General Q's on DCC and layout building

[Suzie]

Hornby points are not 'Insulfrog', and neither are Fleishmann. Only Peco make Insulfrog - and fortunately not for much longer.

[RobinofLoxley]

8 hours ago, Suzie said:

Hornby points are not 'Insulfrog', and neither are Fleishmann. Only Peco make Insulfrog - and fortunately not for much longer.

OK I now need to fess up to a bit of idiocy. I assumed that Insulfrog was a generic term for DC era points lacking the extra connectors required that are present specifically in Peco Electrofrog points. Most of my points are actually Hornby, which explains why I didnt have the reported problems. To be fair to me, this distinction is often not made, because I suppose people assume that everyone knows. Much happier this morning as one big task has crossed itself off

[grriff]

I think there is a misunderstanding of insulfrog points here. The difference between electrofrog and insulfrog points is here:

 

[Suzie]

Insulfrog is Peco's trademark for their particular brand of dead frog points. Only Peco make them. Any other dead frog point is not Insulfrog.

 

Fortunately Insulfrog is being phased out in favour of Unifrog which does not have the same problem.

[grriff]

For those wanting to know the difference between insulfrog and unifrog, the details are here: https://dccwiki.com/PECO_UNIFROG. 

[EsK]

On 19/09/2020 at 08:22, WIMorrison said:

getting it right at the start is well worth the planning effort as it can be much harder to implement after you have built the layout.

Hello Iain,

 

I am butting into this thread but I definitely agree with you on this, what with the occupancy detection troubles that I am facing now!

 

I quote from a previous post of yours

"a) make sure you haven't used thin wire as this acts like an aerial and will 'collect' current from anything nearby."

 

Could you please help with what is meant by 'thin' wire here & how does it act as an aerial?

Thanks,

Krish

[WIMorrison]

Krish 

 

when you use something like 7/0.2 on long runs in bundles beside other feedbacks it is possible to get false occupancy caused by induction from the ‘active’ wire to other wires in the bundle. The DR4088xx is extremely sensitive and this induced current can be enough to cause it to sense occupancy when there isn’t occupancy.

 

the trick is to either use heavier wire, less wire in the bundles, shorter runs  or to use the resistors that we have discussed elsewhere which ‘desensitise’ the feedback.

 

one chap had lots of issues as he had put the cables through metal eyelets to hold the cable runs and the false signals were horrible - removing the eyelets cured 95% of the issues and a few 1k resistors sorted the rest, all without rewiring

[melmerby]

 

13 minutes ago, WIMorrison said:

Krish 

 

when you use something like 7/0.2 on long runs in bundles beside other feedbacks it is possible to get false occupancy caused by induction from the ‘active’ wire to other wires in the bundle. The DR4088xx is extremely sensitive and this induced current can be enough to cause it to sense occupancy when there isn’t occupancy.

 

the trick is to either use heavier wire, less wire in the bundles, shorter runs  or to use the resistors that we have discussed elsewhere which ‘desensitise’ the feedback.

 

one chap had lots of issues as he had put the cables through metal eyelets to hold the cable runs and the false signals were horrible - removing the eyelets cured 95% of the issues and a few 1k resistors sorted the rest, all without rewiring

Twisted pairs help as well. e.g. two insulated wires twisted together to every sensor rather than a lot of common side wiring and one "signal" wire to each sensor.

 

I don't bundle any wiring either.

Edited September 21, 2020 by melmerby

[EsK]

Thanks Iain & Keith. If it's a single wire, then I am using 30/0.1 (which is the same cross-sectional area of 7/0.2) & if it's two insulated wires twisted together, those are 1/0.5 each. 

 

Are these both suitable for N scale, where the longest block is 110 cm and the longest feeder length is 5 meters.

 

Krish

 

Edited September 21, 2020 by EsK

[Michael Hodgson]

On 18/09/2020 at 18:58, RobinofLoxley said:

On 18/09/2020 at 18:58, RobinofLoxley said:

 

Given the end to end I will have locos facing the buffers and have a reversing loop planned in the layout – I would put in 2 but cant find the room. I understand that there is a digital switching device available but my question is, does it maintain loco orientation when the polarity switches?

 

 

 

One thing DC users don't immediately realise is that direction control in DCC is different from DC.  You tend not to have centre off controllers, but a speed controller plus a direction switch.  On a DC layout with that arrangement if you slide the direction switch to the right the train goes to the right regardless of whether it is running chimney first or tender first.  Under DCC you select Forward or Reverse, which means if you turn the loco bodily you have to set the direction switch the other way if you want it to go the same way.  Same as a 3-rail layout.  It's a difference in feel from using an analogue controller that's fairly easy to get used with steam locos which have an obvious front and back, but much less so with symmetrical diesel or electric locos.  If you don't want to determine your intended loco direction by trial and error but don't know the A end from the B end of your diesel, first turn its lights on and see which end lights up!

 

At the risk of stating the obvious - if you are going to modify the wiring of your points, do it before your lay the track!

[WIMorrison]

1 hour ago, Michael Hodgson said:

 

....  Under DCC you select Forward or Reverse, which means if you turn the loco bodily you have to set the direction switch the other way if you want it to go the same way...

With DCC forward is forward always. If you pickup the loco and turn it around then forward is still forward, just in the opposite track direction, but the controller is still set to forwards.

 

The chimney is at the front end and travelling with the chimney at the front is forwards whichever way the loco is facing. If travelling tender first then you are in reverse, always and again it doesn’t matter which way the loco is facing, reverse is always reverse.
 

With DC you need to change the direction switch When you turn a loco around, but not on DCC.

Edited September 21, 2020 by WIMorrison

[melmerby]

1 hour ago, Michael Hodgson said:

 

One thing DC users don't immediately realise is that direction control in DCC is different from DC.  You tend not to have centre off controllers, but a speed controller plus a direction switch. 

 

Quite a few centre off controllers out there, although not as normal.

[Nigelcliffe]

16 minutes ago, melmerby said:

Quite a few centre off controllers out there, although not as normal.

 

And quite a few configurable so they work either as centre-off, or as speed plus separate direction.

[Andymsa]

5 hours ago, melmerby said:

 

Twisted pairs help as well. e.g. two insulated wires twisted together to every sensor rather than a lot of common side wiring and one "signal" wire to each sensor.

 

I don't bundle any wiring either.

hi keith

 

sorry to disagree but twisting wires will certainly introduce capacitance, which is the biggest cause of false occupancy. When I first installed my bdl168s I had no end of false occupancy issues eventually I took the plunge and spent 3 months untwisting every wire. This completely removed all false occupancies and with now 56 bdls installed and over 800 detection sections this would of been a big headache. 
 

on a side note my track feeds are loosely laid next to each other in cable ducting which acts as a trough. No other cables are run next to these wires and they are kept as short as possible. Each zone has its own common wire. Wire used is 16/2 for the feeders and mains rated that feeds the bdls, I will add these feeds are twisted 

Edited September 21, 2020 by Andymsa

[melmerby]

16 hours ago, Andymsa said:

hi keith

 

sorry to disagree but twisting wires will certainly introduce capacitance, which is the biggest cause of false occupancy. When I first installed my bdl168s I had no end of false occupancy issues eventually I took the plunge and spent 3 months untwisting every wire. This completely removed all false occupancies and with now 56 bdls installed and over 800 detection sections this would of been a big headache. 
 

on a side note my track feeds are loosely laid next to each other in cable ducting which acts as a trough. No other cables are run next to these wires and they are kept as short as possible. Each zone has its own common wire. Wire used is 16/2 for the feeders and mains rated that feeds the bdls, I will add these feeds are twisted 

I've had quite the opposite happen.

I use RS-8s for block occupancy and I had one long run where I used just a flat pair of wires, I got false occupancy readings. I read that twisted pairs were preferred (had already done that on most blocks, as good practice during years of electrical work).

Twisted same pair together and false indications went. So proof for me that my original course was correct.

Now all my blocks are fed by twisted pairs and I get no false indications.

[Andymsa]

7 minutes ago, melmerby said:

I've had quite the opposite happen.

I use RS-8s for block occupancy and I had one long run where I used just a flat pair of wires, I got false occupancy readings. I read that twisted pairs were preferred (had already done that on most blocks, as good practice during years of electrical work).

Twisted same pair together and false indications went. So proof for me that my original course was correct.

Now all my blocks are fed by twisted pairs and I get no false indications.

I suspect the differences we are experiencing could be down to the hardware we each use.

[RobinofLoxley]

1 hour ago, Andymsa said:

I suspect the differences we are experiencing could be down to the hardware we each use.

For someone contemplating going into DCC this is the nightmare. Wiring stuff up, having it not work, and having no idea why not.

[melmerby]

1 hour ago, Andymsa said:

I suspect the differences we are experiencing could be down to the hardware we each use.

Is the BDL 168 particularly sensitive? The RS-8s trigger on about 1mA (resistive)

 

I reckon it was possibly crosstalk from another live feed which it, by necessity, had to run close to.

As the RS-8s do not seem to mind a bit of capacitance, twisting did the job. It was using 1.0mm CSA insulated solid conductor and was about 3m/10' long

 

Eliminating crosstalk in signal wires was why we always twisted whilst at work unless it used screened or co-ax cable.

Cat 5E cables have four twisted pairs for that reason.

[melmerby]

1 minute ago, RobinofLoxley said:

For someone contemplating going into DCC this is the nightmare. Wiring stuff up, having it not work, and having no idea why not.

It depends how complex you want to be.

It's best to start with the minimum  needed for proper operation and increase complexity as you progress.

Breaking it into zones is also a good idea so that any fault finding can be limited to certain part of the layout, later these can be fed from individual "boosters"

 

My layout has much more track and many more turnouts and detection blocks than when I started.

e.g. you could wire a long platform for just one detection zone to start but break the track into several separately fed sections but to start just wire them all to just one detector.

Later you could add a couple more detectors allowing three detected sections

I have done that in my terminal station, originally one long section per platform, but wired as a short-long-short to one occupancy detector, now has the two shorter sections at the ends seperately detected.

Same thing with some of the longer plain track sections, where once 4m long, now is 2 x 2m

 

Andy's 800 blocks makes my 100 look puny and I still consider it a complex layout!

 

 

[WIMorrison]

The numbers of twists per meter is also critical and is dependent upon the frequency.

 

Cat5e has a 100Mhz max frequency whereas DCC is 10Khz therefore the twist on Cat5e is too tight for DCC

[Nigelcliffe]

13 minutes ago, RobinofLoxley said:

For someone contemplating going into DCC this is the nightmare. Wiring stuff up, having it not work, and having no idea why not.

 

Which is why anyone contemplating a large automated system is advised to try stuff out on a small test-bed first, to start to get a grip on the issues that may affect their choice of equipment.  

 

For 99% of modellers, the stuff being discussed is irrelevant.  They don't have train detection systems, so don't buy any of the hardware discussed, so don't ever run into the issues being discussed.     
(We're in the same arena as few DC modellers have a clue how an analogue sequencer using a home made rotary selector works, yet Rev Denny used one on Buckingham to run the automated fiddle yard 60-ish years ago  https://highlandmiscellany.com/2017/01/23/a-return-ticket-to-buckingham-part-3-the-automatic-crispin/). 

 

 

- Nigel

[RobinofLoxley]

Do any guidelines exist anywhere, or would it be a case of coming here with a specific equipment list. I am some way off the installation of accessory decoders anyway.

[Andymsa]

26 minutes ago, melmerby said:

Is the BDL 168 particularly sensitive? The RS-8s trigger on about 1mA (resistive)

 

I reckon it was possibly crosstalk from another live feed which it, by necessity, had to run close to.

As the RS-8s do not seem to mind a bit of capacitance, twisting did the job. It was using 1.0mm CSA insulated solid conductor and was about 3m/10' long

 

Eliminating crosstalk in signal wires was why we always twisted whilst at work unless it used screened or co-ax cable.

Cat 5E cables have four twisted pairs for that reason.

in fairness I have never measured current for the trigger, but you can alter the ohm rating from 10k to 22k or vice versa.

 

i still have around another 100 detector sections to install but track to build first

[EsK]

11 minutes ago, WIMorrison said:

The numbers of twists per meter is also critical and is dependent upon the frequency.

 

Cat5e has a 100Mhz max frequency whereas DCC is 10Khz therefore the twist on Cat5e is too tight for DCC

I am using Digikeijs DR4088LN-CS for block occupancy, which are very sensitive. So in my case should twisted wires be avoided all together?

[melmerby]

9 minutes ago, RobinofLoxley said:

Do any guidelines exist anywhere, or would it be a case of coming here with a specific equipment list. I am some way off the installation of accessory decoders anyway.

First thing is to go and see some kit in action, ideally at a dealer that isn't tied to just one or two brands.

Have a try yourself, see how the stuff feels in your hands.

Decide a budget but don't try and skimp too much, you might regret it later when you find it won't do "X" or "Y"

Do you want a handset based system or just computer control? (you can have both)