lost address?

[jward]

i have had this problem with several of my dcc onboard locomotives....

my friend programs the addresses for me with his nce power cab. they run fine on his layout. sometimes they even run fine on mine using a digitrax zephyr for a day or two. then, they won't respond to either the address they were reprogrammed to, or the default address 3. what is going on here?

i was under the impression that an address, once programmed, would stay programmed unless you changed it. some of these locomotives have been sitting in their box for months, they worked fine last time i ran them.

[Jim Banner]

Some decoders are worse than others for this.  I have had very little trouble with Digitrax decoders, for example, but ended up replacing a bunch of older MRC decoders.

The problem often seems to be associated with shorts on the tracks.  One of the reasons for running a bus with lots of feeders is so that any short circuit will immediately shut down the command station.  If this does not happen, then electrical noise on the track seems to have the potential to rewrite decoder addresses.  A similar problem can occur with very long buses ringing and producing spikes out toward the end of the bus farthest from the command station.  I ran into this problem when I was running DCC on one cab of a multicab system.  I would loose complete control of certain decoders in certain place on the layout, all far from the command station.  Sometimes the layout does not have to be all that large to create ringing, for example, if it has a power bus that forms a complete circle

Intermittent loss of power can cause false signals as well.  An example is dirty track or poor pickup of power from the wheels.  This first shows up as flickering lights and as it gets worse, you may lose control of the locomotive.  During this process there is some probability of decoder reprogramming.  Related to dirty track is faulty power transfer from section to section because of loose or corroded rail joiners.

These possible problems do not apply to all layouts.  But they do suggest some solutions.  One is to use a power bus under your layout, with a pair of feeders to every section of rail that is not soldered to another section with feeders.  And if you have an around the room type layout, be sure there is a gap in the bus (and the rails) somewhere about half the way around the railway from the command station.  If you have any doubts at all about your wiring, try the quarter trick that Digitrax suggests, and try it on each and every section of track on the layout, soldered or not.

Another aid to avoiding the problem is to keep track, wheels and power pickups on locomotives clean and lubricated.  The lubrication helps suppress arcing which is another possible source of noise that the decoders can mistake for signals.

Some people swear that turning off automatic conversion to dc helps.  I am not convinced that it does, but I figure it cannot hurt to try.  One thing I am convinced helps is eliminating all those little intermittent shorts that can creep into a layout - the all metal turnout that short when a driver bridges between the stock rail and the point rails, the metal cab that touches a metal tender on certain tight corners, that sort of thing.  I have suspected for some time that an accidental short between switch machine wiring and track wiring could also cause problems if you use dual coil machines, but have never observed it to actually happen.

I am sure others will have other suggestions, but I will give just one more - if your decoders have decoder lock, then lock them.

Hoping that at least one of these suggestions help ...

Jim

[pdlethbridge]

Jim, Some sound decoders come with a capacitor, they call it a stay alive capacitor. Would adding a capacitor to the circuit in the loco help with some of the problems?

[Jim Banner]

Most decoders already include such a capacitor, although smaller than the ones in sound decoders.  But they have to be after the point where the DCC signal is picked off by the microprocessor that reads it or the signal would be filtered out along with the noise.

Jim

[jward]

i am not sure any of the scenarios you describe fit the problem i am having.

what i have are locomotives that run just fine one day, then are put away and not run for a while. when i do try to run them again, they do not respond to any of the addresses they previously held. but a factory default reset will get them running again on address 3.

is it possible for the decoders to lose their memory the way computer chips do when not powered for a long time?

[pdlethbridge]

Jef, I had a similiar problem with some engines on my brothers system. He uses the NCE power cab like I do, but his unit, the hand held, was acting up and going blank on its lcd screen indicating a short. It also was having a problem reading back cv's. What made it hard to figure out was it was only happening with a couple of locos. They ran fine on my layout, would run fine on his until he let them go to move around the layout them the screen would go blank on the hand held. The hand held was getting hot by the consist buttons on the controller. What we originally thought was a loco problem turned out to be a controller problem. Larry, from NCE, picked up the controller from Despatch Junction hobby shop in east Rochester on his way home. We'll keep you informed on the situation.

[Jim Banner]

As far as I know, the only computer memory that normally keeps its data with the computer turned off is cmos memory for the BIOS.  And it can lose the data if the battery goes dead, which usually takes years.  But the RAM in a microcontroller (which typically forms the heart of a decoder) is more like the flash memory you use in your camera.  It requires no battery and only loses its data when you actively erase it.  Or at least that is what is supposed to happen.

There is some indication that flash memory can lose data when hit by static.  Whether or not static can make its way through several layers of electronics between the wheels and a decoder's RAM, I do not know, but I have seen static do some surprising things.  Like our group's layout that operates at the push of a button in our local museum.  The electronics worked perfectly in the spring, summer and fall of the first year and then went insane that winter.  Now keep in mind that our museum, like many others, is both temperature and humidity controlled, which would normally eliminate seasonal static problems.  Or so you would think.  Eventually we tracked it down to people wearing down filled, nylon shelled parkas.  They would charge up, then when they hit the start button, the electronics would develop a migraine.  Switching the start button to a metal shelled unit, grounding it directly to a real earth ground, and using shielded cable finally solved the problem.

I tell you that to ask you this - is there any possibility that you are charging up while putting the locomotives away?  Or even worse, charging up before you lift them off the shelf?  Walking over rugs is one of the best ways of charging up, followed closely by walking around in socks on just about any surface except untreated, unfinished concrete poured directly on moist earth.  You don't need to be charged up enough to be getting shocks off doorknobs - electronics can be sensitive to much lower levels of static than that.

Another thought struck me as I reread your initial post on the thread.  You mentioned storing your locomotives in their boxes.  Do the boxes by any chance include a foam lining?  Styrofoam is the worst, but even the artificial foam rubber used to pack a lot of things can generate static.  What about the clear plastic boxes (Lexan?) that some locomotives come in?  Could they be a source of static too?  Perhaps it would be worth while slipping your locomotives into anti-static bags before putting them away.  These are used for shipping static sensitive circuit boards and are usually free for the asking at places that repair or upgrade computers.  Some of the bags are pinkish poly bags, but the very best ones have a silvery look to them.  Failing that, a wrap of aluminum foil is proof against any static less than a lightning strike.  But if you use foil, use the thinest possible to avoid scratching your locomotives.

I am not at all sure that you are having a static problem, but I think it would be worth a try to take anti-static precautions until you can come to a definitive answer.  With static, that usually means that when you take the precautions, the problem does not exist.  If you were able to show that static is a problem, you would be doing a favour for every model railroader who uses DCC.

Jim

[mf5117]

I had this problem , And had to buy a new controller . I don't know how to explain it .but after a few hours of running . It was like what held the address in the controller or sends it to the loco . would not respond . Like the track was shorting out . The controller would appear to be fine but something in it was weak . I could reprogram it to adress 3 on an isolated part of track . then reprogram it , to the address I wanted then one loco would go dead like it was asleep had lights but nothing else . so I put it up . Thinking Im gonna throw it away . then another loco down . So I would get tired of messing with them and not try to run for a few days or a week . then would seem fine then same S different day . If you had another controller to barrow and see . that might help . I just had the digatil commander from bachmann . And bought another one . And haven't had any problems after that . but all bachmann dcc on board loco's . no custom installs ..... I'm still waiting on my mdt plymoth . there putting a N scale decoder in it with no mod's .The guy put a sound decoder in a thomas engine . believe it or not  I hope to here from them this week  . then i'll let you all know

[mf5117]

please excuse me ... I also had a short in one of my turnouts . When a loco would pass over it , it would not short like you would think . They would just stop on the otherside of the layout. But my track was not anchored down either . and with luck one day I found it with a meter , when I was altering my layout , the metal frog was shorting out on the power wire underneth . That you plug or unplug on the EZ track #5 turnout .... peace to all 

[mf5117]

dis regard what I said . It appears it doesn't matter .I"M removing my account thanks jim thank jw    regards mark f

[jward]

static electricity, eh? intriguing theory. i may have to buy some anti static bootstraps, and make sure that they are worn when running trains. i don't have any earth grounds in the living room with the layout but it should be a simple matter to touch the kitchen sink before running the railroad.

the odd thing is that this only happens with bachmann dcc onboard locomotives, programmed by an nce command station. it doesn't seem to affect any other combination.

when troubleshooting, i was always taught to look for what problem items have in common. that usually points me in the right direction.

[pdlethbridge]

Jim and all, I too was thinking this could be a static problem. I've built computers and they always warn you to wear an anti static strap to avoid problems. But in the models we deal with, as Jim stated, there are many things to happen before a decoder could loose its memory. We can't see or feel all the static discharges we produce, but even a small one could cause a problem.

[Jim Banner]

If the NCE command/Bachmann decoder combination is the key, then the problem should not occur when you change addresses with your Zephyr.  If the decoders are Bachmann non-sound decoders, remember to connect a 1000 ohm resistor across the rails (or across the programming outputs of your Zephyr) while changing addresses.  This is one more thing you can try in your quest for an answer.  I wish one of us had been able to give you a straight forward answer, but your problem has turned out to be more difficult than most, and it may be a while yet before it is completely solved.

Jim

[jward]

jim.
if it were a simple problem, it would have been in the manual an/or i would have been able to figure it out by myself. anybody who knows me will tell you that often ask off the wall questions. i'd like to think that answering them furthers not only my knowledge, many others as well....

i will have to set up a programming track, but as you may be aware my layout is currently under construction. laying your own track is a slow process.

what i have been able to do with my zephyr is to reset the decoder to factory default. i can do that even without the 1000 ohm resistor. any further programming of the bachmann decoders requires the resistor. weird, huh?

thanks alot for giving me some leads to follow up on.