BLI Locos Dropping to Idle in a Particular Block

[eric220]

I had an interesting problem crop up the other day. After years of faithful service, and several laps that day, a reversing section powered by a DCC Specialities OG-AR caused one of my BLI M1's to come to a stop. At first, I thought something had happened to the decoder, since it was still making noise, but it wasn't responding to the throttle commands. As soon as I got it onto another block, the M1 started responding again. Consistently, if I drive the locomotive into the AR block, after a second or two, it reacts as though someone set the locomotive to idle. It coasts to a stop and just sits there, and won't respond to throttle commands. Pushing it back onto a regular block restores it to normal operation.

My first troubleshooting step was to replace the OG-AR with a new one. No change. Next I tired the M1 in another block powered by an OG-AR. It worked fine. Next I tried another sound-equipped locomotive, a Bachmann K4 in the trouble block. It performed normally. Next I tried a different M1. It ground to a halt and behaved just like the first one. Next I tried a BLI centipede. At first, it performed and responded normally, but after it passed all the way through the block and tripped the AR, it began behaving strangely. It was clearly not receiving all DCC commands sent to it; the horn would come on, but would not shut off. One of the two engines continued running, but the other ground to a halt. After repeated commands to change speed, it finally limped back to the other end of the block. As soon as it was out of the AR block, it behaved normally.

I'm kind of at a loss to explain what's going on. The only other data point I have is that the AR is consistently failing to trip before the circuit breaker at one end. Not sure if that's relevant or not.

Any suggestions?

[C855B]

If it were my RR, I'd drag out my (t)rusty old HP oscilloscope and start looking at waveforms. I will bet that something is happening to the square wave transitions in the command sequence, likely in one direction only, which is causing a distortion that the BLI decoders don't like but the B'mann can still read through.

Failing having an o'scope sitting around... , I would go through the entire wiring harness for that block, looking for loose or oxidized connections (including cold solder joints to the track) that might be compromising the signal. That the AR is tripping "late" is a hint that there might be an unexpected resistance, again from a marginal connection somewhere.

[peteski]

Have you changed anything relating to the track, wiring or DCC between when it last worked and when you discovered the problem?

Could you expand a bit on "AR is consistently failing to trip before the circuit breaker at one end."?  What exactly do you mean? How are they wired together?

[eric220]

@peteski I didn't change anything. Quite literally, the loco was running laps and suddenly crapped out. What I meant by the AR is consistantly failing to trip at one end is that when the locomotive crosses the double gapped rails at one end of the auto-reversed block, the circuit breaker upstream of the auto reverser trips and takes down the district before the auto reverser can flip the polarity. Given that the block is connected to the bus with suitcase connectors, I suspect a "high resistance" connection, as @C855B suggested might be at fault. I think that's the next troubleshooting step to take. Tomorrow. After sleep.

[peteski]

With no train on the tracks can you trigger the reverser to trip bu taking a coin or some other metal object and shorting the gap (on one rail only)?  If the polarity is opposite on that end, even single rail polatiy mismatch should trigger the reverser. Try that on all 4 gaps and note the behavior.

The other thing I/m curious about is whether the locomotive can cross the reversing block gap and keep on going when it doesn't need to trip the reverse (the polarity of the reverser block is already the same as the rest of the track).

But like you guys said, the problem is most likely some deteriorated wire connection somewhere.

[seusscaboose]

Eric
I'm curious about the outcome of this

Please post what you find


My main question is was it wire related or something you unintentionally changed or both

[bdennis]

I have has some issues with the OG-Ar not working correctly when it is up stream from a BDL168 block detector.
I found that due to the voltage drop on the BDL 1.2v and with the DCS100 set to N scale there was not enough volts to enable the OG-AR to work.
I changed the switch on the DCS100 to HO rather than N and the issue went away.

Also what main system do you have? Is it a Zephyr? - A note from the DCC Specialties web site - Lower power systems such as the Digitrax Zephyr and the Atlas Commander lack the power to operate OnGuard. (This lead me to change my DCS100 to HO to up the track voltage from 11v (Nscale) to 13v (HO).

As the BLI M1 may be drawing more current due to the sound decoder this may also bring the OG-AR closer to it lower limits?

[eric220]

@bdennis I've got the 8 amp Super Chief set, and I'm using RR CirKits detectors.

@peteski I did try running the M1 into the block after the reverser had thrown the polarity to match. I also tried dropping the M1 onto the tracks in the middle of the block. No improvement. The symptoms would seem to indicate a power and/or signal dropout somewhere specific to this block.

[seusscaboose]

For me... The curious thing is "it just stopped working".  That doesn't just happen (usually).  Something changed. 

At least... That's where I would head.  What changed in that short time period between when it was and was not working.

It is possible the board just gave up

Happened to me once

But on a 4 year old board


Have you tried turning the system off and resting it for 10 minutes, then bringing everything back up?

Just brainstorming

It could be anything

[eric220]

@seusscaboose I've swapped out the board, and I've tried it on two different days.

I agree that something must have changed, since everything ran successfully for years up to 10 minutes before the problem first showed up. I still think the suitcase connectors are my prime suspects at this point.

[peteski]

How about the other tests I've mentioned (shorting one rail gap at a time and observing the results)?

If the locomotive is placed in the middle of the track powered by the reverser and it is not working then there is some serious signal degradation.  Can you remove the AR board from the equation and simply install couple of alligator clip jumpers on the top of the layout to feed that track section from the track on either side (try both sides) and see if the loco behaves correctly on that track section.

[John]

Also make sure these hints are being followed

Helpful Hints

When Isolated and Reverse Sections exceed 10 Ft. in length install at least 2 sets of track feeders for that section. Lack of feeders causes a voltage loss.

For reverse sections, we recommended that the gaps be staggered about 1/8". Perfectly aligned gaps may reduce performance.

If your Power Sections and Reverse blocks are greater than 10 Ft. long be sure to have at least 2 sets of track feeders for that section. Insufficient feeders will cause a voltage drop.

One end of the reverse section will be aligned with normal polarity track power while the opposite or other end will have a polarity mismatch and require the reverser to act.

Note: Not for direct current (DC) applications.

[John]

you might also try removing the block detector from the circuit .. both ahead and in the block .. if the detector is failing, it may not provide the power .. do that after you try the peteski suggestion of just jumpering in

[GaryHinshaw]

Eric, if I understood you correctly, you have an OG-CB upstream from the OG-AR.  When I asked DCC Specialties about hooking these up last year, they told me do not run these in series.  You should try wiring the AR board (which has its own circuit breaker) directly off the bus from the command station.

[eric220]

How about the other tests I've mentioned (shorting one rail gap at a time and observing the results)?

If the locomotive is placed in the middle of the track powered by the reverser and it is not working then there is some serious signal degradation.  Can you remove the AR board from the equation and simply install couple of alligator clip jumpers on the top of the layout to feed that track section from the track on either side (try both sides) and see if the loco behaves correctly on that track section.

I hadn't done them because I hadn't been back to the basement yet!   

When Isolated and Reverse Sections exceed 10 Ft. in length install at least 2 sets of track feeders for that section. Lack of feeders causes a voltage loss.

There are three sets of feeders.

For reverse sections, we recommended that the gaps be staggered about 1/8". Perfectly aligned gaps may reduce performance.

Interesting, but it ran fine for years with non-staggered gaps.

Eric, if I understood you correctly, you have an OG-CB upstream from the OG-AR.  When I asked DCC Specialties about hooking these up last year, they told me do not run these in series.  You should try wiring the AR board (which has its own circuit breaker) directly off the bus from the command station.

No, there is no OG-CB upstream.  There is a PM24, but again, it's worked that way for a long time.

Edit: to clarify, the PM24 upstream of the OG-AR is protecting the entire mainline as a power district, not just the reversing section.

New data points:

• I crossed each of the gapped tracks as Peteski suggested.  No perceivable reaction, including a short, so I presume the OG-AR is functioning as expected.
• I flat-shorted the AR section.  The layout did not go down, so I presume the built in short protection did its job.
• I busted out my multimeter.  There is no voltage difference when comparing the AR section to the next block.  (Interestingly, there is a significant voltage drop at the other end of the layout, around 100 wire feet away, but there are no operating problems there.
• I brought out the convincers and clamped down the suitcase connectors, which appeared to be nice and tight.
• I removed the OG-AR and jumped the wires.  The problem persists, so I guess the OG-AR really doesn't have anything to do with it.  (Guess I need to change the title of the thread.)
• While I was testing, my DT400 crashed.  It stopped responding, and when I unplugged/replugged it went to "idle."
• Remembering that my DT400 had been plugged in and selected to the problem loco address for all the tests, I unplugged it and repeated the test with my UT4.  The problem persists.
• I remembered something that happened recently.  Last week, right before all the progress on the Pi, I accidentally left the layout on for three days.
• Just yesterday, after debating whether to get a second booster, or get a new command station and relegate the old command station to booster duty, I went with just getting a second booster.  I suddenly feel like murphy is watching.