DCC Santa Fe FTA unit not burnedupdue to dc-now runs out of control

[rako]

 ???To Jim Banner and others that offered help whenI asked about dc burning up dcc units,I now have a runaway machine.Now,instead of being dead on the tracks,it's a runaway!!!Before I could try to reprogram on the EZ command module,it started to run at full peed and at any and all 10 positions.Only way to stop it is to hit the stop button.
This is getting weird and humerous.Does this mean that the decoder logic is kaput and sending a signal to run,but why does the speed override the command module's rheostat?
Rako

[ABC]

It may be getting both DC and DCC power.

[rako]

ABC-no dc on the tracks when I set the unit and off it went.
Rako

[ABC]

The decoder is likely the problem then, I suggest you seek the advice of the omniscient (infinitely wise) Jim Banner who assisted you previously. But, before you do anything be sure to seek his advice, he really knows his stuff, and would be much more helpful then I could ever be.

[Jim Banner]

I have been thinking about this problem since you posted it yesterday.  The group I model with has a large scale portable layout that we set up for shows.  One of the fellows has an 0-4-0 side tank porter that has a similar problem to your original problem.  From time to time, it loses its address.  This sometimes correlates with a derailment somewhere on the layout momentarily shorting the rails together.  Other times, we have no idea what causes the address change.  And being an older MRC decoder with no read back, we cannot determine what the new address is.  Once in a while, this particular locomotive will suddenly run at full speed and will not respond to the throttle (sound familiar?)  We put it down to the decoder reprogramming itself to a phantom address in our Digitrax command station.  A phantom address is one that was accidentally left in the command station with a non-zero speed, even though the locomotive that had been run on that address is no longer on the track.  At that time, and with that decoder, the only way we could have tested this theory would have been to try all ten thousand possible addresses, so needless to say we did not try.

Your E-Z Command does not suffer from phantom addresses in the sense that Digitrax does.  But it does have me wondering what would happen if a locomotive running on E-Z Command were to reprogram its address to one of the addresses not used by E-Z Command and then or later receive a signal sent to that address telling it to increase speed.  It would, I suspect, go to the speed that it was told to go to.

Did you ever get a chance to reprogram your grandson's locomotive from an unknown address to a known address?  If not, you could still try that.  What I am hoping is that as soon as it receives a signal telling it that it is being reprogrammed, it will stop running.  I know this works with Digitrax.  The first thing it does in program mode is send an "all stop" signal to stop all locomotives that might be on the track, main line or programming track, depending on which command station you are using.  If that did not work with E-Z Command, I would be asking my local hobby shop, model railroad club, or other local model railroader for help with the reprogramming.  This is the procedure we have always used to bring our one recalcitrant locomotive back to sanity.

Sorry I cannot give you a definitive answer on this.  I do not have any decoders that will lock up on demand to try this cure out on.  If you lived closer, I would be only too happy to invite you over so we could try out a few of these ideas.

One thing I can explain is why the locomotive can ignore the E-Z Command's speed control.  The old rheostat speed controls varied the voltage on the rails.  Turn the power down to zero, and the train had to stop.  Sort of like having your dog on a leash.  You nail your end of the leash to the ground, your dog isn't going anywhere.  But with DCC, there is always power on the track.  Whether or not the locomotive uses it is up to the decoder inside the locomotive.  When you send it a signal it understands, it normally obeys and does what you tell it to do.  Just like your dog.  If you let him off his leash, you can still control him with voice signals.  What I think has happened in your decoder, it has switched its address and will no longer respond to any of the addresses you can try.  This is sort of like you telling your dog "Sit, Rex."  "Come, Rex." or "Go Home Rex."  and he doesn't obey because his name is Fido.  Or for some reason he thinks his name is Fido.  For your locomotive, it would work this way:  Let's say you were normally running it on address 7 (button 7 on your E-Z Command.)  Then for some reason, your locomotive gets a signal (or thinks it gets a signal) telling it to change its address to 39.  Being a good little decoder, it changes its address to 39.  Now when you send a signal to 7, your locomotive ignores it because it is now looking for signals addressed to 39.  You can try all of the addresses from 1 to 10 and your locomotive will ignore all of them because it has been told that its address is 39 and so it ignores all other addresses except 39.  The speed control is one of the controls that helps define exactly what signals are sent out, but if the decoder ignores them, there is nothing it can do about it.

So how does this accidental reprogramming occur?  Our trains operate in electrically noisy environments.  Every time a wheel loses contract with the rails, and every time a short circuit occurs, there are many, many ups and downs of the voltage on the track.  If you could listen to this noise, it would sound a lot like wind in the trees.  Now I don't know if you have ever heard voices in the wind but it is a fairly common phenomenon.  Some parts of the noise seem to sound like voices.  In the case of the decoder, sooner or later some of the pulses seem to make sense and the decoder reacts accordingly.  If you don't like looking at the problem that way, think of the story about the monkeys all typing randomly on typewriters.  With enough monkeys and enough time, they will write all the plays of Shakespeare.  Make enough electrical noise, and sooner or later you will produce a combination that passes as a valid digital command.  The simpler the command, the more likely it will be to produce an interpretable command.  I don't think at this point you will be surprised to learn that changing address and setting speed are both some of the simplest commands.

The solution is simple in concept.  We can either start calling our dog "Fido" because he thinks that is his name or we can convince him to go back to responding to "Rex".  Both will work.  Only trouble is, with 10,000 possible names (addresses,) finding the one that works will take a long, long time.  But changing his name back to Rex (or 7) should be a much simpler process.  And once you do that, telling old Rex (or 7) to stop and go should find him back to his usual cooperative self.

Jim

[rako]

Jim B and ABC As soon as I get to my grandson's I will try the reprogram again.I got confused before and tried to stop the engine first rather than see if the unit stops with the reprogram step Will let you know.

[rako]

Jim Banner.The reprogram did stop the runaway, but no luck on assigning a new address.The unit just keeps ignoring any new addresses and only stops if a reprogam is attempted.I guess we have to assume that the decoder has an address that overrides other commands.
I'm an old electronic engineer(now retired at 66 yrs) and took this as a nice learning experience and challenge with help from experts such as yourself.I understand how it could pick up an extraneous code the way you described and will now punt and buy a replacement Santa Fe for him.I really appreciate the help to date and look forward to more dialog with you and other model train hobbists.
Thanks again Jim.
Rako