DCC Bus wiring - to twist or not?

[mu26aeh]

Typical TRW thread drift, but in regards to bus wiring, what are all the cool kids doing these days attaching feeders to the bus ?  This is all good timing as I set out on the big rebuild

[Bill H]

Interesting discussion re twisted bus and issues with block detection on this page.

https://wiringfordcc.com/track_2.htm#a52

Noting however, that the BD20s (NCE) are wired to a local bus.

Thoughts?

Kind regards,
Bill

[jagged ben]

My thoughts are that ya'll are trippin' and this almost certainly makes no difference to model railroad wiring function.  Twisting might in some cases reduce RF interference with radio devices, perhaps including model railroad throttles, but I think the effect would be cumulative and only show up on large layouts.

[peteski]

My thoughts are that ya'll are trippin' and this almost certainly makes no difference to model railroad wiring function.  Twisting might in some cases reduce RF interference with radio devices, perhaps including model railroad throttles, but I think the effect would be cumulative and only show up on large layouts.

Yes, this is being *SERIOUSLY* overthought (if that is a real word).   It is not like this is for some museum-size layout (um, except for Mike's layout, which is in that category).  But for an average home layout, IMO all that minutia doesn't' really matter.

[nickelplate759]

I wasn't entirely joking when I made my comment about twisting the rails.

If twisting were necessary to maintain DCC signal integrity, then the signal couldn't be successfully transmitted via the rails.

[jagged ben]

Yes, this is being *SERIOUSLY* overthought (if that is a real word).   It is not like this is for some museum-size layout (um, except for Mike's layout, which is in that category).  But for an average home layout, IMO all that minutia doesn't' really matter.

As someone who helps maintain the electrical for a museum size club layout I'd say it doesn't change that much with size.  I've never seen an issue with communication between decoders and boosters on the DCC wiring.  And we even run PWM DC throttles on the same bus with a common rail.  So there!   Communication between throttles and the DCC command station is a whole other issue with frequent problems on both the physical and software layers.  Radio signal is one of them, sometimes.  It it possible that some RF interference is coming from the untwisted DCC buses?  I can't completely rule it out of course.  But there are lots of other possible culprits, including multiple scales using multiple radio controlled systems in the same building. 

[peteski]

Yes, I also already mentioned in this thread couple of times that there is no need to fear that some RF signal will interfere with the high power DCC signal on the bus (and rails).  It is not like DCC bus is carrying microvolt-level signal, but 12V or more, and several Amps. Any interference picked up by the bus conductors acting like an antenna will be likely nicrovolt-level, having no effect on the DCC signal's power pulses.

The only real concern are the voltage spikes on the edge transitions of the DCC pulses.  They could be in excess of what the decoders can safely handle.  I'm still very curious as to what Mike sees on his oscilloscope probing the long DCC buses on his layout.

[C855B]

A month ago I promised I'd do this. The roundtuit landed on my desk, so here you go:



First (left or top) is the signal straight out of the DCS240. Second (right or bottom) is at the end of a 60-foot run of #12 at six twists per foot. Both unterminated. Second shows +/– 1V ringing ahead of the dropoff (trivial) and 14V P-P at worse. I don't see a problem.

@John , the example traces in the YouTube vid absolutely have a reactive load in the signal path. Maybe a bare motor somebody is trying to run on 00? All bets are off with that as a baseline. Should that be the case, it would be yet more justification to keep 00 out of one's operating plan.

[peteski]

So what you are showing is that twisting is not really needed to keep the signal clean?  That was the original question.

[nickelplate759]

A month ago I promised I'd do this. The roundtuit landed on my desk, so here you go:



First (left or top) is the signal straight out of the DCS240. Second (right or bottom) is at the end of a 60-foot run of #12 at six twists per foot. Both unterminated. Second shows +/– 1V ringing ahead of the dropoff (trivial) and 14V P-P at worse. I don't see a problem.

@John , the example traces in the YouTube vid absolutely have a reactive load in the signal path. Maybe a bare motor somebody is trying to run on 00? All bets are off with that as a baseline. Should that be the case, it would be yet more justification to keep 00 out of one's operating plan.

Science!

[Jbub]

So what you are showing is that twisting is not really needed to keep the signal clean?  That was the original question.

With out seeing the result of 60 feet of untwisted wire, all we can deduce from Mike's comment is that twisting doesn't hurt the signal.
I think what the question should be is, does twisting the pair preserves the signal?

[Chris333]

I think you guys got your signals crossed.

[peteski]

With out seeing the result of 60 feet of untwisted wire, all we can deduce from Mike's comment is that twisting doesn't hurt the signal.
I think what the question should be is, does twisting the pair preserves the signal?

True, but we kind of already expected that. Yes, seeing the waveform from both sides of untwisted bus would be nice to show if the twisting maeks any appreciable difference. Bot using a zip-cord and 2 individual conductors.  What would also be interesting is how a snubber added at the end of the bus affects the signal.

[C855B]

No, I was addressing the "snubber" issue from the linked video John posted. The waveform being presented there was bloomin' awful, but the assertion was long runs need to have some sort of termination. My response is "no", there was something else going on in the vid.

Since I have the 'scope out I might just take a look at the signal at the end of a random pair of loose wires to address the twist/no twist question. I doubt there will be much difference, if any.

[peteski]

No, I was addressing the "snubber" issue from the linked video John posted.

Would you be willing to make a simple snubber,  connect it at the end of the bus and see how it changes the signal shape?