Hi All,

I am thinking about adding some automation to my layout via contact/circuit tracks connected to a feedback module to my Central Station 60215. Can someone explain to me what is the difference between the L88 60883 and the S88 60880 as they seem to do similar things.

regards

Paul

Thanks for the information Tom. I guess the 60883 will be for me then. The only advantage is it bring cheaper!

Regards

Paul

That's how I read Tom's reply!

Topic moved into 'Digital" forum where it's more relevant - definitely MR related

You only need one 60883 to connect to the CS. If you need more feedback modules you can use 60881 and connect them with a network cable. The advantage compared to "old" s88 is that it is less sensitive to interference. LDT have been making this kind of modules for a long time, so this is almost as much a standard connection as the s88 bus. CAN-bus is less common in this context, but that is a more reliable way to provide feedback to the CS than s88.

Another advantage of the new "88" for Marklin in the future will be the opportunity to develop the next CS in a more simple way, without S88 connector & without internal S 88 circuit . (it smells a more standardized item)
Costs of production increase with the number of connectors and internal circuits ....
The current CS2 60215 is quite "rich" in terms of connectors/connection...

With regards to feedback, there really is not much to it. The S88 LINK is just the new way of connecting S88 modules. It is not some sort of new feedback technology. The S88 LINK is simply a new entry point for connecting further S88 modules (be it AC, DC or the old type). The S88 LINK itself delivers the first 16 ports (same as an S88 AC). You can, however, configure those ports to represent 64 input buttons, giving you the option to create a real switchboard with buttons that you can use for switching turnouts, setting routes or signals. The latter is quite useful, IMO.

The CS3 only has support for S88 LINK. CS3Plus comes with a built-in S88 LINK (without the 16 ports).

The original question is in the thread subject: " Feedback Module Difference - S88 v L88"




Cat5 is twisted pair cable - it has 4 pairs (8 wires) of which the S88 system only needs 6.

The CAN bus vs S88 vs S88n I'm gleaning somewhat out of a Stummi-a CS2 S88 Thread.



I cannot comment on that. Maybe they were problematic in only a few cases. There has to be a reason the electrical interconnection was changed from the 6 parallel ribbon-cable scheme (CAN bus? - which is generally reliable, yes) to a twisted pair Cat5 electrical interconnection. It may be that its easier to wire up the the connections between modules as obtaining and making Cat5 cables of arbitrary length is very easy and inexpensive these days.

You certainly benefit from greater interference rejection with Cat5 vs the parallel ribbon approach, especially over longer distances in RF noisy environments. If your layout is not in an RF noisy environment, then you wont see the issues. But the amount of RF energy in our homes has change dramatically over the past several decades with the addition of so many other devices: wifi, bluetooth, zwave, zigbee, cordless phones, and likely more. Maybe its to provide two separate busses, I don't know. Others (you perhaps) can explain the issues related to deploying S88 detection over a larger layout space better than I can.

your terminology is indistinct and imprecise - and inaccuarate.

The Cable is not twisted
it is the pairs that are twisted - with different tpm

Which "S88 system" ?
Yes the 6 pin flat cable only had 6 wires
the RJ45 implementation (the so called S88-N) may only have 6 signal lines of functional distinction but these are defined and utilised across 7 wires of the cable.



I very much doubt there is a single reason. A number of items have changed in the last 30 years
- Cat5 was not ratified when the 6088 was first designed.
- the design capability of S88 ports was only 256 - for 6050/1 (limited to 48 for the 6023)
- the "electrical" spec only offered a 2 meter bus length for the string of S88s
- and was based on a 5V power rail.
- polled v interrupt driven
- OEM manufacturers wanting to be "seen" to be different

The debate about "need" v "want", when it came to determining whether the S88s actually needed to live out on the layout is worth bringing up.

For close to 20 years I have been quite happy with my S88s close to the control desk, AND to use Cat5 cables to wire the individual sensors from the layout to the S88 desk. (8 ports per cable, 2 cabled for a 6088(0) )
I found using a single cable to group 8 ports to a certain area of the layout as extremely effective and tidy/disciplined , especially when it cam to debugging.

With the new version of S88s nothing has changed , the 60881(S88 AC) fits my current regime, the 60883(Link S88 = L88) offering with its "matrix mode" is primarily for control desk wiring rather than layout wiring anyway.

What is NEW, and invites a change of thinking, is the 60882 (S88 DC) whereby it IS more prudent for the module to be in close proximity to the power feed lines to the track.
However, if you have no intention to use the 60882, then I would not go out of my way to place the other S88s "out on the layout"


Yep, here I quite agree, using industry standard cables and connectors is great for their price benefits.

That said the ITC industry has been going through a bit of turmoil, as certification of cables includes the quality of the plug termination. For a number of years data centers wishing to meet certification needs have stopped making up their own "made to length" patch cables in lieu of certified ones.


There are certainly benefits to be had ... whether the S88 clocking and data frequency make the Cat5 the best chocie is debatable - it is just a convenient medium.

- It is relevant to note that the S88 clock and ground use a common pair (Blue - Blue/White)
However I am curious to note that ...
- the high frequency data line is paired with the power line and not a ground
- The configuration where the PS(Load) signal (a very low frequency pulse compared to clock/data) has its own ground as a paired signal really does not appear to be making the most of the twisted pair benefits.

My conclusion is that the tpm of the pairs in a Cat5 cable are not really matched to the frequency being used, and as such the benefits are because of better quality wire in the cable and , more importantly, the electronics in the sender and receiver S88s.

If this was seriously being investigated the physical implementation would be differential pairs and not single ended. Correct termination would also assure benefits, and if any sort of auto termination is currently in place, it is within the device and not the cable.


My experience is that keeping it simple is the best path forward.
If you compromise on the quality of your wiring , or do not understand and adhere to discrimination of the return signal where needed then you end up on a downward spiral.

Personally I would recommend minimizing any sensing that use track "0" return (i.e. circuit tracks), and separating out these to their own isolated S88.

I have had two L88 modules on the club's modular layout. One module for each end of the layout. The only wire we had to draw was the Märklin Bus wire and some 60125'es. It worked perfectly. I got the impression from the CS2 that adding another S88 LINK module meant adding additional S88 buses. You address each S88 LINK module first, then the contact number.

I think the amount of modules is only limited by how much the CS2 and its hardware ressources (ability to check all the S88 contacts quick enough) can handle.

Thank you, Peter.

At least, you have reminded me where the "1024" value was quoted.

Unfortunately, I don't have such 64 S88 devices to test on internal CS2 S88 bus if it's true or not.

Cheers
Fabrice

Interesting. I didnt know Link s88 decoders had the ability to control a switch. Am I understanding this correctly?

Ok and the control software still uses an m83 for example to switch the turnout?