Hello,

although CATrain with some changes of the locomotive's wiring (elimination of the direction switch and installing some diodes) can also be used for Marklin HO, the balanced output of the PWM modules causes some difficulties with the isolation of the rails, very especially with M-tracks.

Has anyone with good electronics experience may be considered whether the balanced output can be converted into an unbalanced one, with a common ground for all PWM outputs. Thus, the mass problem of rails, switches, etc. would be largely solved.

And also the isolation of DC systems would thus be simplified, since only one track must be isolated, except for return loops.

The question is however, how to accomplish such a conversion, while maintaining the functions of the sensing lines of the L-298.

I would appreciate your answers very much.

Hello,

using the PWM module with the Märklin "M" track is effectively not advisable.
Isolating every block from its neighbors is not easy.
It might be possible to modify the PWM modules to have a common ground, but I have never tried this.
The L298 has "Full-Bridge" drivers but it is always possible to use each driver in an "Half-Bridge" configuration. Only one output of each driver would be used.
In this case a dual supply is needed (e.g.: +16V, -16V), for powering the PWM modules and the train layout.
The processor must have its own supply (0, +5V), with the logical ground (0V) connected to -16V.
This way, the logic should still operate normally. The current sensing will only be active on the negative supply.
The positive supply will have no more current sensing capability.
This should not be a problem for detecting a locomotive but a current limiter is needed on the +16V to avoid destroying the emitter-follower in case of short-cut on the track.
(The L298 is very reliable, but this "half-bridge" configuration is not the standard way to use it!).
All the 2k2 resistors for the EMF measurement must be removed because the mean track voltage related to the signal ground is no more 0V but +16V!
Anyway, measuring the EMF is useless, Märklin AC motors do not produce EMF...

This is only a suggestion, this needs to be experimented...!

Best regards,

Joel

Thanks so much, Joel, for your detailed explanation.
I will see, what is easier, use isolators in both tracks and mantain an isolation gap between the metalic base, or modify the PWM modules.
By the way, considering the higher energy demand of Märklin loco motors (field magnet coil instead of permament magnet), do you think the L298 will be enough, or should it be used in dual parallel configuration ?

Hello,

No problem, the L298 is powerful enough to drive an old Märklin AC motor with a coil magnet inductor, there is no need to use two L298 in parallel.
You must use "Modparam" to set the current limit to its maximum value for every PWM module.
If this is not enough, (e.g.: because a loco has a smoke generator), you might have to add one or two 2.2 ohm resistors in parallel on the existing ones (R8-R11, R22-R25), to raise the current limit.

Kind regards,

Joël