Hi, I've got a strange problem.
I use an ecos 50200 and also have an old Marklin cs1 connected to the sniffer port on the ecos in order to get an extra couple of control knobs.

This works fine with both locos and accessories.
However, in order to control MFX locos the loco must have a sniffer address set.

I have eight MFX locos, some have Marklin decoders and some have ESU M4 decoders (Lokpilot V4).

Most of these work fine but a few of them won't let me set a sniffer address. I set it in the loco edit screen but it doesn't update to the decoders (other changes to the settings update to these decoders ok).

The problem is not limited to one type of decoder - I have ESU AND Marklin decoders that take a sniffer address and both types that don't.

Has anyone got any ideas please?

All non MFX locos simply have the sniffer address set to zero and work fine.

I have raised the voltage on the ecos to just over 19volts as I read somewhere that MFX like a highish voltage to accept changes. I don't know if this is true but it made no difference.

The software version on the ecos is 3.7.

Thanks for any help,
Paul

Edited by user 19 February 2014 13:38:36(UTC)  | Reason: Added ecos software version

What about the railcom feedback setting ?, same suggestion as Peter, get in contact with the English version ESU support forum, what I find strange thou is the Märklin mfx decoders do it as well.

John

Thanks for the replies and suggestions/explanations everybody.

I am aware that mfx decoders have an address stored for operation on non mfx controllers. That address is storing ok.

I too assumed that the sniffer address was not stored on the decoder but only needed to be entered in the controller.
However I tried deleting the locos and then re-detecting them and some still showed the sniffer address whereas others (the ones that don't work with the controller attached to the sniffer port) had returned it to zero.
I assumed then that in some cases it was being stored in the decoders since it still showed.

Maybe they hadn't deleted properly.

Incidentally I tried all this with rail com both on and off. I also tried it with the asymmetric track current both on and off.
Niether of these made any difference.

I will make sure that all mfx locos are on the track and that none of them have any functions or speed ste[s set and that none are selected in a controller. Then I will use the 'delete all mfx (m4) locos feature of the ecos. I will then take them off the track for a while then put them back on and let them all re-detect and see what the settings show.

If I still can't get them all to work when I enter sniffer addresses I'll take some relevant screenshots and post them up.

Thanks chaps,
Paul

This is the classic symptom of the CS1 "recovering" from a situation where poor programming from a PC/mobile App over the Ethernet is not designed correctly.

Do you use a PC or MobileApp ?
I have found the CS1 in particular DOES NOT LIKE THE BEHAVIOR of these remote apps WHENEVER a configuration change takes place.
More to the point the warmstart will happen after an accumulation of errors (lacke of state change handling) and any changes you have made will be lost - having a battery in the CS1 has "debatable" effect on this.

More too the point of this thread I cannot see how a CS1 configuration would not allow for the ECoS sniffer values not to save as you describe. Unless you meant to say that it did save but didn't work (because the CS1 was sending a different address than you expected.

Just a final update on this problem.
As I suspected it was something that I had done on the CS1 that was confusing it - two things in fact .

Firstly, I had two BR216 loks registered separately but also registered together as a consist. I had been running one of them singly. Whilst the ecos doesn't seem to mind this the CS1 seems to prefer that you delete the consist if you want to run the loks singly. The mystery change of address that the registration on the CS1 of the troublesome mfx lok experienced was to one of these BR216 addresses - coincidence perhaps but I doubt it.
The second thing I had done wrong was that I had the programming track output from the CS1 plugged into the sniffer input on the ecos instead of the CS1 main output. I don't know how much difference this would make as it still seemed to work ok but it can't have helped things.

I have since made sure that it was all plugged together correctly and done a factory reset of the CS1. I reloaded all the locos and they all work fine including all the eight mfx locos that are auto detected on the ecos. All eight have a sniffer address set and the CS1 connects to and controls them just fine.

So, the moral of the story is - check what you're doing before you blame the equipment

Thanks for your help and suggestions - it was a good learning experience.

Paul

This shouldn't make any difference as there is a relay inside the CS1 that normally connects the programming track output to the main track output.