My first big decoder issue- DN163I1B in IM FTB

[thomasjmdavis]

Per the subject line, having difficulty installing a DN163I1B into an Intermountain FT-B.  This decoder is the "drop in" for the loco.  Using JMRI with DCC++.
This loco is one of an FT ABB set (one B was purchased 2nd hand).  The A and the other B (which has an identical decoder) work perfectly- not quite "drop in" because headlight has to be wired and the mounting holes in the board do not quite line up with the holes in the frame, but close.  No difficulty in programming.  Work perfectly.  Also have done a TCS CN, DN126 and DN166I-something in an FP-7 with no difficulty I could not solve.

This B, well.....
PROBLEM:
On either an operational track or the programming track, the loco takes off running.  Runs several inches to a couple feet.  Does a little "dance" rapidly reversing directions several times- takes off running again- another shuffle back and forth- etc. etc.

If I hold it in place, I can program it. 

Things I have looked at:
swapped decoders with the "good" B unit- the problem follows the decoder.
Looked at CVs of the matching units, and other than address, they appear identical. With one exception- CV7 (the "version" #) of the bad decoder occasionally returns either 255 or 127, instead of 0. 

Checked everything I could think to check looking for shorts (before I swapped it with the other unit), and wrapped some kapton tape over the brush caps just to make sure nothing was touching the frame.

I have a spare decoder- exhibits the same problem.

So....am I experiencing a quality issue with this particular decoder (2 out of 3 failures), or is this some widely known phenomenon? Is there a simple solution other than buying new decoders?  I haven't been working long with DCC. Thankfully, so far, the decoder and throttle functions of JMRI really are intuitive to me.

[central.vermont]

Tom have you tried to do a factory reset?
Definitely sounds like its a problem with the decoder and not your install.

Jon

[MK]

I had a similar but not exactly the same problem.  Digitrax decoder for Kato and not IM.  It would program fine on the programming track.  But as soon as I drop it on the main, it takes off at full speed.  I didn't even selected the loco yet let alone gave it throttle.

Put the decoder in a different loco and problem followed the decoder.  Inspecting the decoder really carefully I spotted a tiny solder bridge on one of the chips near the center of the board.  Fired up my soldering iron and removed the bridge.  No longer off to the races as soon as the wheels touch the main.

Digitrax has a one year No Goof warranty.  Just send both back for replacement or repair.  No need to buy new ones.

[thomasjmdavis]

@central.vermont
I did a try a couple resets- I did discover in one of the earlier installs that the decoders do not always come from the factory in "factory default" so I do one when I first put it on the program track, and then again if I have some difficulty getting things to work as expected.

@MK
I did give it a look, but (a) just with my bifocals and (b) I was looking for a missing chip or something obviously broken- I will get my magnifier out tomorrow and give it a closer look for little solder blobs.  Thanks for the idea.

[thomasjmdavis]

Checking on the decoders with a magnifier reveals that the ceramic element on the "bad" decoder- just below the screw head on the board on the left of the photo- appears cracked.  Is this a resistor?  If I check across the poles of that part on the decoder that works (on the left), it registers 10 ohms on my little and cheap meter.  Several hundred ohms when I measure across the cracked one on the right. (A cheap and not particularly accurate meter, and possible I am not touching the correct points- although I repeated the test several times with same result)

[MK]

Tom, it sounds like a resistor if you measure 10 ohm (of the good deocder).  Can you circle the component you are talking about and maybe @peteski can help you out?  I may be able to too if it's a component with some markings.

What about the second bad decoder?

[thomasjmdavis]

Tom, it sounds like a resistor if you measure 10 ohm (of the good deocder).  Can you circle the component you are talking about and maybe @peteski can help you out?  I may be able to too if it's a component with some markings.

What about the second bad decoder?

MK,
The crack is more visible in the shot below, and part now circled in orange.  The second bad decoder, same part is missing.  I think it may be that the screw (the original that held the light board prior to decoder install) may have broken the parts when screwed in.  As can be seen in the photos, there is not much leeway between the screw and the part. (These are original run FTs, and the board screws are difficult to start because the holes in the board do not line up well with the screwholes in the frames.)

[peteski]

I don't have one of those decoders (or at least not very handy), but that component looks like a carbon-film resistor. Those have ceramic bodies with carbon film applied on top surface.  I agree that it was likely cracked while tightening the screw.

The resistor itself is cheap, but unless you are comfortable working with small components, and you still have receipt for the decoder, then sending it back to Digitrax for repair makes sense.

The metal washer under the screw looks dangerously close to the metal pads and components on the decoder - I would either remove it, or replace it with a plastic or fiber washer.

[thomasjmdavis]

Thanks all for ideas/suggestions.  I only need one of the 2 non-working decoders, so I think what I may do is send one back to Digitrax, and then attempt a repair on the second- if I can find appropriate parts- will let me test out the really tiny tip on my soldering iron.  And will be good practice.  Worst case, I will ruin a non-working decoder.

Tom D.

[peteski]

Are there any numbers imprinted on that resistor, (like on most other resistors on that board), or is is just blank?

[thomasjmdavis]

Are there any numbers imprinted on that resistor, (like on most other resistors on that board), or is is just blank?

It is so small that there is nothing visible with my magnifier lamp- just appears to be a small black rectangle. On my monitor, the photo is approximately 5 times the size of the N scale frame and decoder.  Maybe with a microscope?

[peteski]

It is so small that there is nothing visible with my magnifier lamp- just appears to be a small black rectangle. On my monitor, the photo is approximately 5 times the size of the N scale frame and decoder.  Maybe with a microscope?

The numbers would not be very tiny. They would be large enough to occupy the black rectangle (like they look on the other resistors which have numbers on them).  I just thought hat maybe the numbers were printed with a darker ink which was not readily visible in the photo.

Measuring the resistance in-circuit can lead to inaccurate readings - I just wanted to make sure that you find the correct value for the replacement resistor.  It could be 10, 12, or 15 ohms.  Multimeters are not very accurate in the low ohm range, and the rest of the circuit affects the measured resistance value.  Also make sure that when you short the test leads, it shows zero ohms.  Then measure the good resistor twice, flipping the test leads (so you measure it using both multimeter polarities, then take the average value).

[thomasjmdavis]

@peteski

As it happens, today's mail brought some DN163I1C (decoders for IM F7s). These are identical to the ones for FTs, except for the position of the posts underneath that contact the motor lugs. The numbers on this are much easier to read.  The part is question has the number 103.  Which if I remember the coding should be 10x1000= 10,000 ohms.  And at 10,000 ohms, wattage negligible (144/10,000 assuming 12V).

PLEASE correct whatever part of the above I got wrong, since I probably did.  Which reminds me, I need to call my son, the robotics engineer.  Although most of what he does is programming, he did plenty of circuit board design in college.

[thomasjmdavis]

Here's a pick showing numbers-
[ Guests cannot view attachments ]

[MK]

Tom, 103 is indeed 10k ohm.  And your wattage calculation is also correct.

Carry on!