Hi Robert,

well I just got my CS2 back from repair (and actually got another unit in return), and this one is (off-course) updated to the last SW-issue (1.6.4(3)). With this unit I can NOT have the "T" connected to ground since that causes the "mess" you describe. Thus I've now disconnected this connection and it works fine again. My new unit is a 60214 and I guess it has the "passthrough" ground connection and that might be why it now works without this connection (NOTE: this is my own assumption!)

Therefore my suggestion is: Put out a few S88-monitoring components in a layout-window of the CS2 and see what happens with one "t"-connection connected to ground (brown) and compare to what happens iwthout this connection. When reaction of S88:s is normal then the current state of this connection is the correct one.

If "T" is to be connected to ground then only one should be connected.

I always connect my S88 outputs to ground ("0"), but a connection also to "B" is possible (or at least was possible with the 602x central).

For the 3:rd question I'd say: Not the S88-units themselves - but the cable is, considering you use the original flat type of cable. I now use twisted pair cables only!

Sorry about the "T"-connection confusion, but it seems somethings changed in the different versions of the CS2 (my previous one was a 60213).


/Fredrik

Thanks to Fredrik and Robert for your help so far. This is what is really great about this board that there are so many people willing to share their experiences.

I am going to stick with earthing one s88 as without that very little works, which is worse than having too much working.

I only have one flat cable in use (from the CS2 to the first s88), otherwise I am using standard 0.5m and 2m extension cables which are twisted but not shielded. The distances between the units are not great (CS2->10cm->s88/1->150cm->s88/2->10cm->s88/3->50cm-> s88/4) but the cables do trail around the various 24V DC relays which are part of the set-up (which I did not mention earlier) so this may be where the noise is getting in and I will try rerouting them.


What do you mean by this, Robert? Why do you want to isolate one rail or do you mean that when you want an isolated section, you always put it in the outer rail?



I think that now may be the moment to explain just how I have got everything fixed up. If signalling is not your thing then don't read on!!!
1) Each signal has one button for every different route that leads from it, so to set a signal to proceed in a particular direction, you press the appropriate route button.
2) Pressing the route button picks a 24v minature relay which puts an input into the first s88 unit and this triggers the CS2 to set the particular route. Each route is set up in the CS2 to set the points to the correct position. Note that this just sets the points, it does not clear the signal.
3) I use Viessmann 4554 point motors which provide a positive electrical output depending on their position and each point motor is connected to a 5552 AC relay. Because the 5552 units only have a limited number of contacts, I use them to drive more 24v DC relays. When the points have moved to their correct position for the route, a feed passes through the various relay contacts into s88 no2 which says to the CS2 "Route set - Points lined".
4) I have another set of memory functions in the CS2 set up to clear the signals and this second s88 input triggers the CS2 to clear the signal to the appropriate indication.
5) As we know, Marklin signals have contacts on them for controlling track sections, but I use one of these contacts to provide a feed to s88 no3 which says to the CS2 "Signal cleared". This feed also passes through the point relays so this 'signal cleared" feed is specific to the route set.
6) This third s88 input triggers the CS2 to clear the appropriate distant signal (Vorsignal) to the correct indication.
7) Signals are turned to stop by trains occupying contact sections, but once again the feeds from these sections are routed via the point relays so that only the correct signals are turned to stop. These "Turn signal to stop" commands are sent to the CS2 via s88 no 4.

Why on earth is he doing all of this, you may ask?

Well, signalling is a particular interest of mine and all this overcomes some shortcomings of the CS2 and the Marklin way of doing things and makes it much more like the real thing. To me this is just as important as having the correct colour of handrails on a locomotive! Particularly -
1) Points are proved to be in the correct position before the signal can clear
2) Signals clear to the correct indication according to the direction in which the train will go
3) Distant signals only change once the signal to which they apply has been proved to have changed
4) Distant signals which apply to more than one stop signal take their indication from the correct stop signal
5) Contact tracks which can operate on several signals only operate on the correct one
6) All of this logic can (and will) be driven by a proper interlocking which prevents conflicting routes being set and thus stops train crashes on the layout.

One day, when all of this starts to take shape on my "proper" layout (as opposed to my test setup), I promise I will post pictures of it all, including the model-form SpDrS route-setting signalling control panels. However, for the moment I am confined to a good, old-fashioned "lash-up"!

If you have got this far, then thanks for reading and keep the helpful tips coming. I am always greatful.

FWIW


Only one is needed and only then if there is no internal connection to the Central Station (6021 dosn't need it CS1 does, CS2 your mileagge may vary)


Brown !
Red will appear to work as the digital signal passes through 0V as it pulses between +22v to -22v. The issue with this is that the s88 sampling "may" poll at different times of the pulse resulting in multiple triggers.


I have found the inputs from sensors to the s88s very tolerant while IIRC somewhere Marklin has indicated that the flat grey cable between s88s should not be more than 2m. Personally I have all my s88s close to the controller with as short a cable as possible, and then run Cat5 cables out to the layout (providing nodes of 8 inputs).

Other points in this thread
- I , too, isolate the "Inner" rail of any curve I use as a contact track. Any small loco that "leans" due to centripetal forces may loose contact with the inner rail and stop if the outer is used as a sensor and does not provide a brown return for power. I can live with bouncing sensor, I can't live with a loco stopping/stuttering.

- The connector to the track from the CS1/CS2 has reversed polarity over time. Any modern loco dosn't care so it appears to work. However if you connect the s88 earth to the rail when it is actually the Red you will get unpredictable results.

- Similarly if the plug on the s88 grey cable is damaged such that one wire is not connecting you may also observe unpredictable results if more than 1 sensor is activating because of an invalid ground reference. usually seen as all inputs showing active.

- The "T" people talk about should more appropriately be called an "upsidedown T" as it is a stunted variant of the "ground" (masse) symbol

- I suggest architecting your layout so that locos do not stop and hold a sensor active unless you know you have interlocking safety logic. For example, a train in a ladderyard which enters and triggers another train as it comes to rest on top of a sensor, may inadvertantly re-trigger the same sensor to pulse as it jolts to move off next time.

Robert,

I will give you my answer to the cat 5 question. Here is the US we have chains of large mega-hardware stores such as Lowes and Home Depot. In the electrical section, they sell bulk cat 5 cable in different lengths, 50 feet, 100 feet, 500 feet, 1000 feet. I usually buy 100 feet for around US$30. The cable comes with no connectors on them for people who are doing installs of the cables through walls or attics. You can then add your own connectors. I had used it when I wired my entire house for Ethernet.

The cable as you probably know, is 4 pairs of twisted cable, giving you eight total wires in the sheath. When I want to run multiple cables from one location to another, I use this cable cut to the length I need. Then just strip back the sheath and strip off the ends of the four pairs and connect them like any other wire. I use terminal strips a lot. For S88s, I just screwed Marklins little connectors on (boy, do I hate them ... LOL!!). I also have some S88s from LDT and then it is just a matter screwing them down on the terminal strips.

Although most cat 5 I find is solid, I have gotten a small roll of stranded cat 5 from a hobby electronics supply store. Haven't used any of it yet.

If you can't find bulk cable, I guess you could buy a ready made cable and cut the ends off ... might be a little more expensive that way.

Good luck!!

Robert

Hello Robert,



After 1 year, 1 month and a couple of days, "proper" layout achieved? I would definitely be interested in pictures of it all.

Aloha,

Elliott