/the correct length and feeder wire locations.

While at York, I had a long, helpful, and very nice discussion with one of the Gargraves representatives.  He told me that, for various reasons, Gargraves no longer installs the wires that you needed to cut.

The connector wires under the Atlas switches are absurdly tiny, so I wouldn't be surprised if the switches were blocking some of the signal.  In fact, I have noticed that the signal is worst in the area where the switches are.

Even though I probably don't have to do a major rewiring, I am going examine how I have set up my feeds from the TIU to the tracks.  Fortunately, when I wire track and accessories, I use color coded wires and label the wires as well.  Additionally, I maintain an Excel spreadsheet that identifies the position of each accessory on the AIU's and the power source dedicated to each accessory, e.g., I have a list of the accessories that throttle A on my ZW controls.

This discussion is very useful.  Unfortunately, some discussions on this forum are more emotional.

Bear, It's the TIUs that should not be stacked, for they interfere with TIU-remote comms.  I am not award of any consequences of stacking AIUs.  Sorry if I wasn't clear.

When you say 35 track sections, I don't know how long the sections are.  My longest blocks, using Gargraves 36" tracks, is probably 10 feet.  The shortest are a few feet.  In your 35-section loop, 4 or 5 blocks should do it.

I try to put feeders in the middle of each block for electrical distribution reasons--current has to flow through less track.  GRJ is correct when he says use jumper wires rather than soldering rails together; soldered track joints tend to break as the track flexes from train weight.

Rodger Pitcairn, since your setup had been working, you most likely have loose connections somewhere.  I have to retighten my TIU Output connections occasionally.  One of my sidings had a bad signal area for a long time and finally found a loose screw terminal for that area.

You can always get a few pieces of thin plywood or 1/4" masonite to hang vertically below the layout and attach the AIU's back to back with the wood or masonite in between. It would also allow you to keep your wire lengths short.

And we still have two original TIU's in use on a layout.