N Scale Runaway

[justjames]

Greetings,
 
I purchased a Kato U30C at the local trainshow yesterday. It had been converted to DCC by the previous owner. CV8 = 36 so that reads to be a DCC Concepts decoder. My command station is a Digitrax EVOXD with a DCS210+ and 602D throttle.
 
When I got home and tried it and it barely ran so I cleaned the wheels and contact strips and did a lube job. It ran nicely although the start was a little slow to my liking. I placed it on my programming track and changed the "03" address to a four digit matching the cab number and decided to adjust CV2 up from 0 to 20 to see if the start speed improved. Once I entered "Write" it immediately launched off my programming track at full speed.
 
Next I placed it on the mainline and it instantly takes off at full speed. Back to the program track and it will sit still until I attempt to "Read" or "Write" anything and off it goes. My internet sleuthing says its probably thinking DC mode but without the ability to get it to sit still for programming, what to do?
 
Suggestions and help needed, thanks in advance!
 
James

[peteski]

I'm not  familiar with DCC Concepts decoder brand (looks like it is Australian), and I don't  believe I ever heard a model actually running in  programming track when doing Service Mode programming. I don't use Digitrax command stations, but programming track should not have any DC voltage, and the only voltage there should be DCC programming packets which are only sent while the decoder is  being programmed. But then again I don't use Digitrax.


As for the DC mode, all you have to do is check if it is enabled in CV29 and then disable it and see of it makes a difference. But if there is a proper DCC power (signal) present on the track, even a decoder with DC enabled should not start moving.  The typical "DC type" runaways happen only when a command station is being powered up and sends some DC voltage to the track as it initializes (not a desired thing in a first place).

If you are not familiar how CV29 works, you can use https://www.2mm.org.uk/articles/cv29%20calculator.htm to decipher the value in CV29.

Another way to check for DC operation is to touch a fresh 9V battery to the wheels and see of the loco runs. But there could be certain types of decoder failures (besides the DC mode being enabled) that would make the model run in DC.

[ednadolski]

If you haven't already done so, do a full factory reset from the programming track.

Ed

[justjames]

Thanks for the ideas.

The motor runs when using a 9V battery. It also runs uncontrollably when on the mainline with no throttle input. Once on the programming track with a Digitrax system, the options are to "read" existing CV's on the decoder or enter any new values and "write" the changes to the system.

As suggested, I've removed the shell, turned the loco to a wheels up position and tried "reading and writing". The engine goes wide open instantly when either command is entered. I've tried both commands and allowed the motor to run for several seconds while hoping this would allow the programming to complete it's cycle. No luck.

[peteski]

Maybe I misunderstood the original problem description.  Or maybe some details  were missed.

At its current state, if you place the loco on the dedicated DCS210+ programming  track, it starts running immediately, or  pnly when you attempt to read or write the decoder's CVs?  Again, not knowing the nuances of how Digitrax powers its dedicated programming track I'm just making educated  guesses.

When you went to make the initial changes, did you try to read any CVs, or you went right to programming the address?  Did the command station maybe read some CVs (liek the address) first?

As Ed suggested, try resetting the decoder (on the dedicated programming track, according to what Digital Concepts recommends as it might no be CV8=8).   Also when you power it with a 9V battery do you have control over the direction?  For example does the loco run forward and in reverse by flipping the battery's polarity?

[Maletrain]

To keep things under control while you read and write on the programming track, how about removing the shell and putting jumpers with alligator clips from the rails to the frame halves so that it can be run without resting on its wheels?

Then check that CV29 is set to not allow DC ("analog") operation.  Also check to see if it allows 28/128 speed steps - if not, it is using 14 steps, which might result in a somewhat fast start.

A CV2 setting of 20 is pretty fast for starting a Kato loco.  Mine often crawl OK with a 0 setting, and usually end up in the 3-5 range when trying to speed match with locos with mechanisms that don't crawl as well as Katos.

I suggest that you try to find a list of the CVs used by DCC Concepts decoders.  I have found that many, especially older, decoders do not stick to the NMRA standards for which CVs are implemented and in some cases how they are implemented.  (For instance, I had a loco take off in reverse after adjusting CV95 from 130 to 120 -- turns out that CV adjusts forward trim only for values above 127 and values below 127 adjust reverse trim - - so I inadvertently set my reverse trim from 0 to 120!  That decoder does not even have a CV66 that NMRA says adjusts reverse trim.)

Not knowing how your DCC Concepts decoder uses CVs is like fumbling in the dark.

[C855B]

Question: Is it the only locomotive on the railroad? I ask because I had a situation with a runaway where another loco on the rails was doing something to the DCC encoding which rendered the one in question uncontrollable. Removed the second "bad" decoder, suddenly everything was copacetic.

[justjames]

peteski,

I'm sure my description left a lot of room for confusion. To answer your questions, when placed on the program track, it starts running only when I attempt to read or write to it. ( It starts immediately if on the main.)

I did initially read CV's 1-6. The address was 03 and CV's 2-6 were all 0. Next I changed the address to a 4 digit address using the "Quick Decoder" function which is somewhat automated. I did not attempt to read this as it normally is automated. I then changed CV2 to 20 in an attempt to give the start a little more pep. With Digitrax, that data is entered and then the "Program" function is called "Write". As soon as I hit "Write" the runaway began.

At this point I can't read or change any CV's. This includes resetting.

Your 9V question is interesting. I can only get the motor to run in one direction. I've tried changing the polarity and tried it on all four strips. What does that tell you?

Thanks for helping.

C855B,
 
I've got around a dozen loco's on the layout at any given time. I will try removing them one at a time to see if that helps.

Maletrain,

I have the shell off and jumpers available. At this point, I can't read or write and CV's.

[Maletrain]

This is beginning to seem like some sort of short from the frame directly to the motor.  I have read that some Digitrax command stations do provide enough current to the programming track to "test the changes", but have not verified that, myself.  So, maybe that is why the read and write commands set it off on the programming track and it just takes of on the main.  Also, a short would sabotage the communication from the decoder back to the command station that is based on a very brief current draw from a brief application of voltage to the motor from the decoder.

I don't see that writing a value into CV2 could logically cause this behavior.  But, perhaps the little jerk that the decoder does with the reads and writes was enough jostle to make a connection that was just a hair's breath away until that write.  Or, perhaps it was just pure coincidence.  I have gone down several rabbit holes because of coincidental events that were functionally unrelated to the action I had just taken.  So, it does happen.

Edit:  On second thought, a short from the frame to the motor would not cause motion on DCC track.  Perhaps a short from somewhere on the decoder board, after the rectifier to something connected to the motor?  Maybe something wrong with the actual electronics on the board, or maybe some lack of insulation between board components and the motor wire?  I don't know what to make of the observation that the motor only runs in one direction with a 9v battery being applied (I assume) to the wheels.

[ednadolski]

I purchased a Kato U30C at the local trainshow yesterday. It had been converted to DCC by the previous owner. CV8 = 36 so that reads to be a DCC Concepts decoder. ....

I'm not  familiar with DCC Concepts decoder brand (looks like it is Australian)...

Not to spread FUD/innuendo, but I think JoeF does his homework.... anyways, read the whole thread and decide for yourself.  My point being that if this is some odd-duck sort of decoder, it may turn out to have some inherent quirkiness (e.g. not NMRA conformant, etc...), or simply just be damaged/broken:

https://forum.mrhmag.com/post/dcc-concepts-accessory-decoder-availability-13811446


If the latter, then it may just be worthwhile to consider swapping with a know-good decoder, just to save the debugging time & hassle. I don't suppose you had the chance to test it at the trainshow when purchasing?

Ed

[justjames]

After reading the above referenced thread and considerable rabbit hole digging I think it's time to punt.
Thanks to all for suggestions and tips. If I luck into an unexpected fix, I will post it.

James

[Maletrain]

James, Just to make sure that there isn't something we don't know about the command station, I would put another loco on the program track and make sure I can read the CVs without getting the runaway effect.  And, I would check whether the command station allows for throttle control of a loco on the programming track, such as lights on/off, horn, and motion.

[peteski]

If the loco runs in the same direction regardless of the DC (9V) battery polarity tells me that the decoder is toasted. That's an important clue. Either the microcontroller (the decoder's brains), or the H-Bridge circuit which is used to produce the variable voltage of changeable polarity has a problem.

This is confirmed by the programming track behavior. That track only turns on power when programming packets (voltage) are sent to the decoder, so the model (somilar as  when in DC) starts running in one direction as soon as the power is present. Same happens when placing the loco on DCC powered track.

The constantly running motor also explains inability to program the loco on programming track. After the command station sends a programming packet it expects acknowledgement from the decoder in a form of a current pulse. The decoder does that by briefly turning on the motor (causing the model to twitch). That produces the current pulse the command station is expecting to tell it that the decoder received the programming packet.  With the motor running as soon as the programming track is powered, the current consumption is higher and constant, even before the programming packet is received, and no acknowledgement pulse can be generated.

This has nothing to do with DC bit in CV29.  As I mentioned earlier, a normally behaving decoder with the DC bit turned on will not run away when valid DCC pulses (of constantly alternating polarity voltage) are present. The runaway occurs then DC power is sent to the track by the DCC booster (usually when the DCC system initializes after being powered on.

But either way, I don't see that there is anything you can do  to fix it.  Why the decoder failed when it did I  have no explanation. Programming track is usually safe for the decoder.  I suppose it won't hurt to check if the motor leads are isolated from the track pickups (or models frame), but again the fact that the motor runs only in one direction to me seems to prove that the motor is isolated from the track pickups.

[justjames]

James, Just to make sure that there isn't something we don't know about the command station, I would put another loco on the program track and make sure I can read the CVs without getting the runaway effect.  And, I would check whether the command station allows for throttle control of a loco on the programming track, such as lights on/off, horn, and motion.

I can read and program to other locos on the programming track but there is no throttle control while on the programming track for lights, sound or motor. I've tested this to be sure this morning and all seems normal with other locos.

[justjames]

If the loco runs in the same direction regardless of the DC (9V) battery polarity tells me that the decoder is toasted. That's an important clue. Either the microcontroller (the decoder's brains), or the H-Bridge circuit which is used to produce the variable voltage of changeable polarity has a problem.

peteski,
Your detailed post makes sense to my feeble brain but its getting way beyond my skill/knowledge set. I appreciate your input as it is more evidence of the value and knowledge base on this forum.
Thanks!