I have started to automate my M track layout with CS3+ and focused on use of contact tracks instead of (finnicky) toggle tracks. The latter sometimes get stuck and wagons with lighting sliders complicate the trigger. I have not tried reed switches on my layout.
I discovered after many crashes that the standard m track contact rail (5105 and 5104) has an isolated length that is less than the wheel distance of my modern trains (like ICE). Sometimes it senses occupancy, sometimes not. I can use a contact extension track and two 1/2 railroad crossing trigger tracks, but that is a lot of specialty track for one sensing segment (and hard to find, and expensive).
Is there a clever way to make one's own contact segments with M rail? I experimented with cutting rail, but it remains grounded to the metal base regardless.
Some cutting and soldering is fine. Thoughts, Marklin peeps?

One crazy approach is to use two M-to-C tracks and interrupt the rail on one side. Downside is giving in to plastic. Upside is no shorts from metal base. Pricewise about the same $15-20 solution.
(yup, answering my own query as my 'research' continues)

Went straight to the shop to try this.
Wow, shrink tube for isolation works great; see pics. I can make the sensing track any length with this method. I just added marklin stock interrupt for testing, but will modify 2 straights to make my own.
Great solution, Martti!

The alternative to cutting 5200 series track is to use current sensing s88 units link the Viessmann 5233 (see https://viessmann-modell.com/en/product-range/electronics/digital-electronics/117/feedback-decoder-with-track-occupancy-detector?c=242 ). Instead of using one of the outer rails for occupancy detection, it uses the center rail. I have used this solution for decades.

Other manufacturers include LDT -> https://www.ldt-infocenter.com/dokuwiki/doku.php?id=en:ab_rm_88_n_o_3lg . This one has the later style connectors for CS3.

In both cases, you just have to isolate the middle rail at each end of the section and attach a feeder to the s88 unit from somewhere in the middle of the section. I will look for the marklin number shortly, but it is a single red wire with a middle rail tongue tab.

Hi Adrian


Current sensing is OK for 2 rails operation (you have no choice AND you may have resistive axles - graphite allowing you to detect ALL cars).

In 3 rails, it is a disaster: the system sees ONLY the current consuming cars (With a slider). How many cars / wagons have a slider, not to mention the noise all these make. This means that in 3 rails, detecting cars with an extra rail makes hell a lot of difference I think. Any car is detected unlike with with current sensing.
Just my opinion.
Cheers
Jean

I used current sensing using an LDT s88 unit, on my previous M track layout. It worked very well. Now, Märklin has an s88 unit at a reasonable price:

https://www.maerklin.de/...ts/details/article/60882

Thx for the alternative suggestion of using current instead if mass sensing. However, I believe this requires s88 DC modules, whereas i have l88+s88 ac modules on my M track layout.

Making my own contact tracks has solved the length limit. Connecting two homemade contact tracks does take soldering of a feedback wire and cutting a small space below one connector to avoid metal base contact (just like stock m contact tracks have).

I'll eventually try computer based automation (rocrail?), but for now stick with CS3 software control.