Joos Buijs

[Joos Buijs]

Hi Edwin,

great to read that upgrading helped with the majority of your problems.

There is indeed a lag between the sensor trigger and the train actually stopping.
A few things I learned that might help:
1. I try to set the min. speed of trains as low as possible
2. Place in sensors in the third track piece from a switch for instance to allow for enough room. If a block is always travelled in the same direction then the enter sensor is best placed as early as possible in the block.
3. I disabled BBT (Block Brake Timer) and set Rocrail to go to min speed as soon as a train is entering a block and should stop (block settings -> details -> stop -> min).

For me this worked in preventing crashes 95% of the time.

Hope this helps!

[Joos Buijs]

Hi Edwin,

Could you upgrade to the latest 0.6.1 version? There were several significant improvements made in the meantime.

I ran 4 hubs on one MTC4BT controller for over 90 minutes this week so it should be doable, although I don’t have experience with the MTC4PU code and stability.

Please keep us posted!

Kind regards,
Joos

[Joos Buijs]

Hoi Martin,

De voorkeurstaal is idd Engels, maar Duits en Nederlands mag ook 🙂 Anders post ik wel een slot-post met in het Engels het probleem en de oplossing 🙂

Kun je de foto’s ergens anders uploaden en de links hier delen? Dan kunnen anderen ook makkelijk meedenken en -lezen.

Groet,
Joos

[Joos Buijs]

Hoi Martin,

Welkom op het forum! Leuk dat je je Lego treinbaan wil automatiseren met MattzoBricks, hopelijk kunnen we je helpen het aan de praat te krijgen.

Ik neem aan dat je een esp32 als pcb hebt gekocht om een MTC4BT aan de praat te krijgen die je sbrick kan aansturen?

Kun je aangeven waar je in de stappen zoals hier beschreven vast loopt?: https://mattzobricks.com/controllers/mtc4bt

Met vriendelijke groet,
Joos

[Joos Buijs]

Hi Luuk,

Thanks for reporting your issues!

If I recall correctly I had similar issues back in October and November. For me the cause seemed to be the ESP8266 variant, which seemed to cause issues at the MCP23017 side. Another ESP8266 with the same MCP23017 board worked fine.

Does this help? 🙂

[Joos Buijs]

Hi Hakan,

I guess it does not matter that much, as long as you can connect via USB to the EPS8266 for the power supply. I guess the charging port is handy but not a strict requirement.

Hope this helps!

• This reply was modified 2 years, 3 months ago by Matthias Runte.

[Joos Buijs]

Good to see you got some steps further to a working state.
Your log however says ‘connecting to 3 hubs’ which seems odd to me. Could you double check that you’re using the correct configuration file? There are several config files.

I can dive deeper into your issue next week if necessary, weekend obligations 😉

[Joos Buijs]

Hi Max,

Great to read that you’re willing to try out the MC4BT.

From what I read from the logging is that you have a configuration set up which sets up LEDs connected to the esp32. The one on pin 23 seems to cause the esp32 to crash and disconnect, causing the final communication error.

Could you remove the LED configuration from your config file all together, or at least verify that it’s correct?

Hope this helps, please keep us updated in your progress!

Kind regards,
Joos

[Joos Buijs]

Hi Dirk,

Happy new year, and great to read you’re happy with the project 🙂

I used the MTC4BT controller extensively and the main benefit is connecting hubs: the order, and even the time the hubs are on does not matter any more. I have changed batteries while the other locos remained running, and without rebooting the MTC4BT the train connected again and was ready to continue its journey.

Hope you find the MTC4BT as useful as I do!

Kind regards,
Joos

[Joos Buijs]

I’m happy to report that I solved the issue!

The cause apparently was not the board firmware version, MCP library version, resistors, or USB power source. This week I ordered a new (and different type) ESP-8266 and this one does not have this issue (with the same MCP, resistors, power source, code, etc.).

So, for me this one did not work in combination with an MCP23017:
Ai-Thinker NodeMCU-8266 – ESP-12S
This one works:
ESP8266 NodeMCU V2 which has an ESP-12F (in my case)

If other users have a similar (or different) experience I would be happy to know!

[Joos Buijs]

Hi Edwin,

Thanks for your message. I did not let my layout run for longer periods yet, but could you indicate after which time period you notice a delay? (minutes, an hour, 2 hours, etc.).

As I will encounter this issue as well soon I might look into it further in the coming weeks (no promises though :)). I’ll keep you posted!

[Joos Buijs]

This morning I found that the resistor I added via a Dupont cable sometimes lost contact, causing the MCP to go crazy and sending all kinds of noise signals. As a remedy I connected RESET directly to 3V3. I can now run my layout for a couple of minutes before it goes crazy again, but I’ll buy a better resistor and see if that helps.

I hope that anyone with similar issues reads this in the future: check your RESET resistor!

[Joos Buijs]

Hi Mattze,

Thanks for your reply.
I haven’t been able to troubleshoot much but I managed to get the sensor inputs working with a small sketch using similar code as the layout controller, so no hardware or library causes. The layout controller works fine until I add the third sensor on the MCP23017. Then it goes crazy with a stream of false triggers…

I was going to add debug messages to the code anyway to further debug so I might as well do this on the 0.5 version and help you guys. If you could send the details that would be great!

Kind regards,
Joos

[Joos Buijs]

Hi, unfortunately I cannot help you via teams for the coming week.

What worked for me yesterday when my layout controller couldn’t connect anymore was to clear the wifi flash. If you’re using the Arduino IDE this is an option under ‘sketch’. Usually this is set to ‘sketch only’ but you might want to try to set it to ‘sketch + WiFi’. If that doesn’t work you can try the ‘flash all’ option but note that this also clears your Mattzo ID, which means you need to change the rocrail settings (except for the MC4PU, so you’re lucky).

Hope this helps!

PS I’m not behind my computer so the names between quotes might be slightly different in the Arduino IDE.

[Joos Buijs]

Hi vanderwel,

This is a known issue in the 0.4 firmware. For my MC4PU I’m using the 0.3 version of the controller and Mattzo library. For the layout controller 0.4.1 works fine.

Hope this helps!

[Joos Buijs]

Hi all,

I’ve been playing around with the MC4PU and MattzoLayoutControllers for some weeks now, and I’m gradually increasing the difficulty until I can control my full layout for a show early December.
I already got the MC4PU working, as well as the layout controller without MCP23017 with up to 6 sensors and a (Trixbrix) switch.

I’m currently experimenting with extending the ESP8266 with the MCP23017. I have connected everything almost as shown on the layout controller page, except for the red wire bridge going to ground instead of 3.3V, which I believe is a mistake in the wiring diagram.
This is my current setup, with the resistor (temporarily) via a dupont extension, and a male-male dupont cable as sensor replacement (I have reed relais in the rails but this is easier for desk-debugging).

However, without the sensor on port A0 connected I get a lot of debug messages mentioning that sensor <x> got triggered, where <x> is one of the sensors I configured if I configured more.
Alternatively I sometimes get no triggers at all, even when I connect the sensor to GRND.

I have started the configuration file from scratch, and checked whether all Mattzo libraries and the layout controler files were identical to firmware 0.4.1, which they are, except for the occasional added debug message.

This is my configuration file for 1 sensor on port 0 (A0).

If anyone could point me in the right direction for further debugging that would be much appreciated.

Kind regards,

Joos

• This reply was modified 2 years, 6 months ago by Joos Buijs. Reason: replaced file download links

[Joos Buijs]

Interesting idea but this would mean that every train needs a Mattzo controller. For me, as I’m using powered up trains, which don’t need Mattzo controllers on the train itself, this would not be an improvement.

[Author]

Posts