point switching problem with Roco/MERG combination

[chaz]

A very strange fault has arisen today on my On30 layout.

 

First a description of my setup.

 

My DCC system is a Roco Multimaus  with two handsets, one master, one slave.

 

My pointwork is all handmade using the Fast-tracks jigs and all have free-moving tie-bars and blades.

 

All the points installed (eight so far) are thrown by Tortoise motors which are connected to the DCC with MERG accessory decoders - three in total.

 

To work the points the button on the MM that switches between locos and points is pressed so that the display will show the number of the point last selected and its position - straight route or curved.

 

Until today the points have thrown perfectly with the Tortoises working as they should and the blades seating against the stock rails properly.

 

Measuring the DC fed to the MERG decoders with a meter shows 15V with no load.

 

Now the fault.

 

Except for one point all the points hardly move, they just start to move and then just stop. The one exception works as it should. The handsets show the points positions as they should be. It's as if the MERG decoders have been reconfigured to momentary pulses, rather than the continuous output necessary for the stall motors in the Tortoises.

 

The only different thing recently connected is a turntable drive - the motor is controlled by a Lenz 1017 decoder wired to the track feed so it acts effectively like a locomotive (with the address number 99) and does not take power from the DC point power. This motor (not currently connected) draws much less than an amp so should be no problem. 

 

This fault has got me baffled. And why all but one of the MERG outputs? Very strange.

 

Has anybody had anything like this happen?

 

What do I do to correct the fault?

 

Chaz

Edited September 18, 2017 by chaz

[chaz]

I've just had a thought - could it be that some PoM setting of CVs of a Soundtraxx Econami has affected the MERG boards? And if it has how do I reset them? I should say that I am no longer a MERG member.

 

Chaz

[chaz]

The more I think about it the more I am convinced that the MERG boards have had a CV value changed inadvertently by some PoM work. I won't be doing that again. 

 

I have got some MERG paperwork relevant to the accessory decoders so I will have a look at that to see if the relevant CV details can be gleaned. Meanwhile if you know what I need to do and can save me some reading please post details.

 

Chaz

[PaulCheffus]

The more I think about it the more I am convinced that the MERG boards have had a CV value changed inadvertently by some PoM work. I won't be doing that again. 

 

I have got some MERG paperwork relevant to the accessory decoders so I will have a look at that to see if the relevant CV details can be gleaned. Meanwhile if you know what I need to do and can save me some reading please post details.

 

Chaz

Hi

 

Depends what has changed but I would try setting

 

CV3 = 0 Output 1

CV4 = 0 Output 2

CV5 = 0 Output 3

CV6 = 0 Output 4

 

Which should set them as continuous output.

 

Also check CV2 = 255 all outputs on.

 

The above assumes you are using the steady state decoders kit 53.

 

Cheers

 

Paul

Edited September 19, 2017 by PaulCheffus

[chaz]

Hi

 

Depends what has changed but I would try setting

 

CV3 = 0 Output 1

CV4 = 0 Output 2

CV5 = 0 Output 3

CV6 = 0 Output 4

 

Which should set them as continuous output.

 

Also check CV2 = 255 all outputs on.

 

The above assumes you are using the steady state decoders kit 53.

 

Cheers

 

Paul

 

 

Yes Paul, they are steady state decoders and have been used with Tortoises for some years. Thanks for the info', most helpful. I will be hooking up the decoders to my laptop this evening and will report back.

 

Chaz

[chaz]

OK - I have sorted it out.

 

Hooking up the laptop and ESU programmer interface to the DCC input of each MERG accessory decoder in turn I read the values in CV3, 4, 5 & 6.

 

The values read 40, 20, 40, 0.

 

When I used PoM to set the CVs in a new loco decoder guess what values I put into CV3, 4 and 5? That's right - 40, 20, 40. I didn't change CV6 - which explains why one of my points was unaffected and worked as it should - it is driven by output pair F4.

 

So you can't use PoM programming on your locos if your system is like mine - a Roco Multimaus DCC with MERG accessory decoders - as you risk resetting all the MERG boards to unwanted CV values.

 

Two solutions are possible.

 

If you must use PoM programming then put a switch into the DCC bus connection to the accessory boards so that they can be disconnected while you do it. 

 

Or - and this is my preferred method...

 

Use software in a computer via a programming interface to a dedicated programming track - which is completely isolated from the layout.

 

The only remaining question is why the MERG boards reacted to a loco with address 18 having CVs written to when none of them have that address.

 

Thanks again Paul for your info' which helped me on the way to the solution.

 

Chaz

[Nigelcliffe]

Are you absolutely certain you were in "programming on main" within the MultiMaus.   Because it sounds like you went into "service mode programming" in order to set CV's to all devices on the layout, or somehow got the MultiMaus to issue a "broadcast to all devices" POM instruction. 

 

To the best of my knowledge, the MERG devices should ignore POM instructions.

 

 

- Nigel

[chaz]

Are you absolutely certain you were in "programming on main" within the MultiMaus.   Because it sounds like you went into "service mode programming" in order to set CV's to all devices on the layout, or somehow got the MultiMaus to issue a "broadcast to all devices" POM instruction. 

 

To the best of my knowledge, the MERG devices should ignore POM instructions.

 

 

- Nigel

 

 

No, Nigel I'm not. I might well have used service mode. I must admit I wasn't aware of the distinction between these two methods. I'm not sure that the MM makes any such distinction when you go into programming mode. I will in future always use a dedicated programming track so no such confusion can result. Using ESU's programmer and software to read and write CVs via a dedicated track will avoid any such problems.

 

Chaz

[Phil S]

Firstly, the Roco/Fleischmann Basic Multimaus CAN be set to either 'service mode' ( labelled as CV in menu ) or On-Main programming (labelled PoM in menu)  An error will be reported if PoM is used incorrectly.

However, when used with the BASIC AMPLIFIER 10764 there is no separate programming track - so either add a switch or a pluggable connector to allow programming on s separate piece of track and not the whole layout INCLUDING the accessory bus (as you have discovered !) ... unless you also add an isolating switch for the accessory bus.

 

Accessory Modules/Decoders which use a 'LEARN' button as their ONLY method of being programmed will be unaffected if they are left in circuit - hopefully the sum of all their currents will not affect any programming current sensing.

'Read Back' prior to writing IS possible with the basic multimaus BUT ONLY if connected (as a slave) into a MultiCentralePro or Z21 .... because THEY hold the master software internally, whereas the Amplifier is just that ... the handset plugged into the Master Socket IS the Master - using the 3rd pair of the six wires ... hence it is also the memory of locos, speeds, functions and point  positions.

[chaz]

Firstly, the Roco/Fleischmann Basic Multimaus CAN be set to either 'service mode' ( labelled as CV in menu ) or On-Main programming (labelled PoM in menu)  An error will be reported if PoM is used incorrectly.

However, when used with the BASIC AMPLIFIER 10764 there is no separate programming track - so either add a switch or a pluggable connector to allow programming on s separate piece of track and not the whole layout INCLUDING the accessory bus (as you have discovered !) ... unless you also add an isolating switch for the accessory bus.

 

Accessory Modules/Decoders which use a 'LEARN' button as their ONLY method of being programmed will be unaffected if they are left in circuit - hopefully the sum of all their currents will not affect any programming current sensing.

'Read Back' prior to writing IS possible with the basic multimaus BUT ONLY if connected (as a slave) into a MultiCentralePro or Z21 .... because THEY hold the master software internally, whereas the Amplifier is just that ... the handset plugged into the Master Socket IS the Master - using the 3rd pair of the six wires ... hence it is also the memory of locos, speeds, functions and point  positions.

 

 

Phil, thanks for that useful information. Taking your first paragraph I had gone to the menu item "service mode" by selecting CV - thereby inadvertently causing the problem. If I wanted to use the Roco MM basic system for programming I would certainly arrange the wiring so that the track output would go either to the layout or a dedicated, electrically isolated programming track.

 

However, as I said above, I would prefer to use my PC laptop and the ESU programming interface and software. This allows CVs to be both read and written. The laptop (running 7) is used for nothing else now as I now use a Mac for all my computing so it can be left under the layout and brought out when needed. I will probably disconnect it from the net to avoid the endless plague of MS updates.

 

Chaz

[Phil S]

By contrast 8-) ... On our H0 transportable layout, which is run with MultiCentralPro, I have a separate loop which is powered by a basic Multimaus and Amplifier - so that if one system fails, there is still a train running  (inside the helix which connects the 3 levels **) - but also so that programming / reprogramming  does not stop trains on the main layout. Also, it is much faster to simply write a new value.    

IF I did want to log and modify the complete set of CVs, then I would connect my sprog and update my JMRI software..... but as I use illustrated stocklists to identify locos - which mostly use the loco numbers or classes - I havn;t had a reason to use this for years.

** younger visitors can also safely run this train without affecting the remainder of the layout - whilst with G Scale we let them run a sturdier train

 

One reason I have only recently started using our Z21 ( and on G Scale) is that it will take a long time to enter all our locos and photos ! ... although I have used the 64 loco library in each Multimaus, we often end up reverting to loco number entry through habit 8-(  [ It was after I realised that I started to use them at t ALL locos - ie with long or short addresses could be stored in the handset's library at all! - the manuals still refer only to 'CV1' even with the LAN maus  8-)