whiskers

Thanks all I have the answer! It turns out the code, microcontroller, power supply, DCC decoders, and track are all absolutely fine. The issue was that the H-bridge part I was using was triggering its overcurrent protection due to the inrush current when starting up with more than one device on the track - it would have been fine running the locos after startup but didn't like the initial demand and would shutdown repeatedly. Appreciate you all pitching in with ideas - gave me some things to try at least. :-)

Hey Mike, thanks for the input! The DCC signal/track voltage is 12V (that graph doesn't very clearly show the tick marks) so should be adequate. I'm running this with a large bench power supply that can deliver 150W (over 12A if needed...) so that shouldn't be an issue...

That's a very valid point - it really is odd. I'm going to try to get hold of a stock DCC controller and see what happens with that, can't understand what's happening here.

It's possible, I just can't see how! I'm not blindly blaming anything though, I've done a lot of testing and checking! I think the fault is in the non motorised part of the twin set, that's where the whining sound comes from when the failure occurs, I wonder if it is shorting the rails somehow... I'm going to take it to my supplier today and ask them to check it for me. I can't currently get hold of a controller, they seem to be out of stock everywhere...

It's a microcontroller running code that wrote connected up to a h-bridge that drives the rails. Since the microcontroller can never "know" that there is a second device on the track then how could it change anything to respond to it being added? I'm almost certain the fault is on the track side.

The DCC spec doesn't use -ve voltages. The rails are energised alternately (one +ve and one at 0V) this means that the current flow changes direction but there is never a "negative" voltage.

I have it hooked up to a dedicated bench power supply that can output 150W (over 12 amps at 12V), the second decoder is really just turning on a couple of LEDs should be using very little power...

I've heard the name Lenz from some videos but the other two mean nothing to me. I'll grab a couple of Lenz ones to have a go with, thanks!

That is fine, I did say so in my original post. :-)

They are both taken from the same setup within a minute or so of each other, nothing was changed apart from placing the non motorised portion of the twin set on the track alongside the motorised part.

The decoders are these: https://railsofsheffield.com/products/rails-connect-ros-n18-rails-connect-decoder-next-18-4-function-decoder which I assume are just rebadged products made by someone else. The DCC system is my own code based on the DCC specification which works fine with a single loco on the track but when a second loco is added there is some sort of electrical fault that occurs (see the traces I posted) - this fault isn't coming from my side of the system (it has no way of knowing a second loco was just placed on the track...). "Does the motorised car work as it should when it is the only vehicle on the track? " Yes! "Have you tried both decoders in the motorised unit?" Yes! And I've just done it again to make sure. I ran the motorised unit with the decoder it had in it (address 4 after reprogramming) and it is fine, I swapped the decoders and ran it again (address 5 this time) and it works fine. Have you any other decoders that you could try? Sadly not. If you could recommend a Next18 decoder that would be compatible (I'm new to this..) then I'll buy a couple to try out!

Originally they did - which I believe should be fine according to the spec. Subsequently I programmed them each to have different addresses and the same issue still occurs.

I have a small N-gauge loop (new to the hobby) that I'm playing with. I bought a loco + dummy pair (https://railsofsheffield.com/blogs/news/n-gauge-lner-nrm-40-hst) and installed DCC decoders in both (the dummy car allows lighting control). If I place either on the track on their own then the DCC signal looks fine (as per a logic analyser and a scope) but if I place them both on the track at the same time then the DCC signal seems to be killed completely and I can hear a faint whining sound from the DCC decoder PCBs. I've attached images below of the good and bad states in case that helps. The top trace is a digital representation of the DCC packet and the bottom trace is the analog voltage. Any idea what could be happening here? Any other information I need to provide? I'm well versed in electronics and have pored over the DCC spec to see if there are any clues to no avail... This is the DCC signal looking great with a single loco... And here it is when the second device is placed on the track... It looks almost like it is shorted out with brief pulses which I guess is where the whining noise comes from.