Feature Request for wagon lightning only

[Space-Farmer]

Hello I am thinking currently about the possiblility of using your project as basis for an dcc controled wagon ligning.
Means that the Lighning of the wagon can be controlled like lightning or smoke of a lokomotive.
I hope you got my intention.

For that I want to prepare a pcb that is based on your work that would fit in the roof of the wagon.
Has LED with resistors and transistors on it to allow PWM dimming.
This PCB could be enlarged/shortend to fit the wagon.

Unfortunatly I am not that skilled in Software.
Therfore I kindly asking for your assistance to add this possiblity to your Software.

Does this sound reasonable?

[gab-k]

I mean, the functions do exist in the current version. From a software perspective, the decoder doesn't care whether the motor is used or not, so you could simply use the existing function outputs. I'm not sure what you're asking. Are the space constraints the challenge? You could, of course, remove the motor driver and its circuitry, which would not require any software changes I can think of. I hope this answers your question. Please feel free to clarify if I have misunderstood you.

[Space-Farmer]

Maybe I missed something. During the startup of my last HW batch the micro controller is checking the back-EMF Voltage before it goes to normal operation. And it stuck there until this process is successful carried out.

-> Yes from space and cost perspective I would remove the motor driver

Iam thinking of implementing a resistor to emulate the current needed for the DCC feedback communication.
Because using the LEDs might not generate enough load.

I hope this post clarifies my first request, sorry for being that imprecise.

[gab-k]

Maybe I missed something. During the startup of my last HW batch the micro controller is checking the back-EMF Voltage before it goes to normal operation. And it stuck there until this process is successful carried out.

I'm not 100% certain but briefly looking into the code I can't spot where this is done, in older software versions there is definetly a calibration process which requires the back-EMF Voltage however this is omitted for quite some time now. When did you try this and can you try the newest software version?

Iam thinking of implementing a resistor to emulate the current needed for the DCC feedback communication.
Because using the LEDs might not generate enough load.

This should work, but it might not even be needed, as the CV gets programmed anyway, you just dont get the confirmation.

[Space-Farmer]

When did you try this and can you try the newest software version?

It is quite a while ago, sorry. 12/2023
I flashed my batch with an I2C/SPI programmer and did not update them.
Sorry that I did not read the release notes.

I will try to update and double check. Might take some days.

This should work, but it might not even be needed, as the CV gets programmed anyway, you just dont get the confirmation.

Ok, I will mark this as optional.

Do you have some input what you would like to see on HW side?

[gab-k]

Do you have some input what you would like to see on HW side?

I mean it depends on your use case, lighting up a few LED's isn't really that big of a requirement i suppose.

[Space-Farmer]

in older software versions there is definetly a calibration process which requires the back-EMF Voltage however this is omitted for quite some time now
I tried the latest release and now I can confirm that there is no calibration process at 1st startup.

I mean it depends on your use case, lighting up a few LED's isn't really that big of a requirement i suppose.

I am currently struggle a little bit with the huge input voltage of 12V to 24 V
For that I see currently 3 options:
1.) DC/DC converter to ~6V to be able to have 2 LEDs in a row per string
Con:
- Might get some issues with interference of the DCC signal
- lots of additional parts
Pro:
- Best solution in regards to flicker reduction caused by connection interruptions

2.) Current source IC per LED string to limit the current
Con:
- Interference with PWM controlled LED strings possible
Pro:
- Part is easy to place like a resistor
3.) Usage of the WS2811 or similar, with integrated PWM and current sink
Datasheet
Con:
- SW afford to integrate library and CV to WS propagation
- Not clear if this IC can drive 3 white LED instead of RGB
- Maybe additional power regulation necessary
Pro:
- All in one solution (PWM generator, current source, load driver)
- Possible solution to light whole train with only 3 lines (+,-,Signal) and no further intelligence in the other wagons.
All LEDs are still fully controllable

Layout:

• The PCB is currently 260mm long
• on both sides there are "ears" that could be removed to fit into smaller wagons
• On both ends of the PCB there is the possibility for connecting the rail voltage
• On both sides there is the possibility to connect a head and a rear light with integrated resistors

Comments are welcome

[gab-k]

Sorry for the late answer maybe i can still give some tips:

I wouldn't worry too much about interference for now.

I think voltage regulation is not the smartest here, you definitly should prefer constant current sources, either use an IC like you said or perhaps something like this.

I havent really looked into the WS2811 datasheet into detail but i think there is no way for the WS2811 IC to know what color the LED has.

[Space-Farmer]

Hello @gab-k ,

no issue I was stuck in a quick Halloween project...

I am still not finished with my decission.

With the simple linear current regulation you might have less issues and complexity, but the thermal aspect has some downside.
With 12V as minum operating voltage, only 3 LEDs in a row makes sense.
With 3 LEDs you can have 2.6V -3.4V depending on the bin.
Having a maximum voltage on rails of 24V - 7,8V (LEDs )= 16,2V over linear regulator
Results with 20mA max. LED current to 324mW power dissipation.
Which results with an SOT223 housing

To a junction increas of ~55°C towards ambient temperature.

-> Should be no issue for the device

On point I was a concerned of is the total waggon warming.
Only LEDs and and linear regulator might result into 1,92W (24V 12 LEDs), which is more than the established lighning bulb that was used over decades.

-> I will reduce the number of LEDs or the max current, no real issue

You really should have a look to the WS2811 "neopixel" or "pixel LED" stuff.
I used this for my hallowee project together with WLED

Maybe this datasheet is easier readable: WS2812 LED

This technic opens infinit possibilities:

• Controll the brigness of the LED per CV entry per waggon section/compartment
• Control the brigness of LEDs along a set of carriges with only 3 wire connection
• Add indivdual effects per LED like fglashing( defect of flurescent tube), periodicly off/on (simulate passenger leaving compartment) ...

There are already many projects around that are using the "neopixel" with RP2040 like: Link

Maybe you can think about an integration of something like this.

Regards

[gab-k]

Hello @Space-Farmer,

That's an excellent suggestion. You are correct that using WS2811/WS2812 LEDs would be a much more efficient and flexible solution, though it would indeed require a DC/DC buck converter on the hardware side.

While I'm excited by the possibilities, I am also a bit concerned about adding too much complexity to the core project. My goal has always been to keep the software relatively simple and maintainable. A feature this advanced would add significant code and a more complex CV structure, which could be overwhelming for users who just need basic functions.

Perhaps the best way forward would be to implement this as an optional, compile-time feature. This way, users who want the advanced lighting could enable it, but the default firmware would remain lean and simple for everyone else.

I still don't have the bandwidth to develop this myself, but if you were interested in contributing, I'd strongly encourage you to approach it as an optional module. I would be happy to provide guidance on that structure and review a pull request.

What are your thoughts on that approach?

Regards