Welcome to the forum   
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

Share
Options
View
Go to last post in this topic Go to first unread post in this topic
Offline marklinist5999  
#1 Posted : 12 December 2024 13:20:39(UTC)
marklinist5999

United States   
Joined: 10/02/2021(UTC)
Posts: 3,864
Location: Michigan, Troy
I chimed in too late yesterday on the webinar with Curtis Jueng and Rick Sinclair to ask questions.
Curtis was explaining how bytes are programmed in decoders for the Motorola and mfx format as well as dcc. He mentioned that check bits are sort of a double repeat code so that devices or trains having the same address don't crosstalk data that is only meant for one. He said it's more complex with MFX. I have a fully MFX signal which I installed as MFX, not mm, or DCC. It's a home and distant combination signal. The CS 3 chose address 03 for the main home signal. I don't have any other signal connected in sequence with it.
Moving on, I have a Marklin br 101 Starlight Express digital fx from the late 90s with dip switches. I set it's address to 03. It only has long distance headlights and acc/deceleration delay on/off functions. When I ran it, and selected one or the other, the signal setting and lights changed. This seems odd to me.
I know it's a complicated thing, and I may email them about it. Has anyone else experienced a similar thing?
Offline rhfil  
#2 Posted : 12 December 2024 17:10:26(UTC)
rhfil

United States   
Joined: 05/09/2014(UTC)
Posts: 674
Location: NEW HAMPSHIRE, Somersworth
I don't have any signals but the obvious and easy solution to me would be to change one from address 3. That address is one most Marklin non-mfx locos were set on at the factory so I would change that one.
thanks 1 user liked this useful post by rhfil
Offline JohnjeanB  
#3 Posted : 12 December 2024 19:32:20(UTC)
JohnjeanB

France   
Joined: 04/02/2011(UTC)
Posts: 3,548
Location: Paris, France
Hi
Originally Posted by: marklinist5999 Go to Quoted Post
I chimed in too late yesterday on the webinar with Curtis Jueng and Rick Sinclair to ask questions.
Curtis was explaining how bytes are programmed in decoders for the Motorola and mfx format as well as dcc. He mentioned that check bits are sort of a double repeat code so that devices or trains having the same address don't crosstalk data that is only meant for one. He said it's more complex with MFX. I have a fully MFX signal which I installed as MFX, not mm, or DCC. It's a home and distant combination signal. The CS 3 chose address 03 for the main home signal. I don't have any other signal connected in sequence with it.
Moving on, I have a Marklin br 101 Starlight Express digital fx from the late 90s with dip switches. I set it's address to 03. It only has long distance headlights and acc/deceleration delay on/off functions. When I ran it, and selected one or the other, the signal setting and lights changed. This seems odd to me.
I know it's a complicated thing, and I may email them about it. Has anyone else experienced a similar thing?


DCC or MM2 being one direction signal (one sender, others are receivers) there are NO cross-talk issues. So there is NO problem in using the same address in DCC or MM2 on 2 locos provided the digital signal is strong enough and the electrical distribution correct.

MFX is a different animal so it seems with Märklin's MFX implementation (I cannot say for other manufacturers trying to emulate MFX (M4, etc):
- when posing an MFX loco on the track, the master CS or MS ("Master") checks if the loco is already registered BOTH sides (it may be registered on your CS3 but, been used on a different unit (CS, MS, etc).
- if a loco needs a new registration, the master tries to find an unused address and preferably with the same MFX address as used to drive the loco in question (with another CS / MS).
- when some locos are not present on the layout or are power OFF (i.e.: stopped by a signal's insulated section), it may happen that an address previously used on the master is re-allocated to the new-coming loco (thus "stealing" an already-used address)
- in this case, all is fine until the "other" loco with the same address is powered ON / installed on the layout. As soon as the Master detects that 2 locos have the same MFX address, it decides to re-register one of the 2 locos so that ALL MFX numbers are unique (no address duplication). The re-registered loco is disconnected (placed on stand-by) until the user "welcomes" the new loco (the solid red oval marker is replaced with an hollow red oval, the loco is proposed at the start of the locos list. If you select the loco in editing mode, access its parameters, then the loco is "Welcomed" and takes back its place in the list (name or address listing) and the red oval marker becomes solid again.

MFX accessories: the numbering is independent of the loco numbering

This is only what I have observed using many MFX locos with multiple layouts and CS3, CS2, M2.

Cheers
Jean
thanks 1 user liked this useful post by JohnjeanB
Offline revmox  
#4 Posted : 12 December 2024 21:55:54(UTC)
revmox

Australia   
Joined: 26/05/2021(UTC)
Posts: 198
Location: Australia, East Maitland, NSW
Hi Jean,


I can't offer any solution to your problem with address 03 other than to say that for some reason it seems more prone to issues - possibly because it is a default address in some protocols. I got a Brawa loco preset to DCC address 03 and found every time I moved it all my supposedly Marklin mfx locos would move as well (previous post, sorted by moving it off address 03, mfx locs apparently still listening out on DCC 03).

My opinion about lost and stolen addresses, backed by what I have seen with my mfx recorder/analyzer -

As far as the error checking for mfx transmissions goes - it is a 16 bit checksum (CRC) giving only about 65000 possible checksum combinations. However, a typical mfx command may be around 48 bits - several hundred quadrillion combinations - so many, many false combinations (corruptions) could possibly go undetected. That said, with careful choices about how the CRC is calculated, it can do a pretty good job of detecting the occasional bit of burst noise expected on a model railway. But it is far from what would be called robust in the communications world.

The same can generally be said about the RDS feedback signal. - pretty good but not foolproof.

IF the mfx signal to and from devices was 100% reliable then the controller SHOULDN'T forget registered locs or reallocate existing addresses. - but it seems clear that it occasionally does.

This is definitely not a hate session against mfx - I think it probably does a really good job of handling half a dozen loc on a modest layout. Maybe when you go to more than a dozen locs on a large layout, lots of accessories, and mixed protocols the "shortcomings" become more apparent.
thanks 1 user liked this useful post by revmox
Offline Kiko  
#5 Posted : 13 December 2024 04:53:42(UTC)
Kiko

Canada   
Joined: 13/02/2017(UTC)
Posts: 159
Location: Ottawa, ON
Hello,

Sorry if this has been covered elsewhere, but please keep in mind that the address space for "locomotives" is completely separate from the address space for "accessories". These two address spaces do not compete but reside in their own environments in the control system.

Cheers,
Andry
Andry
// Marklin HO K track; ECoS; TrainController Gold; Marklin & ESU decoders; Arduino controlled Switches, Semaphores & Accessories (DCC); Win 11 //
thanks 1 user liked this useful post by Kiko
Offline marklinist5999  
#6 Posted : 13 December 2024 13:07:19(UTC)
marklinist5999

United States   
Joined: 10/02/2021(UTC)
Posts: 3,864
Location: Michigan, Troy
Thanks guys! I did change the address of the loco to omit the issue. It's not MFX. Simpler than removing the signal from the CS 3 and reloading it. Maybe my house is posessed by rf gremlins. Back in the summer, the garage door began opening and closing at will. It's a 375 MHz belt drive from 2007. Locking out the wall switch button stopped it. I reset the limits and it hasn't done it during the day. I lock it out at night and if we go away, I unplug it. Logic boards are no longer available as the new standard is 950 MHz.
A logic board costs half the price of a new opener anyway. I was told that locking out the wall switch doesn't stop ghost opening and only disables the remotes. I could still operate it with the exterior keypad though, which is also a remote.
Offline clapcott  
#7 Posted : 16 December 2024 02:48:38(UTC)
clapcott

New Zealand   
Joined: 12/12/2005(UTC)
Posts: 2,448
Location: Wellington, New_Zealand
This has been covered before and is a known bug.

Yes the Loco and Accessories, using MM SHOULD be a different address space (frequency band also)
But the signals do not honor this.

I would suggest setting the signals to use DCC

Peter
thanks 1 user liked this useful post by clapcott
Users browsing this topic
Guest
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

| Powered by YAF.NET | YAF.NET © 2003-2025, Yet Another Forum.NET
This page was generated in 0.654 seconds.