Forum Replies Created
-
Posts
-
Ich hoffe die Bilder, die ich Dir per E-Mail gesendet hatte, haben ein bisschen geholfen.
Ich habe in den letzten Tagen ein paar Regressionstests gemacht und bin auf einen witzigen Bug in der Firmware gestoßen, der in speziellen Konfigurationssituationen auftreten kann. Dieser ist mit Release V1.0.1 behoben. Bitte Firmware updaten, falls Du auch die orangenen Servo-Motoren für die Brücke verwendest.
Happy to hear that. I have reviewed all of the MLC example files. There will be a lot of changes, updates and improvements in firmware version 1.0.1.
Jepp, it’s the bug introduced by the MLC example configuration files in the bascule bridge configuration.
You need to do two things:
1.
Please remove these lines (in the bascule bridge configuration section)://Must be defined, but are zero
#define NUM_SIGNAL_ASPECTS 0
#define NUM_SIGNAL_LEDS 0
#define NUM_SIGNAL_SERVOS 02.
Please change (in the signal configuration section):>>
// Number of signals
const int NUM_SIGNALS = 8;
// Maximum number of signal aspects (e.g. red, green, yellow)
const int NUM_SIGNAL_ASPECTS = 3;
// Number of signal LEDs (usually equal to NUM_SIGNAL_ASPECTS)
const int NUM_SIGNAL_LEDS = 3;
<< into // Number of signals #define NUM_SIGNALS 8 // Maximum number of signal aspects (e.g. red, green, yellow) #define NUM_SIGNAL_ASPECTS 3 // Number of signal LEDs (usually equal to NUM_SIGNAL_ASPECTS) #define NUM_SIGNAL_LEDS 3 << Then it should be working. Cheers, MattzeMeanwhile, could you check if you still have the following code in your controller_config.h in the level crossing section?
// BASCULE BRIDGE CONFIGURATION
// General switch for bascule bridge (false = no bridge connected; true = bridge connected)
bool BASCULE_BRIDGE_CONNECTED = false;
// Must be defined, but are zero
#define NUM_SIGNAL_ASPECTS 0
#define NUM_SIGNAL_LEDS 0
#define NUM_SIGNAL_SERVOS 0This was the problem that I fixed yesterday. You need to delete these defines from your configuration, as these values override the mentioned parameters by accident.
Hi Fabian,
you are doing everything fine. There is a bug in the example files for the MLC in release V1.0. This is the reason why the light signals do not work. I just found bug myself this evening while uploading the firmware to one of the “MLC megas” that I am presently building.I just fixed the bug and pushed the bugfix it to the development branch. The fix will be included in the next release 1.0.1, which we are going to release soon.
If you know how to check out the development branch, just go for it. If not, drop me a message, then I will help you.
But thanks for your feedback anyways!
Cheers,
MattzeEine IO-Datei habe ich leider nicht. Von dem Prototypen gibt nur die Bilder auf der Website und das Video.
Wenn ich Zeit und Lust habe, baue ich aber noch eine anständige Brücke auf Basis der Trixbrix-Brückenelemente, die voraussichtlich 6 oder 8 Standard-Geraden lang sein wird. Vielleicht mache ich dazu eine IO-Datei.
Wenn da Lampen blinken, nimmt der MLC ja zumindest schon mal Kommandos von Rocrail an. Das ist schon mal sehr gut und 90% der Miete. Funktionieren die Weichenmotoren denn am MLC ohne Port Extender?
We have already done some preparations for that, also on the hardware level. We are planning to work on this matter in the future.
Happy to hear that. Somewhere on the website there are some tipps for improving performance of the automation. Did you find that?
May I also propose to shift the discussion to the MTC4BT topic? This is the MTC4PU forum. 🙂
Thanks!
We have optimized some internal settings within MTC4BT to speed up processing. We will publish a new release soon and also plan to open up the git repository to the public. Stay tuned…
We are working on that for both MTC4BT and MTC4PF. The aim is to correct the train speed for decreasing battery power as long as possible.
Hallo Volkmar,
wir haben das Problem identifiziert und mit der neuen Firmware 0.6 hoffentlich auch gelöst. Hintergrund war, dass abs() keine unsigned integers schluckt (eigentlich logisch), und daher bei einigen Compilern Fehler wirft.Schau mal, ob es jetzt funktioniert, falls das Problem noch aktuell ist.
Gruß,
MattzeHello Karoly,
we investigated the abs() problem again and we found the solution for it (or Hilbert did, to be precisely).I have just released firmware 0.6. The problem should be fixed now.
Cheers,
MattzeTrue. If you’ve never worked with microcontrollers before, it’s a steep hill to climb. But it’s worth it and many people have valued up their layouts with our train automation system. We appreciate every new user and layout, so: welcome to the club!
Hi Max,
that sounds pretty cool. I have a lot of narrow gauge tracks, switches and crossings here, but I just played around with them a little bit – nothing serious yet. I didn’t have time to get into the science of building rolling stock for narrow gauge yet, so I have no practical experience in that yet. But I have a strong desire to dive into that matter later. So, maybe I ping you one day for exchanging a bit more of your experience.Did you check out the track planning section on Mattzobricks yet? I have added narrow gauge tracks to the Trixbrix library for Bluebrick, so Track Planning should be quite easy also for narrow gauge.
Hi Max,
ad 1: well, while traffic also does play a role, it’s more the number of connected Bluetooth devices that is relevant. It’s no deterministic science as of now, but we are tracking the problem down as good as we can.Ad 2: will be hopefully resolved soon.
Thanks for your patience.
Hi Max,
I believe your problem relates to a bug in the MTC4BT that we are presently tracking down. At present, I would recommend not to use too many PU or SBrick units with one MTC4BT. Using only a limited number of hubs improves the stability of the controller significantly. Our hope is that we get the problem fixed soon, then it should no longer matter.So, the short answer is: invest in one or even two more controllers.
Cheers,
MattzeFor the Arduino based controllers (MLC and MTC4PF) it should be quite simple: just add username and password to the connect method of the PubSubClient. Check https://pubsubclient.knolleary.net/api#connect
I will ask Ray to check how this would be done for the MTC4BT.