I have just installed a new Digitrax DZ146 into a split-half Atlas GP35. It will not run and all CVs read back a value of 255. Only the orange and grey wires to the motor and the black and red wires to the frame halves are connected.
Have you verified that the motor brushes are isolated from the chassis? Is the bottom side of your docoder insulated? Does it rest against the chassis? Sounds like a short.
No idea if you wiped out the deocder but just in case.
Did you do an ohm meter check between each motor contact and rail pickup?
With one decoder install I was not sure on a split frame, top and bottom setup, and found out one motor contact was not insulated. Both motor brushes where connected to each motor frame half. I had to use Kapton tape between the motor frame and loco frame or I would have probably wiped out the decoder. I also used nylon screws to mount the motor.
My NCE Power Cab on program track might have caught it. It did one time when I swapped the orange and red wires when the lighting was not the best. Always best to use the program track, first.
19 for CV29 is reverse direction, short address, 28/128 speed steps, using a speed table. Not suitable for a decoder reset. You want 6, or 2 - 6 for short address, DC on, 28/128 steps, or 2 for shooort address, 28/128 no DC (no dc is preferred unless you really need to run your loco on a friend’s DC layout - turning it off in the decoder helps prevent the odd runaway when powering the system on.)
And not running is the other clue there is likely no connection between the motor and the motor terminals on the decoder - or a short. This seems to come up fairly frequently on N scale installs witht he board repalcements. Somewhere there seems to be a tolerance issue, perhaps with the thickness of the PCB. The power input contact on most of these is made by the locoo frame halves touching traces ont he decoder PCB. If this is a loose fit, there is likely no contact. It seems common now to have to add some solder blobs to get a tight fit. Second place for having an issue is that with the original light board, the motor and the track pickups are usually in parallel, the tabs from the motors come up and wrap over the board, and with DC there’s no issue if the motor tabs touch the frame. For DCC, this is death for the decoder - usually the instructions mention applying some Kapton tape to prevent contact where the motor tab folds over to read the connection points on the decoder. Used to be, a lot of N scale decoders would include small pieces of Kapton tape, not it’s mostly up to the modeler to supply their own.
Make sure both of these are taken care of and it should work fine.
Some decoders do seem to accept 0’s in CV 2, 6, and 5, and treat that as an unmodified straight line profile, though intuitively a flat line should be CV2 = 0, CV6 = 127, and CV5=255.
As a followup note, the decoder is hard-wired to the motor and the frame halves. The continuity between the wheels (or track) and the pick-up wires to the decoder is good and there are no shorts between the loco halves or between the motor leads (orange and gray) and the the loco halves. I checked these before I submitted the first post.
The Zephyr will not read anything if it cannot “see” the decoder. To me, reading 255 means it is “seeing” it.
Seeing 255 usually means a short. In readback mode the system starts at 255 and reads decoder info. 255 is the first thing your system is seeing and it stops. Have you tested frame to motor ? There should not be any continuity between the orange and red (frame) wires or gray and black (frame) wires. Or any motor connection to the frame!
Can you read any other decoders? If they all return 255 as well, I would suspect that the program track connections have been damaged and your Zephyr needs a vacation to Florida (sending it to Digitrax for repair).