Computer Control - Roco Multimaus & JMRI, Arduino, CMRI

[James 1707]

Hi,

 

I'm very new to DCC and I've got my self a Roco Multimaus 10810 thats working great with my single DCC Class 26. I also got a Lenz LS150 accessory decoder that I can change 2 points I've got connected, great fun.

 

I'd like to link my lapout to my PC and use JMRI to run my train automatically and change the points using the JMRI scripting (I'm ok with that bit)

 

My question is it possible to connect a Lenz LI-USB interface to the slave port on the 10810 to achieve this? Is there any other interface that I can connect my 10810 to my computer.

Has anyone ever done it?

 

Thanks very much. Jim

[Nigelcliffe]

JMRI documentation page:

http://www.jmri.org/help/en/html/hardware/XPressNet/index.shtml

 

Shows that a MultiMaus supports most things other than hardware feedback (ie. output from sensors around the layout; if using those you need another mechanism to get the sensor data back into the computer)

 

 

Interface options;

the Lenz adaptor is a bit expensive though should work. Lenz have said its due to be replaced (but Lenz have announced things before then never released them..)

The XnTCP solution looks better, has good end user comments and is cheaper, but I've not built one. As DIY electronics projects go, its probably about as complex as a decoder hard-wire and lighting installation.

 

 

- Nigel,

[Ron Ron Ron]

Roco DCC kit is Lenz designed or made.

Most of the Roco electronic add-on modules (accessory decoders etc,) are just re-badged Lenz ones. Roco's own computer interface may well be exactly the same or similar to the Lenz module.

 

Incidently, Roco have their own layout control software to go with the MultiMaus.

It's called RocoMotion and is made by Freiwald, the producers of RR&co.

 

.

[buffalo]

...The XnTCP solution looks better, has good end user comments and is cheaper, but I've not built one. As DIY electronics projects go, its probably about as complex as a decoder hard-wire and lighting installation.

 

I have an XnTCP interface and can confirm that it works well with JMRI and is very easy to build given basic electronic soldering skills. The instructions on the web site are very complete. It uses an ethernet connection to the PC and an XpressNET connection to the DCC gear. JMRI sends network packets directly to the XnTCP and no other software is needed. I've only used it with Lenz kit, but it is claimed to work with the Multimaus.

 

Nick

[James 1707]

Hi,

 

Just want to say thanks for all the info. I didn't know anything about XnTCP but it looks great and could be fun putting it together.

 

Many thanks. Jim

[James 1707]

Hi All,

 

Just thought I'd give an update. I built the xntcp board as suggested. Quite straightforward even with my shakey hands. A multimeter was a must for testing the links.

Couldn't get it to power from the Roco command station, so I got a plug a 9v battery in and got it going.

 

Linked first time to JMRI and I got a simple script running making a train going back and forth automatically. Great.

I don't think my Roco can do feedback so I might need to upgrade to a Lenz 90.

 

Thanks for all the help.

Jim

[Tim Hale]

Roco DCC kit is Lenz designed or made

 

Hi,

 

May I correct that small rumour, Roco ceased to exclusively use Lenz with the introduction of the MultiMaus.

 

The MultiMaus was a product of an independent Austrian electronics company which was having difficulty providing support for their DCC control system, the company approached Roco who offered to take on the MultiMaus and sell as their own. To complicate matters even further, Roco* have been using ESU LokPilot 1.0 decoders with the introduction of their BR64 and have recently used ESU re-badged decoders in their sound-equipped locos.

 

However, Roco did use Lenz products and I still have an accessory Roco decoder with Lenz instructions.

 

DesA

 

*As do Bachmann- the later versions of the 36-553 are programmable LokPilot 1.0's.

[Ron Ron Ron]

Thanks for that correction Derek.

 

I see that Roco still sell Accessory decoders and other modules made by Lenz, but was unaware that the change over from the LokMaus to the MultiMaus spelled a change in supplier for the control system.

 

Do you happen to know who supplies their new PluX16 & PluX22 decoders?

[TWG]

Thanks for that correction Derek.

 

I see that Roco still sell Accessory decoders and other modules made by Lenz, but was unaware that the change over from the LokMaus to the MultiMaus spelled a change in supplier for the control system.

 

Do you happen to know who supplies their new PluX16 & PluX22 decoders?

 

Roco have changed their decoder supplier so often that it is hard to keep up. They have used products from ESU, Kuehn, Lenz and Uhlenbrock. I have read on a German forum that the PluX decoders are supplied by Zimo.

 

This is the first time that I have heard the information about the Multimaus. I feel sceptical. Which Austrian firm made or makes it? I am interested in digital control and model Austrian railways, so I am surprised that this is new to me. There is a firm called HKE, an established electronics maker, which brought out an excellent digital system; it did not meet with the commercial success that they had hoped for so it was withdrawn from the market. Their controller was nothing like a Multimaus. It was a radio handset which could also be plugged in and was (is) first rate - I have one. It looks exactly like a CT Elektronic handset with different buttons. There was a firm called ConRail, which announced a mouse like controller a few years ago (if I remember correctly it was called a Wildmaus), but it did not look like a Multimaus and was considerably more sophisticated. Regrettably the firm collapsed in its early stages. So I am curious to know which firm is meant.

[Nigelcliffe]

Just thought I'd give an update. I built the xntcp board as suggested. Quite straightforward even with my shakey hands. A multimeter was a must for testing the links.

Couldn't get it to power from the Roco command station, so I got a plug a 9v battery in and got it going.

 

Linked first time to JMRI and I got a simple script running making a train going back and forth automatically. Great.

 

Thanks for the feedback on making the XNTCP board. Its good to hear someone get it working. And its far cheaper than the Lenz official product.

 

 

- Nigel

[James 1707]

Well it's been a few years since I started this question but I thought I'd add an update as to where I've now got too.  Computer control was the original plan and I got the command station talking to the computer running JMRI.  The next 6 years on and off has been spent trying lots of things to get a cheap solution for point control, signal control and block occupancy.

Over the coming weeks I'll update on how this has gone and going to see if there is any interest and also to get some help with some stuff I've yet to solve.

 

For starters I've attached a diagram of the bits so far.

 

 

Computer Control Model Rail.pdf

[James 1707]

Arduino CMRI

 

I’ve taken the fantastic ideas of Michael Adams at http://www.utrainia.com/65-arduinocmri-and-rs485 to configure my CMRI nodes to control the servos and block occupancy sensors.

 

As in Michaels article I’ve chained 3 nodes together using the MAX485 chip.

 

In my set up I’ve used 3 x Arduino Nanos (which cost me about £1.50 each) and two MAX485 chips again not very much money on Ebay.

 

I’ve attached a picture of my CMRI control board.

[James 1707]

I've now attached a servo to control a point attached to the CMRI2 arduino.  I'm using the SG90 Servos that are very cheap.  I've mounted them on a bit of plywood, it seems the way some people have done it. 

 

I've connect the servo to a separate 5v supply rather than drawing power from the arduino.  You will see the brown control wire attached to digital pin 3 on the arduino to send the signal to control the point.

 

I've seen it suggested of adding a 330ohm resistor to the control wire to protect the arduino and would welcome a view on this?

 

I've attached some images and the arduino script I've pulled together. It all works great from JMRI when I created CT2003 to open and close the point.

 

I now just need to think about scaling the script up a bit to support all the other points.  Again I'd welcome a view on coding (an array?) rather than just repeating the same code over and over for the other servos.

 

Thoughts welcome Jim

script1.pdf

[James 1707]

An update now on Block occupancy.  I spent days reading loads of articles on the internet and trying various circuits and spent a bit of time trying the ACS712 current sensor that from China is cheap. 

The N Scaler website has done a fantastic blog on all this http://thenscaler.com/?p=514 and I got it to work however it did not work great when a loco sat idle in a block.  However its great research and worth a look.

 

Oh forgot to say my Arduino Nano cmri nodes each support 17 inputs/outputs.  So with my 3 I can use it for a total of 51 servos or sensors or a mixture.  Not bad really for £15 or so.  By just adding another MAX485 chip and another Arduino you have another 17.

 

Right so after moving on from the ACS712 and other circuits I finally settled on Rob Paisleys vt5 circuit that uses transformers http://home.cogeco.ca/~rpaisley4/DccBODvt5.html Rob will send you the circuit boards and its easy to set them up.  My soldering skills have improved and it takes me about 50 mins to out one together.  Parts are cheapish the transformer being the most expensive bit at about £1.25.  I reckon I’m getting a block detected for about £2.50.  Not sure if that’s good value or not but the vt-5 is very reliable (so far).

 

I’ve added some photos and w ill put a bit about the code on the Arduinos next time.

[James 1707]

Arduino code for block detection.

 

Well this took me a long time to get my head round especially as this all links back to JMRI on my laptop.

 

So I’ve wired the D3 (Green wire in picture), D4 (Grey in picture) and the Grd (Black wire in picture) back to the VT-5.  The VT-5 also needs some power to drive it.  I’m using a 5V supply from an old mobile phone charger. Once I’m ready I with all my blocks I’ll add a better supply, but for testing this is just fine.

 

The Green and Grey wires connect to the open collector outputs on the VT-5.  Don’t forget the ground.

 

Code now.  So I’ve just messed about with the CMRI library input and output example a bit.

#include <CMRI.h>

 

CMRI cmri(0, 24, 48, bus); // I never actually use all 24 inputs and 48 outputs on my Arduino Nano

 

pinMode(3, INPUT); // listening on D3 for the open collector output from VT-5, my green wire.

pinMode(4, INPUT); // listening on D4 for the open collector output from VT-5, my grey wire.

digitalWrite(3, HIGH); // set D3 HIGH for starters

digitalWrite(4, HIGH); // set D4 HIGH for starters

 

so that’s the setup function stuff.  No the main loop funtion.

 

cmri.set_bit(1, !digitalRead(3)); // this now binds D3 to the CMRI bit 1

cmri.set_bit(2, !digitalRead(4)); // this now binds D4 to the CMRI bit 2

 

Now for the crazy bit.  In JMRI I’ve already added my CMRI nodes 0 , 1 and 2.  Remember 0 is the node I’m interested in here.  I’ve then added the sensors in the sensor table. CS2 and CS3. 

 

It took me a while to get my head round the relationship to the pin on the Arduino. CMRI and JMRI however I think its because the CMRI inputs starts a 0.

 

Arduino Pin                ->                CMRI bit           ->                JMRI sensor

D3                                           1                                              CS2

D4                                           2                                              CS3

 

Anyway its working and is consistent.    Thoughts and comments welcomed. Jim