Problem with AR-1 (Auto Reverse)

[MtRR75]

I have a simple layout with a reversing section.  I have an inner and an outer loop, which are connected to each other.  Then there is a diagonal track across the inner oval, which allows a train on the inner loop to cross over and reverse direction.  (I have isolated this diagonal track as the reversing section, and connected it to the AR-1.  All the rest of the track (both loops) is connected directly to the power supply (PWR 56).

I have tested the AR-1 with two of my locos.  Loco 304 is a Bachmann 2-8-0 with an LM-2 installed in it.  Loco 1396 is as an old Mantua 4-6-2 with an LM-3S and a Keep-Alive.  Both are steamers.

Test #1: I ran #1396 continuously around the outer loop, while shuttling #304 back and forth on the diagonal reversing track, from one side of the inner loop to the other side.  Both locos ran smoothly, and were not affected when the AR-1 flipped the polarity of the diagonal track (which I could observe as the AR-1 indicator light switching between green and amber).

Test #2:  I reversed the locomotives.  I ran #304 continuously around the outer loop, while shuttling #1396 back and forth on the diagonal reversing track, from one side of the inner loop to the other side.  This time, #1396 ran smoothly across the reversing diagonal, and was not affected by the polarity changes.  However, every time the polarity switched on the reversing diagonal, #304 (on the outer loop, where the polarity never changes), stalled more a moment, then quickly restarted.

Any ideas what is going in here?  Obviously, the Keep-Alive is a factor.  But can LM-2 locos run smoothly on track that is not part of the reversing section – when another loco causes the reversing section to switch its polarity?

Details:  Running an HC-2 with HC-2 software version 2.06.  Both locos are running LM software version 2.09.

[G8B4Life]

Could you do a mud map of the layout and how the PWR-56 and AR-1 are hooked up (something simple in MS Paint will do)?

Does the PWR-56 give any indication when either of the loco's cross the AR-1 boundaries? what you describe sounds like the PWR-56 is detecting the momentary short that occurs when an AR-1 switches polarity. I shouldn't think this should happen unless something isn't hooked up correctly.

The software versions of the HC and LM's should be irrelevant in this case thought there is 2.09 available for the HC-2.

- Tim

[MtRR75]

Sorry for the delay in responding.  Lots of family stuff came up.

The PWR-56 indicator light stayed on green through all of the tests and never flickered.  The indicator light on the AR-1 flipped between green and amber, as the polarity was switched.  It also never flickered.

Attached is a diagram of my 4' x 8' layout and a hand-drawn wiring diagram for the PWR-56 and AR-1.





Any other ideas as to what is going on here?

Does anybody else have this problem?  Surely, others have run multiple trains while one train is crossing a reversing section.

[Alan]

All of the yard tracks (except the bottom one) and the two spurs to the left of the purple track should be wired to the AR.

Train total length cannot be longer than the reversing section if there are metal wheels on your cars. Might be a challenge on such a short reversing section.

[MtRR75]

Thanks for the response.

(1) All of the yard tracks and spurs that are connected to the purple track are wired to the AR-1.

(2) I am aware of the problems that metal wheels on long trains can cause.  However the tests that I outlined in the original post were done with just the locos (and their tenders).

[G8B4Life]

Hmmm, perplexing to say the least, I can't see anything obvious from your diagrams that could be causing an issue.

Since it looks like we can count out shorts going by the LED's on the PWR-56 and the AR-1 then the only thing I can thing of, and as far fetched as this sounds, is that the outer loop is somehow loosing power momentarily when the AR-1 switches polarity.

Did you try this as an experiment? Have 304's information screen open on the HC when 1396 goes across the AR-1 boundary and the AR-1 switches polarity and see if the information screen reports anything. It's a real long shot but it might provide a clue if you haven't done it. You could also try hooking a multimeter up to the outer loop (measuring both voltage and current draw) and watching that as well when the AR-1 switches.

With the exception of the AR-1 section I'm guessing that all the "sections" in the outer and inner loops are just "rail level" gaps (no electrical isolation between sections) and not also "bus level" gaps (you can isolate each section electrically).

Is this a new phenomenon, ie the AR-1 has been in place for a length of time but it's only started happening? or this has happened right from the outset of the installation of the AR-1?  Just wondering if the AR-1 could have developed a fault if it's been in place for a length of time.

Unfortunately I don't have an AR-1 myself to to try and duplicate what your seeing.

- Tim

[MtRR75]

Thanks for the response.

I will try the info-screen and multimeter experiments (maybe today or tomorrow) and get back to you.

Your other questions are best answered by giving you a short history of this layout.

I inherited this layout as an old classical, 4’ x 8’, DC, 2-cab layout in very rough shape.  Over the years, I have completely rebuilt and rewired the layout.  Then I discovered RailPro.  I am slowly converting my locos to RailPro.  But I still have more DC locos than RailPro locos.  So I installed a master switch that converts the layout between DC and RailPro.  It is completely idiot proof.  There is no way that both power systems can be on at the same time.  (But I still have to be careful which locos are on the layout before I flip from DC to RailPro, or vice-versa).  To run RailPro, I flip all of the block switches to Cab-A, then switch the power from DC to RailPro.

So, to answer your question… All of the block insulators really are insulated joiners.  But all of the reversing track is wired together in the control panel, and all of the non-reversing track is also wired together (in a separate bundle in the control panel).  However, if I accidently leave one of the block switches set to Cab-B or OFF, the train just stops when it enters that block.

As for the reversing section, when I was researching RailPro, I exchanged numerous emails with Tim Ring before I purchased the system.  One of my questions to him was, could I use the DC directional switches to operate the reversing section in RailPro mode.  He said that I could, but it would be cumbersome.  I have operated in this manner for a couple of years.  But recently I had to rebuild my control panel to accommodate several new yard tracks and turnouts.  So I took the opportunity to add an AR-1 to my system.  The addition of the AR-1 was one of the last things that I did.

So, to answer your question… The AR-1 was just installed last week, and these were the first tests of the AR-1

[Alan]

I believe an important clue is in the operational differences between the different locos. The system must be sound if one loco works correctly.

Possible the wheel treads of 1396 are ever so briefly shorting the insulating tip of a frog (heel point)? The short detection in RP power supplies is really fast, both acting and resetting. If the AR were to flip at the very same moment a frog tip is shunted then the PWR does what it is supposed to do - shut down and reset. Peco switches are especially known for this phenomena. Nail polish is a common fix.

Slight differences in wheel gauge or wheel tread width would then explain why 1396 does and 304 does not.

Watch super carefully as 1396 crosses a frog. Does 304 stutter just as a 1396 wheel is on the frog tip?

[G8B4Life]

I like your theory Alan, sounds better than mine! Could a short be so long as to stall the loco on the track but fast enough to not drive the short indication on the PWR-56 or AR-1? I guess it depends on how long that stall moment is when 304 stalls.

I looked up 1396 on the 'net, a 4-6-2 with insulated pony, pickup on one side of the drivers, insulated trailing and pick on the opposite side of the tender. I wonder if this issue happens at both ends of the reversing track. I neglected to ask...

Running 1396 tender first from the reversing section into the inner loop using the insulated side of the tender to go across the frog and then the non insulated side to go across the frog should partially validate whether it's the switching of the AR-1 or not that causes the issue. Of course if MtRR75 comes back and says that the frogs are one big hunk of plastic like old Peco Insulfrog and old Atlas then it's a moot point.

MtRR75, you could also watch the info screen of 1396 as it passes out of the reversing section and see if that give any warnings or such.

- Tim

[MtRR75]

I figured it out.  Thanks for all the help.

Details later today, when I have more time to post.

[MtRR75]

I think I have figured out what is going on.

SHORT VERSION:  The problem was inadequate power to a section of rail.

LONG VERSION:  As Alan suggested, the problem is NOT with loco #304 – which is the one that is stopping when #1396 crosses one of the ends of the reversing section.  The problem turned out to be #1396 – the one that was not stopping.  Loco #1396 has an LM3s with a Keep Alive attached, which keeps it going..

As suggested, I started watching the voltage on the HC-2.  In order to do that effectively, I slowed the speed from about 25 to about 15.  At first I was getting erratic results.  Most of the time, #1396 was crossing in and out of the reversing track without affecting #304.  But occasionally, #304 would suddenly stop (and restart) as #1396 crossed the end of the reversing track.

Suddenly, I noticed that #1396 was SOMETIMES slowing down a little as it crossed the end of the reversing track.  I was able to reproduce this and watch the voltage.  The voltage was dropping gradually as #1396 slowed down, indicating a loss of power, which would make #304 stop.  So I appear to have a section of track that does not always deliver full power.

The first thing I did was to test all 8 possible situations where the loco could cross the end of the reversing section. (2 ends of reversing track X 2 directions of travel {in or out of reversing track} X 2 loco orientations {pilot first vs. tender first} = 8 combinations).  The problem only occurred in one of these 8 situations (loco exiting reversing track, tender first, at the SouthEast end of the reversing track – near Turnout #9 – see diagram in earlier post.)

Because the Keep Alive kept the loco going until it regained full power, I was able to figure out exactly where the loco was when the voltage began to drop and where it was when it when it regained full power.  This section turned out to be the point rail in Turnout #8 and extending SouthEast to the insulated junction at the end of the reversing section.  Ohm measurements confirmed that the resistance between this section and the rest of the reversing track fluctuated wildly as the point rail was wiggled.

As an aside, why did the train run smoothly across this section when travelling loco first?  When the loco is in this direction, the tender pickups are on the defective track rail.  The two trucks are farther apart than are the 3 drive wheels that pick up power on the other side.  The two tender trucks straddle nearly all the way across the defective area.  When the loco is reversed, there is a longer time that the three drive wheels spend in the defective section.  Also, the loco is much heavier than is the tender.  So the drive wheels push the loose point rail in Turnout #8 out of alignment more readily.

Installing jumper power wires to the points and an addition power drop to the short track section between Turnouts #8 and #9 should fix this problem.  I probably won’t get this done for a while, as I won’t get much train time in the next month.

Thanks for all of your help.

[Alan]

Glad to hear you are solving the problem.

Yet another testament to the old adage - a feeder to every piece of rail.

The hobby world seems to have an irrational fear of wire or a misconception they can get lucky enough that power dropouts won't happen to them. But in the end, the story is always the same. Forced to go back and add feeders. It is like the old Fram filter commercial - pay me now or pay me later. Your choice.

I chuckle every time someone posts on my blog saying I am insane for having 8 feeders on a turnout. The popular word used is "overkill". Guess who has never had to lay upside down under a layout adding feeders!