Well,
what is really known about this? Very little has reached my eyes and ears (except pure speculations). So let's speculate :-)

It appears that the first units which will be able to handle the new protocol are the mobile stations, the mfx decoders, and the CS. I get the impression that the mobile station don't react unless a mfx decoder registers itself during power setup; therefore, I guess noone has yet been able to see any new packets, as the mfx decoders are not available?

It has been denied from Märklin that the new protocol is DCC; it's also clear that Märklin for marketing reasons can't tell the whole truth, as well as can't lye completely. It could be possible to have a new protocol very similair to DCC which is not approved by NMRA, and then it isn't DCC by definition. But it seems a little bit unlikely to me, even if I think it would be smart, as in that way it would be easier for Märklin to reuse existing components.

The specs of the new protocol, or rather, what is told about the new system is that there are 16 k adresses, 16 loco functions, 128 speed steps, and a register function activate at power startup; the digital system senses what mfx decoders are present, and make sures even that two identical locos (with mfx) can be adressed individually. Mark my words, it has not been described as features of the protocol. Also the existing protocol can be used for some of these features, for example as the new BR55 with 8 functions, implented by using two adresses!

So. Perhaps the first question is, is the new protocol an enhancement of the Motorola protocol (1), or is it something completely new coexisting with the Motorola protocol (2)?
1. could be thought of as Motorola New vs Old. I don't think so, it seems very difficult at least to fit 16 k adresses into it.
2. seems more likely, either a third "frequency" opened up and Motorola Style, or something more like DCC (which with Uhlenbrock and lots of decoders coexist), or something completely new. The nature of this is important; with possible new frequencies it might be that the old boosters can't be able to handel the new protocol. And it might be that IB+Daisy can't be software updated to do it either. Any thoughts from those more skilled than me in electronics?

Regards,
Lars Westerlind

Hello Perz, Lars, everybody,
I was told that Selectrix is a high speed protocol. It is compatible with Motorola and DCC and used in this way in the Intellibox.
Märklin should keep support to SX as well, since they can't leave the TRIX customers abandoned.
So, in my opinion, another possibility is an enhancement of the SX protocol to support 16K addresses, 128 steps and 16 functions.
Only my speculations, of course.
Regards,

Jorge,
there are two reasons why the Selectrix protocol is faster than the other protocols. First, the pulses are shorter. SX: 50uS per pulse, Motorola: 196uS per pulse, DCC: 200/116uS per pulse
Second: Selectrix works with groupaddresses. Other protocols send an address per loco, and data for that loco following the address.
SX sends a groupaddress, followed by the data for seven successive loco's, so saving transmission time for 6 addresses.

Selectrix is not compatible with motorola and DCC as you stated, but the IB sends messages for MM, DCC and SX in sequence. When for instance a Selectrix message is send the DCC and Motorola decoders don't react because they don't understand the "language", and so vice versa.

Lars,
A weird speculation from my part.
Guess the 16k addresses in fact are not decoder related, just aliasing addresses.
What if a short shuts down the system with "dynamic address assigning"? How does the system resume operations in this case? How do decoders know about their identity in a digitally controlled environment with old style power off block sections?
How long does it take to identify and dynamically assign addresses for, let`s say, 80 locos?

To Jorge,
Don`t care about the Selectrix users group. They are excellently taken care of By MÜT, Rautenhaus Digital and Doehler&Haass. Far better than any time in the past!
BTW, MÜT is probably the best address when it comes to digitalizing
Z-scale.

Regards,
Manfred

Fredrik...are you lucky enough to have more than 80 Locos????
David

Yes, I think you are right. But still guess that they won't change the accessory protocol.

About more news, yes, I guess we all would like that. Including what possibilities there will be for other manufacturers, and how they (including Uhlenbrock) reacts to that.

In my thinking it's more important to know if the mfx-locos will be fully usable with for example Intellibox and Daisy, if the new protocol may co-exist with DCC (nearly the same question), and if the DC brake signal is still understood by the mfx. Central station and Mobile are of less interest, as they represent closed systems.

Regards,
Lars

Good point Fredrick. With more and more items now working under digital control you are right about 256 switches.Also Lars point about mfx is interesting although at present there will only be two mfx locos available as far as I can see. Will Marklin be able to change fx and C sine locos to mfx if they wish thus making full use of the new system.
David

Interresting information from yahoo Marklin Bar and Grill

"
From: "Prof. Dr. phil. Dr. h.c. mult. Dipl.-Ing. Andreas Kreuz" <kd4vs@c...>
Date: Tue Aug 31, 2004 10:58 pm
Subject: Reflections on the M. Mobile Station

If one has an opportunity to spend a little quiet
time with one of the new M. Mobile Controllers,
one will learn the following.

1. Inside lies a "CAN bus" driver chip.
Therefore, one may conclude that the bus which
will connect the various "stations" is a "CAN
bus." (CAN = Controller Area Network.) Please
see
<http://www.semiconductors.philips.com/buses/can/>
for further details of this simple, two-wire
differential serial bus system.

2. The heart of the MS is an MC9S12D64 CPU of the Motorola HCS12 family.

3. The MS also employs a "PT2579-S" chip. The
great Oracle "Google" reveals to her devotees
that the PT2579-S is an RDS (Radio Data System)
decoder. RDS is an "SCA" system with a
subcarrier frequency in the range of 67 kHz, used
by FM Broadcast stations to send ancillary data
such as a station's call letters, format, slogan
(e.g., Jazzy 100), etc. It is reasonable to
conclude that M. or its developer has harnessed
the same, mature RDS technology to enable
decoders talk back to the control station.

Beste Grüße,
--

Prof. Dr. phil. Dr. h.c. mult. Dipl.-Ing. Andreas Kreuz
"

As John says this is interesting info. My problem is I havent a clue what all these chips and systems are that the Prof is talking about. I think I have reached the stage where if it works the way I want then I am happy and whether it is Chips French fries or just potatoes doesnt really matter. Terrible thing this getting old!!!!
David

It was already known that it uses a CAN bus, but that superimposed RDS stuff at 67Khz sounds exciting... A 3rd "channel", so to speak... Loks, accessories and now also feedback...

Can anyone explain to me what this ciruit does? Is there an antenna present also; wouldn't it be easy to identify? If not, does it mean that the circuit interprets messages with the same coding as RDS-messages, but not necessarily transmitted in the air? Would radio transmission be a problem due to the differen national regulations over the world?

A side note: in fact, the Motorola protocol initially was built with Motorola circuits coding a signal in the way we know, primarily intended for transmission by IR, for example control of TV sets etc. Thereafter Märklin has design it's own circuis, and extended the old protocol.

/Lars

OK.
Is it clear that this circuit is a RDS de-coder then? So that we speak about the "feedback" signal coded like this? Couldn't it be an en-coder which produces the signal for the new protocol as well? 16 k adresses an the like?

And could it be explained how the feedback is done electrically? Should the decoder transmit eneregy to the track in pauses? I guess not, should be harmful if several decoders AND booster transmit energy at the same time, shouldn't it? Is somthing like the booster transmitting a bearfrequence which is modified by the decoder/decoders, and current measured at the booster? I've not taken the time to read NMRA nor ZIMO either, and am probably not knowledged enough to fully understand.

Regards,
Lars

After having gone high tech with MOTOROLA now MARKLIN should buy into GENERAL MOTORS "OnStar" SatNavSystem. GPS on the layout - what a new dimension [:p]

The PT2579-S is an RDS demodulator (receiver) only chip. According to the data sheet, the subcarrier frequency is 57 kHz, not 67 kHz.

Are there really RDS encoder chips? RDS decoder chips make sense since they go into every RDS radio. But RDS encoders are mainly needed in radio stations and test equipment, a much smaller market not worth developing specialized chips for. I have found a do-it-yourself implementation of an RDS encoder, but it requires two PIC processors and a lot of other components. http://members.lycos.nl/rdsencoder/rds1.html

Well, some way the mfx decoders will have to send data that the RDS decoder chip will understand... At least as I understand it... Or am I totally stupid here? Or have Märklin designed/programmed another special custom chip to achieve this functionality?

Yes, that is told in Märklin's official documentation. There is no longer a separation between loco addresses and accessories, and most probably all of those can be aliased. Now you have to switch zillions of dip-switches to get the accessory decoders to the right button on your keyboard. With the Central Station they have to be programmable (please, please) so you can get them on any button you like... Otherwise, if the CS puts them in a random order at startup - how do we ever learn what button switches what accessory?

Fair right - but HO isn't the microcosmos of our computers, so I think we are better off recognizing those outer limits. Also, locos and layouts are bigger than chips so a direct comparison just doesn't work here. Last not least since one single operator can't control dozens of locos or more at the same time - he needs computer control.
Well, and from that point on he has lost control and is just spectator - mutation finished [xx(]

Fredrik, sounds great - so you're an locomotive engineer by occupation [:p] ?

Best Regards
John

Great - so you are one REAL PLAYER, enjoy it [:p]

I made some quick oscilloscope measurements on the Mobile Station output signal. It sends what looks like ordinary Märklin/Motorola packages, with base frequency 38.4 kHz. (Yes, it is 38.4 kHz here, while the 6021 uses 40 kHz.) These packages are interleaved with some other kind of packages that look more DCC-like. The base frequency for these packages seem to be 40 kHz. I'm no expert on DCC so I cannot immediately say if it is real DCC or something just similar.

Unfortunately the oscilloscope I am using is a simple portable one, with limited recording functions and limited resolution. So I don't have the equipment to analyze the packages in detail. I will probably find out a way to do that sooner or later, but it will not be today at least .

It seems the intermediate packets aren't exactly DCC either. They look very similar to DCC, but there are noticeable differences. DCC should code a "1" with nominally 58 us low + 58 us high and a "0" with nominally 100 us low + 100 us high ("high" and "low" may be swapped).
I see valid DCC "0"s but the "1"s are a little bit too short, only 50 + 50 us (DCC requires a minimum of 52 + 52 us). There are also occational non-DCC symbols (100 + 50 us).

The intermediate packets are rather long, well over 20 ms. A double motorola packet is less than 10 ms.

In short:
In the future all system components will transmit and receive data, thus being part of an electronic network.

The German text is copied from the lead of their systems promotions tour.

Regards,
Manfred

Is anyone aware of patents that ESU or M* may have on the mfx system? One would expect that if the MFX system protocol is covered by a patent it is described in some detail in the patent literature.
OTOH, if it is not covered by a patent if can be reverse engineered by anyone who would want this, such as Uhlenbrock.

Sorry if this has been discussed before...

Cheers, Bernhard

Welcome, Digitalbox...

It is an interesting question you put here... What software will support the new system from Märklin? Will it be "Märklin software" only, or will the independents get enough information to implement support for the new system too?

Then we are in a "wye"... Should we go the "open" way or should we go the "Märklin only" way?

This is up to us to decide, would we like to follow the current market standards or should we jump onto the Märklin bandwaggon?

Until the Central Station is on the market, and if the "competition" as Uhlenbrock with the IB will "answer", we have no idea about how the market will evolve... If Märklin releases the secrets of the new system, we can have a healthy market, but if they decide to make things like the Intellibox obsolete by keeping their secrets, we will have to choose what seems to be the best for us... Then market forces will either force Märklin to comply and release info about the system, or make other manufacturers to step up another notch and offer the same features that Märklin have with their new digial system, even if another "protocol" is used...