Maximum lenght for the CDE twisted cable

[Dvorak]

Hello. 

In my club of friends, we are having several problems with the intergrity of the DCC signal, specially when we pass of district powered by a booster, to other one.

We are using the z21. I think the reason of this bad working, is because we don´t twist the JK and CDE cables, and we have few boosters in our modular exhibitions. 

 

I think, twisting the cables, we would improve the quality of the DCC, but, someone knows, what should be the maximum distance for the CDE cables (twisted, of course), should be acceptable?

 

Thank you very much.

 

Best regards,

 

Dvorak.

 

 

 

 

[WIMorrison]

I assume that you are breaking out from the x-bus cable connection on the z21 to get these web CDE cables? I ask because whilst the X-Bus supports Xpressnet I believe it is the Roco interpretation of Xpressnet. You may have more success using a Z21 and the B-Bus to connect boosters as that is what the port is designed to support.

 

this link provides details on Xpressnet. I suggest that using commercial twisted pair as found in Cat6 or better is the route to go if you stick with X-Bus.( The cable quality is better than telephone cable and simply don’t use the spare pairs)

 

https://www.opendcc.de/info/xpressnet/xpressnet_e.html

 

[RobjUK]

If it's purely a digital data connection and not a current-carrying cable, anything up to tens or even hundreds of metres should make no difference.

 

Using a shielded twisted pair will reduce the possibility of interference.

 

If you use Ethernet cable as WIMorrison suggests, that should work fine - but make sure you use one pair for the data and a separate wire (or pair joined together) for the ground/common connection.

 

I have seen people parallel pairs in multi-pair cable, thinking that will make it work better; it does not...

 

A single pair RS485 type cable could be smaller and neater than Ethernet. 

 

If you need a lot of the stuff, a reel of this would be appropriate - it's purpose made data cable, rather smaller than Ethernet and stranded core:

https://www.ebay.co.uk/itm/VANDAMME-GREY-SERIES-1-PAIR-100M/351543309187?epid=1139229544&hash=item51d99d3383:g:dgYAAOSw1JVaG~iC

 

(That's cheaper than we have been paying from trade suppliers...)

 

Or, If you only need around 3 - 5m, you can have an offcut for the cost of a stamp.

 

[Dvorak]

Hello, WIMorrison and  RobjUK.

 

Thank you very much by your answers, I appreciate a lot. I will read carefully the links.

 

Best regards,

 

Dvorak

(Francisco Mallols)

[peach james]

I cannot directly help with z21, but my experience with Digitrax is that with about 200' of RJ12, in 50' lengths, the signal is OK.  I haven't O-scoped it to see, but it works for supplying boosters.  I would suspect you would be able to get similar distances (50+m) total length before problems start happening.   I'd also expect that if you have it as a single length without any interruptions, it will work over a much longer length than the multi cable approach I ended up with.  (it's not a good way to go...but it's where I'm at, and it works for me...but I don't recommend it !)  Twisted Pair, in shielding, is going to be far more efficient than RJ12 cable is for preventing interference.   I did find I had to split my layout into 2 sectors for the Digitrax system to work reliably (basically, controller/boosters are on one side, feedback modules/throttles on the other of a loconet repeater). 

 

James

 

 

[O-Gauge-Phil]

I should know the maximum distance of this type of data transmission as I used to design systems like this before I retired but I think it is one or two kilometres.  However that is using the correct data cable.  One common mistake is to use a large diameter cable(to reduce the resistance, this is done on DC  circuits) , this increase in capacitance increases the electrical reactance (= 1/2 pi f c). When dealing with high frequency the electrical resistance is negligible compared to the reactance caused by the frequency x capacitance.  

 

Phil

 

 

[Junctionmad]

As an aside , are there any written up specs for a CDE interface 

[Ron Ron Ron]

25 minutes ago, Junctionmad said:

As an aside , are there any written up specs for a CDE interface 

Yes.

The CDE interface is formally called the “Power Station Interface” by the NMRA and is specified by Standard S-9.1.2

The CDE wires form the “Control Bus”,  i.e. the NMRA universal link between Command Stations and external Power Stations (otherwise known as Boosters).

 

Ron

Edited October 27, 2019 by Ron Ron Ron

[Junctionmad]

Thank you

 

 

[Junctionmad]

2 hours ago, Ron Ron Ron said:

Yes.

The CDE interface is formally called the “Power Station Interface” by the NMRA and is specified by Standard S-9.1.2

The CDE wires form the “Control Bus”,  i.e. the NMRA universal link between Command Stations and external Power Stations (otherwise known as Boosters).

 

Ron

I did some more research and came across discussions on a Dutch and German site.  The CDE interface isn’t identical to the NMRA spec as it contains the E line , emergency stop line which is not specified in the nmra spec , however it’s likely the CD lines use the NMRA spec 

 

essebtially 

 

C is DCC data

D is common 

E is booster short circuit output indecator , pulled to D by an opto coupler used to inform command station that the booster has shorted 

 

dave 

Edited October 28, 2019 by Junctionmad

[Crosland]

9 hours ago, Junctionmad said:

I did some more research and came across discussions on a Dutch and German site.  The CDE interface isn’t identical to the NMRA spec as it contains the E line , emergency stop line which is not specified in the nmra spec , however it’s likely the CD lines use the NMRA spec 

 

essebtially 

 

C is DCC data

D is common 

E is booster short circuit output indecator , pulled to D by an opto coupler used to inform command station that the booster has shorted 

 

dave 

 

It depends. The NMRA spec covers two types of interface.

 

What they call unipolar (I would call single ended) has one data and one common.

 

What they call bipolar (I would call differential) is similar to the track signal and has no common.

 

A lot of boosters will work with the single ended signal. The Hornby booster seems to require the differential form.

 

[Ron Ron Ron]

My understanding is that the C&D conform to the NMRA standard.

The E is optional and provided for in the standard (feedback), but undefined.

 

There’s a bit of history about the relationship between Command Stations  and Power Stations (Boosters) in the original formation of the NMRA DCC standards and the wrangling between certain participating manufacturers, who were resisting the degree of universal interoperability between system components.

Their aim was to preserve as much hardware “ lock in “ to their own products, while still buying in to the aim of a universal DCC set of standards.
D******x were apparently the worst culprit in this.

As such, the original DCC S&RP’s were launched without a defined standard for the Command Station to Power Station interface.

That was eventually agreed a few years later on, but made optional and only for external CS to PS links, as some manufacturers ( e.g. Digitrax) had chosen a different method of communicating the Command Station signal to their own Boosters (e.g. via Loconet ).

 

Ron

[Junctionmad]

23 hours ago, Crosland said:

 

It depends. The NMRA spec covers two types of interface.

 

What they call unipolar (I would call single ended) has one data and one common.

 

What they call bipolar (I would call differential) is similar to the track signal and has no common.

 

A lot of boosters will work with the single ended signal. The Hornby booster seems to require the differential form.

 

The common being D in all the diagrams I’ve seen 

 

 

[Crosland]

6 hours ago, Junctionmad said:

The common being D in all the diagrams I’ve seen 

 

 

 

For the CDE interface, maybe, but the thread has drifted onto discussion of the NMRA power station interface, S=9.1.2, which also includes a differential mode.

 

A differential signal has no common.

 

Here's what S-9.1.2 says "If the input signal is bipolar, no particular relationship of either part of the differential signal to a ground reference may be inferred. Power Stations should interpret the signal at their inputs as differential in nature and should not require any implied 3rd connection as a reference. The Power Station’s response to this differential signal should not be adversely affected by common mode voltages as great as 12V of either polarity between the Command Station's internal reference and that of the Power Station, as happens if both were powered from a common power source.

 

If the input signal is unipolar, the signal will be referenced to signal ground."