Signal system and signals for an N-TRAK layout?

[Walkercolt]

Current detection is considered more reliable by the "signal" gurus, and I agree. On the DCC red-line, we can go with Digitrax's signal system(but not inexpensively) but really, the DCC guys can just follow their trains and go to "restricted speed" if there's another train a module ahead of them. The hard thing is the DC lines. "Retro-fitting" everything for signaling, gets like a real railroad signal system, with all the relay boxes, detectors, and so on. And the wiring gets outragous. to detect a train in a block ahead on one track takes 4 wires. To detect the train in the block next ahead takes 8 wires. To detect the train in the forth block ahead takes 12 wires, plus power and the signal driver wiring. No wonder UP and BNSF have buried fiber-optics along the right-of-ways for signaling and dispatching. Remember the old photos with all the telephone poles and wires on them? That was for signaling and CTC. The "Great One", Bruce Chubb, suggested using an older PC and running coax or EatherNet around the layout, but then I still need to build lots of dollars worth of circuits to put data on the "buss", and take data off the buss to change the signals. With 60'X90' to work on, I'm looking at a grand for cabling...not practical. The "standard" N-TRAK signal system, just gives you a yellow or red for the block ahead of you, and costs about $3 grand to build and much, much extra wiring to hook-up(wrong), and a "black box" between blocks. We tried it, but we really underwhealmed. What we want to do really, is increase the number of trains running on each track at one time to hold the public's interest better. With a 20'-40' long freight train, you've got a lot on your hands to watch your train, talk to people and not worry about "over-running" into someone else's block, with walk-around radio throttles, following your train. I know, radio headsets could help, but that takes away from talking to the public(major N-TRAK goal) and everybody else is running headsets, and you run out of channels, even with the little 420MHz radios at a big show. The inexspensive PRACTICAL solution is what's very hard. Keep those cards and letters comin' in, folks. I've read Bruce Chubb's "Bible" on signaling, and talked to Keith Guttierez(sp???) who wrote up on a computer based signaling sytem in the '80's in MR(using all his componets , now mostly obsolete), but a "standardized" detection board and "standardized" signaling componets would bring the costs way down, and prevent 20,000 miles of wiring to troubleshoot.

[Walkercolt]

Hey thanks, everybody for your input. You all had many good ideas. I'm probably gonna call the Digitrax enginnering guru's and talk to them a bit. Maybe detect on DC track and feed the data to thier Loco-Net and thru an older PC, might be practical and cost effective. ??? ;D

[AlkemScaleModels]

And a great point about bringing signalling to NTrak to bring down the overall cost of signals.  I am trying to add signalling to my layout and it's quite crazy what the commercial products cost.  And even if you just buy commercial targets and scratchbuild the rest it's not cheap.  I am working on a relatively cheap alternative that looks pretty good though...we'll see how it works out.

We hear you. The economy stinks and our etch suppliers significantly raised their prices, more than doubling the cost of our PRR signal kits.

But, we want to promote signals on N Scale layouts. Why, because it's cool!   So, we have lowered the price of our pole signal kit to $2.99 each from $8.95. This is a new permanent price while supplies last, and we have lots of them to sell. We already have our pre-wired signal heads on sale from $23.99 to $16.99. We hope this will make signaling a lot more affordable.

http://www.geocities.com/bkempins/ASMMain/PRRSignalBridges.html
http://www.geocities.com/bkempins/ASMMain/SignalLights.html

On the bad news side, we had to raise prices on our 4 track signal bridge, but Kleins may still have some at the old price.

Hope this helps.

BCK
www.alkemscalemodels.com

[Walkercolt]

"Fearless Leader" wants semiphores at least 6" tall so you can see them from a distance, probably bigger. Detection circuit boards and signal drivers don't have a "scale". You could use them from Z to G. For "portable" layouts, we've got to figure out a way to reduce the wiring. Every connection is a potential problem, and miles of wiring equal miles of possible troubles. Don't want to have a "signal maintainer" crew, along with the "track gang" and the DCC/DC troubleshooters. Why is it never as easy as it looks? ???

[bicknell]

I'm thinking cat3 twisted pair cabling....

8 wires, ground, red, yellow, green "westbound", ground, red, yellow, green "eastbound".  One per track.

Outfitting a module with no signals would probably be around $8 on the cheap end (three wires, install your own plugs, use female to female adapters on one end).

A pair of  4 pole double through relays (perhaps $15 at jobber prices) to a turnout and you can indicate turnout position; make a tortoise or slide switch trigger the relays when thrown, send through all four wires when in the "normal" position.  Send "red" one way and "yellow" the other way when turned.

A signal bridge is simple.  In red wired to light red, wired to out yellow, in yellow wired to light yellow, wired to out green.  Green wired to light green wired to out green.  I see no need for any electronics for a simple signal bridge.

Now, block detection requires opto sensors, current sensors, or DCC transponders and then a little board to light up the right signal wires.  That's where the cost comes in.  Ideally there would be something between each set of signal bridges, but depending on DCC/DC and technology that could get expensive.

In an ideal world all signals would be opto-isolated, which isn't that expensive if someone made the right boards.

[John]

Again, I would like to point you guys to the Railmodel journal article .. they have a complete system described ..


http://www.garymgreen.com/FAQ%20for%20modular%20signal%20system.pdf


http://www.railmodeljournal.com/feb5site/Pages/signals.html


Simple Signal Systems for Modular Layouts
Railmodel Journal, February 2005 page 8
( ELECTRONICS, "FUHRIMAN, GREG", SIGNAL, WIRING, RMJ )

[Walkercolt]

I've read that article John, and it won't quite do what we'd like, and it's not cheap after a while. If someone was mass producing circuit boards, of course the costs would drop dramatically over etch your-own boards and buying componets 8-10 at a time. It's like building a car from parts...it costs 20 times what a finished car costs. I'm getting a little closer to a couple of possible solutions...but none of them are inexspensive, yet. The good 'ole "Twin-T detector" circuit costs over $20 a copy(times , and there's the wiring to deal with. Signaling seems to violate one of N-TRAK's fundamental rules.."KISS"."Keep it simple stupid!" :'( :-\

[John]

I have to disagree with you .. it does not cost 20 to put in a block detector. You can buy one already made for under that ..

Check out Rod Paisley VT5 - its DCC only, but will give you an idea

http://home.cogeco.ca/~rpaisley4/DccBODvt5.html

If you look at the rest of his site, there are signal ciruits also ..

None of this is complicated, nor expensive. You should be able to wire a 1T module for under 20 bucks .. and remember, you dont need do do every module eiither ..

[bicknell]

http://home.cogeco.ca/~rpaisley4/DccBODvt5.html

Bingo.

http://home.cogeco.ca/~rpaisley4/SimpleSignal.html

That's exactly what I was thinking of for NTrak.  Simple, and actually requires only two wires in each direction, so a 4 wire telephone cable per track could carry both directions.  There are boards for both three light signals and searchlight signals.

It would not be hard to write up a quality standard (cabling types, connectors, voltages, etc) to create a RP.  Once adopted, building these two boards could be done very cheaply in modest quantities.  A few bucks a pop. 

[John]

Leo .. the signal drivers  and detectors are pretty easy to build  .. Paisley sells the ckt boards for the detectors, and the parts are pretty easy to buy from the usual sources .

[bicknell]

Leo .. the signal drivers  and detectors are pretty easy to build  .. Paisley sells the ckt boards for the detectors, and the parts are pretty easy to buy from the usual sources .

The key to me is the standard.  I won't invest time and money in signals unless I know they will work with other people.  I want to see layouts like Louisville have working signals all the way around the red line route.

I agree the parts aren't all that complicated, the complicated part is writing it all down in a documented RP and coming to a standard on the module to module interface.  I am willing to help with details or review of that standard, but I don't have the time to drive such a project right now.

[mmyers]

  I am willing to help with details or review of that standard, but I don't have the time to drive such a project right now.

Leo,
We should talk Doug into it

[bicknell]

We should talk Doug into it

+1

I'm all for it!

[mmyers]

I've been looking at the Team Digital SIC 24. http://www.teamdigital1.com/prod_catalogue/sic24_product/sic24.html
One board mounted on a module would handle eight 3 aspect heads. That's enough to handle red, yellow, blue, and green lines in both directions on one board. Keep all of the signal and detection wiring on the module. Connection to other modules would require running a 6 wire phone wire from board to board. If the club is already running loconet, just include the boards in the throttle network.

 Loconet compatible so it can send data to other SIC24's and Digitrax compatible devices. Digitrax DCC is not required though. Can be used with a computer but also not required. To handle DC, would just need analog detection that takes the sensor input low. The NTRAK Lindsay type detection system does that so it should be workable. MSRP at $55 /board so it's reasonable.

Martin

[Walkercolt]

OK, I'm seeing lots of things compatible with DCC(which we don't need) but not so much that will work with DC. We don't need switch position signaling, as we don't change lines during shows, mainly due to the lack of #16-#20 turnouts. But we need to know when there is a train in the block ahead, or two blocks ahead, for a total of eight blocks driving 8 lower quardrant semiphore signals, and no wheelset resistors. We can't demand everyone(who we may have never met before) have a resistor on their last car...not N-TRAK protocol. At $60 each, the Team Digital boards will cost $480, plus wiring, plus plugs, plus power supplies, plus the control circuits for the signals and the power for their lights. 300 feet of even RJ-12 cable and plugs add up fast, and over that length, you need signal boosters, been there, done it with DCC. At Louisville, Digitrax "beta tested" some new equiptment, and found over the length of the loops, the "nominal" 12 volt power supplies wouldn't work, it took 15 volts to be reliable. Signal systems that work on huge home layouts are totally defecent for N-TRAK. Home layouts aren't subject to wild and weird RF interference from everything from police radios to dozens of radio throttles, and radio headsets. The system has to be bullet-proof and if possible, idiot-proof, and not take 4-8 hours to set-up and take down. Any more links or such are welcome, but remember what we're doing. I'm getting the RMJ articles to re-read. We tried the N-TRAK Lindsey signals, and were quite underwhelmed, and the system costs well over a grand(for a system that was built, and is for sale for less than cost to us.)  :-[