PCB LED - should it remain on, or turn off after a while? and loco not running except for the odd jerk

[Ma77h3w]

Hi all,
I was given some of the iLabs version 2 of the RP2040 decoder (without USB) and have wired up one to a small 0-4-0 locomotive with just motor and track pick ups connected.
When I put it on the track the LED on the PCB comes on and the motor turns a little bit and then often the LED goes off and stays off. It comes back on if I nudge the loco.
The loco was running fine on DC analog. But it almost seems like it's losing power like bad pickups.
I've not managed to get the loco running, and it fails to respond to commands on the programming track.
Occationally, but not too often, the DCC controller warns of too much current:
<* TRACK A ALERT 2148mA >
< TRACK A NORMAL (after 20msec/20msec) 0mA.
Also the troubleshooting link at the bottom of this page doesn't go anywhere: Wiki - Getting Started
I've measured the AC voltage of the track and it's 14Vac. I'm using the DCC-EX controller.
On the programming track I ran the diagnostic and it fails to get the Loco ID.

I have a larger Class 4P 2-6-4T, much older engine (80's) and another of these decoders I may try. Although I need to check if 2 Amps supply / decoder is OK on the older models. Although that runs well on DC.

[Ma77h3w]

Update: I've temporarily hardwired the track power directly to the decoder, and it's now working. So I must not have good enough track pick-ups for it to boot up and test properly.
To answer my own question on LED, it takes a couple of minutes of the decoder wiggling the motor and then the LED goes out.

If anyone has any suggestions on improving track pickups or perhaps adding a stay alive capactor that would be great!

[Ma77h3w]

Further update + solutions.
I added a DIY stay-alive circuit. I found I had a 470μF 16V capacitor, 100Ω resistor and a diode laying around. Connected that up to the decoder and suddenly it's running much better! I think I just about have space for a 1000μF ⌀10 x 16mm in this tiny loco's boiler area above the worm gear, so I've ordered some of those. I've also ordered some contact cleaner for the wheels as I'm sure that will also help.
For anyone having similar issues the circuit is the one on the first post here https://www.modeltrainforum.com/threads/talk-to-me-about-diy-style-dcc-keep-alive.205233/
There are off the shelf keep alive / stay alive things but they all seemed to big for this 0-4-0 steam loco.
Further note: elsewhere I've seen mentioned "Disable DC operation in CV 29 & set CV 11 to “0”, when using Stay/Keep Alives. " but I haven't tried that yet.

[gab-k]

I suppose you are using this software?
https://github.com/PontusO/RP2040-Decoder

The software is using a calibration process at factory reset which might be causing you trouble. I removed this in the newest software in this Repository. However you cant just use the software with your hardware without changing pin definitions.

This is the schematic for your board:
https://usercontent.one/wp/ilabs.se/wp-content/uploads/2024/10/99-00299-1_V2.0.pdf?media=1689860297

This is the schematic for my USB version:
https://gab-k.github.io/RP2040-Decoder/sphinx/html/_images/RP2040-Decoder1.svg

I dont really know what the LED does in the iLabs software version in Pontus' repository.
What i can tell you is that currently in the Software in this repository the LED just turns on for every error condition. So no real way to tell which error happened at the moment, for further investigation you currently would need a debugger. I suppose that does not really help you because you are most likely not using the software from this repository.

[Ma77h3w]

You're right the iLabs points to your documentation for software support, but then also links to that PontusO fork so they're using that instead due to slightly different pin-outs and lack of USB. I didn't realise. Looks they they changed the 3.3V supply and used ADC1 to monitor the rectified voltage and then shifted the pins up 1.
So it won't necessarily have the latest updates yours has. The LED was remaining On during that initial calibration, but once it managed to complete that (after I added the stay alive), it now only flickers the LED briefly occationally when going over bad track.