Busses and feedback

QUOTE 1. Is it true that all systems must be passing the same protocol OUT to the tracks to enable us of common decoders for loco's and common decoders for points etc.?

Yes

QUOTE 2. Is it true that the RETURN bus for detectors etc. is proprietory and so all circuits will not be common, because of different protocols?

Yes

QUOTE 3. If I built a detector of some sort, how do I encode the address into it so that the system knows which detector it is?

That's the tricky bit. Essentially you have two different systems. The standardised DCC control system is "fire and forget" and apart from RailCom which is developing at glacial speed (that's speed of flow down the mountain, not the rate of summer induced melting) there is no standard way of getting setting information back to the control unit.

My background is similar to yours, and I suspect that I have been down the same path. I initially short listed the Lenz because it seemed to be the most complete system with regard to feedback. A freak of circumstances led me to postpone buying such a system - out of stock - and during the ensuing months I became disillusioned with the lack of new product development on the Lenz system. I eventually bought an ECoS, not because I was taken by the S88 feedback bus which as far as I can tell is proprietary and difficult to get information on, but because I decided that for the volume of feedback detectors I needed, I would probably need to build them myself so the feedback bus became irrelevant. I liked the basic hardware and "took a punt" on some interesting firmware appearing in the distant future given that it had an addressable pixel based display.

For a feedback bus I did not want to reinvent the wheel so I joined MERG and I have been particularly interested in the CBUS project. If you have not come across the MERG website, I suggest you have a look. MERG website.

As you may have noticed from other threads / blogs on the Forum, the ECoS has good support for shuttle operation. The operating firmware has gone beyond the "introductory" phase and is now starting to add some really interesting features but it is significantly more expensive than the NCE.

I hope this helps

David

Or the other way to go is by using a standalone S88 feedback bus.

This means you could use the NCE Powerhouse Pro as the DCC controller, connected to your PC.

For the separate S88 feedback bus you could use the LDT offerings as shown HERE

They have S88 occupancy/block detectors that can be connected to the S88 port of the ECoS or even the Viessmann Commander(how much money did you want to spend!!) or you can plug them into the HSI-88-USB PC interface which means you could then use any DCC controller which can be connected to a PC for automation.

I've just bought the RM-GB-8 which is a feedback module with block occupancy detection built in.
These come in kit form and just require the parts soldering onto a PCB about 30mins work, but saves £12 on the price!!
I've not actually fitted them yet but by all acounts they seem to have good reviews from others who have used them.

Lots of ways to go really, just depends on what type of DCC controller you really want.

Cheers

QUOTE (Donone @ 2 Oct 2008, 02:16) <{POST_SNAPBACK}>I am a newbie and need help. No I am not (probably) going to ask which DCC system to get, even though that would be helpful.
I have read just about every manual available trying to decide. I rejected some due to the different bus systems.
I am an electronic engineer and happy to build the 'extras' by which I mean, occupancy detectors, accessory decoders etc.
So, I rejected Digitrax because it appeared to me that one has to buy the proper devices for their bus. I similarly rejected several on the same basis, because all the circuits I could find seem to be applicable to the Lenz system, so I thought that I could only get a Lenz compatible system (even though I don't like their unreadable display).

There is the problem. NCE Power Pro 5 does not seem to have a bus to return the signals, but has other wonderful attributes.
Others have different busses to return signals, but no circuits.

I need someone to clear up what may be misunderstandings on my part.

1. Is it true that all systems must be passing the same protocol OUT to the tracks to enable us of common decoders for loco's and common decoders for points etc.?

2. Is it true that the RETURN bus for detectors etc. is proprietory and so all circuits will not be common, because of different protocols?

3. If I built a detector of some sort, how do I encode the address into it so that the system knows which detector it is?

4. If NCE does not have a return bus, does this mean that it cannot be used for shuttle.

My aim is to add a computer and Rocrail or the like and presume the DCC I choose must have a return bus to enable the DCC to then send info about the occupancy to my computer.

Grateful for any assistance in straightening me out. I think I read too many manuals.

Oh! By the way, which should I choose, Lenz 100, Digitrax Supe Chief, NCE Power Pro 5?
Thanks
[EDIT] Should I have said that my aim is auto control of some trains and manual others.

*** All systems have two busses: One power bus and one control bus. There are no exceptions.

From reading your post you are missing or misinterpreting much of what you are reading.

Each brand chooses their own bus format, but all are really simply a derivative of standard computer type communications busses, and all are nothing overly special in data communication or computer speak. an ability to survive the vagaries of miswiring and worn cable choices by modellers is more important to them than speed or sophistication in any way.

ECOS uses a derivative of The CAN bus, as does Zimo
Lenz uses a standard computer type bus they simply choose to call their own name
Marklin used S-88 bus originally, and now use a CAN type derivative

NCE uses what is primarily an RS485 bus - the nearest I can explain it is a high power printer bus is similar. It actually has a big advantage in that it is a very robust bus for complex layouts.

ALL of them are equally capable of being computer linked and operated, and all are quite capable of handling 2 way communication.... NCE even has the computer port on the unit so I have no idea where you got the idea their bus was one way only....

Detectors are always simply dumb devices and all brands can accept input from them and react accordingly - its just an on/off or positive or negative voltage swing that requires an action ascribed to it.

if you want to add intelligence to detection as a home brewed option then you should learn about programming of processors (PIC etc). All of the communication protocols are freely available for those who wish to do their own thing.

Detectors for all Brands depend on an interface device between detector and action device (control system). It is this IO device that decides how the information will be interpreted and acted on, not the DCC system or the detector.

Things like railcom are simply proprietary additions to decoders - exactly the same result can be obtained for other purposes by other technologies - for example, I use RFID type devices as train describers on my loco's so the system can know specifically what is where on the layout.

NCE are just as capable of being used for shuttle etc as any other brand is - they are equally well supported by most computer software and as to "own brand" hardware, they build it for use with own and other systems too... in fact very soon a simple to use add on NCE outboard device will add so much functionality to NCE that it may be in the lead in this area!

Make your decision on DCC system SOLELY based on the ability to drive trains comfortably and well, not the layout control hype that is used. Much of that is smoke and mirrors, as DCC is for driving trains, third party or outboard systems are for doing the layout control! If you lock into a brand for the wrong reasons, then you will almost always be disappointed.

Given your post, take it slowly: Rushing in with half awareness is not a good idea for either your wallet or your eventual hobby satisfaction!

Regards

Richard.

I have thought on it and believe the way to go is ECoS on its own, then connect PC as needed. A question has arisen in my mind about S88. Units can be built from simple shift registers, parallel load serial shift. The manual seems to suggest that contacts on the line can be connected directly to the inputs of the S88 on the ECoS, not mentioning any units between.
It says simply put two contacts etc.
Can anyone please enhance on this issue before I buy?
Thanks

[Edit]
@Richard
Thanks for that explanation Richard, your post did not appear on my screen and so missed it.
I understand what you say, I believe, but am still slightly bothered by for example running shuttle with say NCE but without a PC (which will be added later). For the time being ECoS seems able to do this alone while I cannot see how the track detectors are fed back and accounted for with standalone NCE.
I would appreciate just a little more assistance here and will of course not buy before being sure. I keep bouncing back and forth.
I was considering Power Pro 5 amp.
Thanks

QUOTE (wiggy25 @ 2 Oct 2008, 03:46) <{POST_SNAPBACK}>Or the other way to go is by using a standalone S88 feedback bus.

This means you could use the NCE Powerhouse Pro as the DCC controller, connected to your PC.

For the separate S88 feedback bus you could use the LDT offerings as shown HERE

They have S88 occupancy/block detectors that can be connected to the S88 port of the ECoS or even the Viessmann Commander(how much money did you want to spend!!) or you can plug them into the HSI-88-USB PC interface which means you could then use any DCC controller which can be connected to a PC for automation.

I've just bought the RM-GB-8 which is a feedback module with block occupancy detection built in.
These come in kit form and just require the parts soldering onto a PCB about 30mins work, but saves £12 on the price!!
I've not actually fitted them yet but by all acounts they seem to have good reviews from others who have used them.

Lots of ways to go really, just depends on what type of DCC controller you really want.

Cheers

***I'm really wondering why there is such an obsession with S88 bus.

(Ok, I know its because its been popularised by Marklin and now ECOS - but ECOS uses it only to give themselves access to marklin owners, as does Littfinski really - NOT because its good - its nowhere near as good as most "bus" structures in fact!)

I use it too, but only for a small application and only because the littfinski RFID device is specifically built for it!

Actually its a very very old/early computer type approach to data which is now almost totally dropped due to its narrow usability.

The S88 bus was actually established originally in the dark dim early days as cheap feedback bus for model railway layouts.

The principle is simple: the S88 bus is just a serial shift register with a parallel load input.

Additional modules are added to this bus by simple cascading, creating a long shift register with all bits in one long chain. The advantage of this simple approach comes with a big drawback or two...

the address of each module is defined by its position in the chain and is unchangable. The transmission of data is made unprotected: there is no parity, no checksum nor CRC.

Simple S88 use is generally fine/OK, but complex systems are prone to confusion and communication problems unless lines are properly terminated, CAT5 or better shielded wire is used etc etc.... and because of the simple chain type comms, as the system grows it slows and gets unreliable....

Also - don't run S88 alongside the power bus unless the power bus is twisted about 4 turns per foot - crosstalk will potentially create mayhem when the power bus gets loaded with a couple of amps or more!

Please don't misunderstand - it works, but let not create a myth that S88 is something special - its not!

Richard

QUOTE (Richard Johnson @ 2 Oct 2008, 08:18) <{POST_SNAPBACK}>Please don't misunderstand - it works, but let not create a myth that S88 is something special - its not!
Absolutely!
When I first looked into how it worked, I was surprised to find how simple it was, and how similar it is to the RPC concept which I had devised for myself some years earlier, completely independently. The same type of shift register is used, except RPC incorporates an output channel too for controlling accessories, whereas S88 is input only. This fact also makes it relatively easy to attach MERG RPC kits to an S88 capable controller.
This has been successfully tried with an ECOS and a MERG DTC 8 channel current detector kit (so far). The only problem we found was that the ECOS was expecting a logic '1' for occupied, whereas MERG kits have always adopted logic '0' as the active state. However, it is a trivial matter to turn the logic up the other way.
One of the issues that Richard raised regarding the termination and wiring of such a simple system, particularly regarding wire lengths, is also very relevant. To get round this in the RPC system a special interface unit was created, known as RSE, allowing any logic level shift register based modules to send their information along a standard CAT5 computer network cable with proper buffering etc. Several of these can be daisy-chained together if desired.
Those familiar with the Beckenham Club layout 'Horton' may know that this principle is used extensively throughout their control system, even though they drive the trains using DCC in the usual way.

QUOTE (Richard Johnson @ 2 Oct 2008, 08:18) <{POST_SNAPBACK}>***I'm really wondering why there is such an obsession with S88 bus.

(Ok, I know its because its been popularised by Marklin and now ECOS - but ECOS uses it only to give themselves access to marklin owners, as does Littfinski really - NOT because its good - its nowhere near as good as most "bus" structures in fact!)

I use it too, but only for a small application and only because the littfinski RFID device is specifically built for it!

Actually its a very very old/early computer type approach to data which is now almost totally dropped due to its narrow usability.

The S88 bus was actually established originally in the dark dim early days as cheap feedback bus for model railway layouts.

The principle is simple: the S88 bus is just a serial shift register with a parallel load input.

Additional modules are added to this bus by simple cascading, creating a long shift register with all bits in one long chain. The advantage of this simple approach comes with a big drawback or two...

the address of each module is defined by its position in the chain and is unchangable. The transmission of data is made unprotected: there is no parity, no checksum nor CRC.

Simple S88 use is generally fine/OK, but complex systems are prone to confusion and communication problems unless lines are properly terminated, CAT5 or better shielded wire is used etc etc.... and because of the simple chain type comms, as the system grows it slows and gets unreliable....

Also - don't run S88 alongside the power bus unless the power bus is twisted about 4 turns per foot - crosstalk will potentially create mayhem when the power bus gets loaded with a couple of amps or more!

Please don't misunderstand - it works, but let not create a myth that S88 is something special - its not!

Richard



I didn't know any of that!!
I've only used the LDT items because they were cheap and readily available.

I hope I didn't make it out to be something special, I'm not that clued up on electronics or DCC; it was just a way I knew, of how to provide feedback into the ECoS although there are other ways.

I shall now go and boil, not my head but.....the kettle for a nice cup of tea!

QUOTE (Donone @ 2 Oct 2008, 20:28) <{POST_SNAPBACK}>Thank you for that link Vulcanbomber, very interesting as is the Cat5 tip Gordon H.
Well I just did it after two weeks of daily reading and re-reading and taking note of comments on various forums; ordered the ECoS.
I believe it to be the future way things are to be done instead of sequences of button presses and mini multi-purpose screens on handsets, which after all are just a compromise to utilise the cheapest solution.

I am glad I joined this forum and will hope to get much more from it in the future.
Thanks

Not really - The primary differences are physical not capability.

ie: Train driving systems vs layout driving systems - nothing to do with the cost of the display at all. Actually the majority outside EU and UK have larger layouts so console type controllers are much, much less favourably viewed, vs smaller layouts that can use an ECOS type unit as a whole system controller from one place.

Most users of NCE type gear are focussed on the best possible train control interface and other issues are consciously separated, so they will either use separated panels or PC attachment for layout control, with extreme layouts using a full centralised traffic control computer with trains allocated to each operator.

Its all about how operating the layout is viewed, and nothing to do with technology really. You need to ignore the pretty boxes and specifications totally for a while, and focus on other much more important things, like how you like to relate to the trains, before you even think about a brand.

The biggest differentiator is the way the system is designed to be used followed by the way they interact with you / or the displays are used - NCE communicates so much more than Lenz for example and does it in plain easily understood English using its handset screen, even though the screen is the same as Lenz in size... ECOS uses a larger very nice touch screen and is effectively a linux/windows type operating system with progressive menus but its not portable.

Both are very good systems, its all about personal preferences and layout size/application.

re Andrews comment - there is some ASCii related programming info directly at the back of the NCE manual.... and the NCE main technical manuals with much of what you need to know available as PDF files in the NCE-DCC Yahoo group files.... but the while the data is correct as to system architecture and other issues it isn't necessarily up to date in relation to current system software loads.

regards

Richard

QUOTE (Donone @ 2 Oct 2008, 23:50) <{POST_SNAPBACK}>I know this is now a conversation with Richard, but my view for what it is worth, is that if it works OK then it must be OK, and any argument about which is best is largely irrelevant.
What does matter here is surely passing on as many different opinions and experiences as possible for the greater good.
There is theory and there is practice!

***Agreed totally - there are thousands of S-88 modules out there, and they do work... very well in some applications - but total potential data flow on layouts is a thousand times more now than it was when S88 was adopted, and its very important that its real world limitations are understood if one is to adopt it for larger or more complex applications, or tears will eventually follow.

Its also important to know that there are wider options - Gordons comment is particularly valid, and actually you may enjoy MERG membership as its a very clever group of people and a blessing to anyone comfortable with a soldering iron and components!!

but why is it a conversaton with me?

I am simply answering questions like all others, and where I see it appropriate, making corrections to assumptions that aren't really right. I always choose to do such things directly and to the point... there's no points for dancing around issues when other peoples wallets and hobby pleasure are involved.

ALL inputs are equally valid as to whats possible and suitable - communication styles, perspectives and experiences simply differ.

Richard

Hi All

This has been an interesting read. There is so much you can learn about DCC the mind boggles!

But the good thing is , that you can keep it simple if you want or take it to complex levels.

When on Holiday in the Mid 80's , I saw my first Digitally controlled layout on a display . this was a large Layout (it may have been a Marklin system) and it rekindled my interest in Model Railways (mind you it took me another 15 years to dig out my old Trains) The cost of going DCC at the time prevented me from considering that option , so I set up a conventional DC layout .

When Roco (with Lenz) released the LokMAUSII Digital Starter packs , it suddenly became affordable I bought a second hand set Complete with Loco track and Wagons for $120 including Postage . a couple of DCC loco's at around $90 came next. Then a few conversions of my old Loco's and there you are up and running DCC at affordable prices.

I sold the LokMAUS system and went for the MultiMAUS which I have had for 3 years , I find it easy to hold and use one handed . I can use it on my Grandsons Lenz system (which is Computer controlled as well ) and it does what I want it to do.

The Future prospect of integrating Computer with DCC control .Most systems have either an optional computer interface or have one built in. There are varying degrees of compatibility(and controllability ) between the Software available and the DCC systems

JMRI is a Java based system available for free (you can donate $ to help )

You will find info on compatibility of systems and the program Here

The most expensive but probably the most comprehensive software Is Railroad and Co
You will find info on compatibility of systems and the program Here
Their Programs will Integrate Sounds with the positions of running Loco's utilizing a 5.1 computer based sound system (4 speakers around the Layout and a Sub)

The System you choose will also depend on your personal choice, what looks and feels best for you
Some people prefer a small 1 hand operable control , Knob based or button based
Others a 2 hand unit like the ECoS
Best to try a few systems out to see what feels right for you and compare that with the compatibility of the software

There is also a Forum member working on Voice Command via computer software
this looks to be a great development you will find details Here

Hope this helps

Regards Zmil