Recommendations for block detection and signal control?

[GaryHinshaw]

I've been drooling over Chris Kilroy's installation of block detection, signalling, and CTC on his mega-Tehachapi pike, and I want to start planning mine.  I'm here to solicit recommendations on the how to implement such a system given what I already have in place.

My goal is very similar to Chris': CTC control of mainline routes, with track- or dispatcher-controllable signals at check points and track-controlled intermediate signals.  I have a pretty good idea of the signal locations, hence the mainline block locations.  I have a standard Digitrax system (DS100 command station) with several isolated power districts.  The thing that is a bit non-standard is that every section of track has its own feeders from a 12-gauge bus and each section is essentially isolated because I don't use rail joiners, as in the top half of the diagram below:



I think the Digitrax BDL168 has the functionality I want, but I definitely do not want to detect each section of track separately, since that would be expensive and unnecessary.  Since my signal blocks align reasonably well with my track sections, it seems to me the simplest way forward is to tie the A rail feeders for one signal block together into one feeder and detect the current in that feeder, as in the bottom diagram (for a 3-section signal block).  Does this seem like a reasonable plan?  I'm not thrilled about running what amounts to a parallel bus to make this work, but it's not too bad. 

Any recommended alternatives that interface easily to LocoNet, and JMRI/CATS? Would you recommend an RD2 for remote detection in this application?  The blocks will be physically well separated from the actual BDL168 (assuming that is the approach).

Thanks in advance,
Gary

[John]

I did mine like the bottom one in your drawing ..  I am using a combo of BDL168s and SE8C .. but also have a bunch of RRCrkits BOD

[jagged ben]

We are using RRCirkits Tower Controllers on the club layout.  As far as I can tell they are the most versatile and affordable components for a complete approach.  The CT detector coils are good for being non-invasive (i.e. not actually wired into the track power circuit, and thus not consuming any DCC power), and be remote from the Tower Controller connected by light wire (we use Cat-5).   The tower controller can do pretty much anything you'll need it to do, and usually in multiple ways.   

As far as your scheme of running what amounts to parallel busses for track feeders, I think something like that is going to be necessary for any sensible type of block detection.  It does certainly beat spending more money on unnecessary decoders.

I think you may find that you need/want more block detectors than you initially imagine.  For example to have a proper underlying 'early tumble-down' on Tehachapi you'll need 'half blocks' in the sidings.   Another thing to consider is automatic train operation.  Maybe that's something you have no interest in, but if you want to have any kind of future option for it, it's worth planning ahead and making separately detected 'stop sections' in front of every signal.  For that matter, those could aid in human controlled train operation in areas with compromised visibility, such as your helix. 

[C855B]

... Another thing to consider is automatic train operation.  Maybe that's something you have no interest in, but if you want to have any kind of future option for it, it's worth planning ahead and making separately detected 'stop sections' in front of every signal. ...

I'm a little lost here, probably not knowing all the details - why would it be necessary for stop sections to be separately detected? If your loco fleet is programmed for brake-on-DC, simply having an isolated section on a relay would supply this function in the face of a red board. I've seen this work in a hands-off display setting protecting an at-grade crossing, although the example I saw the trains were running too fast and jerked to an exaggerated stop.

At any rate, braking sections wouldn't work well (i.e., wouldn't work at all!) with running helpers as Gary does, at least not without putting transponders in everything including rolling stock, and a whole bunch of complicated programming. 

The design in my head for the GC&W is to have detection for whole blocks and the full length of the siding, but, importantly, separate detection within the fouling area of each switch in signaled territory. That's what I'm familiar with in 1:1.

[jagged ben]

I'm a little lost here, probably not knowing all the details - why would it be necessary for stop sections to be separately detected? If your loco fleet is programmed for brake-on-DC, ...

With separately detected stop sections the computer just stops the train when it reaches the stop section.  No loco programming is necessary.  Transponding wouldn't be strictly necessary either, although probably helpful.

Either way one has to do extra wiring for the stop section.

At any rate, braking sections wouldn't work well (i.e., wouldn't work at all!) with running helpers as Gary does, ...

I agree that automatic train running probably wouldn't work well at all with helpers, period.  Probably one wouldn't try that.   With that said, the computer controlled method with detected stop sections could have a chance of working, whereas brake-on-DC would not.

The design in my head for the GC&W is to have detection for whole blocks and the full length of the siding, but, importantly, separate detection within the fouling area of each switch in signaled territory. That's what I'm familiar with in 1:1.

I agree completely.  For example, without separately detecting the turnout, you could have the end of a train on the main fouling the turnout, while meanwhile the turnout is lined for the siding and a train behind it has a red-over-green.  Crash.

[GaryHinshaw]

Excellent - thanks for the feedback.  The RRCirKits components look really good, reasonably priced, and relatively straightforward to set up.  Automated control is not really on my radar.  I'm lucky to have plenty of crew members on call, and when I'm alone, I'm content to run 1 or 2 trains by myself.  I will eventually want protection for the lower deck swing gate, but a simple interlock should be fine there.

Cheers,
Gary

[John]

Gary .. on the M&O, I use a combination of the old RRCkts BOD-Hs .. and the digitrax BDL16 .. I have two of the BDL cards, and probably another 20 or so of the older detectors ..

The BDLs are on the lower level .. I put them in strategic locations so that the power runs arent too long .. digitrax also makes a remote diode thing that will allow you to send detection back to the BDL using three Cat5 type wires .. I have a couple extra if you need them ..

The BODs I use on the upper level, and use them to send occupancy info to either the SE8C or DS64s ..

I run a master power bus along the layout .. then use a aux bus for each detection section .. and run the track feeder from that .. remember, you only need one aux wire .. the other track feeders can go onto your "common" line ..

make sense?

[John]

Of course, all this stuff feeds a JMRI computer, which then controls the signals off SE8Cs .. I have 4 of those .. and will probably need another ..

you should be able to do the 2 light signals that UP uses out there

[babbo_enzo]

Hi all
it's some time I also was analysing what is on the market - what/if I can scratchbuild- to add signals on my layout.
So I'm reading with interest all post here.
Aside all cards some of you mention, nobody have experience with this product:
http://www.cmlelectronics.co.uk/index.php/sigm20
I have an NCE system, and seems all solutions ( well, nearly all ! ) take the path to have a parallel accessory bus using Loconet. Comments on this ?
Cheers

[mmyers]

Enzo,
 You will have to run some type of parallel system because quite frankly, NCE has never brought their signal controller to market. They do have detection and those detectors work with other controllers.
CML has a nice system as does RRCirkits. All are compatible with Digitrax components. At some point there will need to be a signal controller and if it is to report via the network, you will need a network. I believe there may be a way to tie your NCE network to a Loconet network through a computer but don't have any practical knowledge on that.

Martin Myers

[babbo_enzo]

Martin
short answer is "yes" : using a LocoBuffer or a PR3 driving the Loconet around the layout and bridge to a PC running JMRI.
It depends on what you want to automate or display.
I'm not sure to want to automate the turnouts and signals from a PC and display the panels on the LCD screen...
As my first step, I want to proceed as I've done with turnout control: I use Tam Valley octopus and servo but no decoders (for the moment) just front panel manual switches and Leds to simulate the prototype human manual job.
For signals what to do?
- Block detection first
- Then signals controllers (signals 3 LED) and front fascia leds repeaters.
- Maybe later, but not mandatory block to stop trains and feedback to display tracks on a computer pane.
My layout is not big and no plan to have remote CTC...

Good to know you give a "star" to the mentioned controller.

[GaryHinshaw]

I run a master power bus along the layout .. then use a aux bus for each detection section .. and run the track feeder from that .. remember, you only need one aux wire .. the other track feeders can go onto your "common" line ..

make sense?

Makes perfect sense, John.  Thx.  If you were starting fresh, would you go with Digitrax or RRCirKits components for the detection task?

For the group, do you add resistors to your rolling stock wheels?  Any favourite methods?  Do resistive wheels ever become an issue for long trains with all those resistors in parallel?  If I have one 10K Ohm per car and a 50 car train, I've got 200 Ohms across the rails, which is naively giving ~50 mA of additional current (which is what I'm detecting if there is no loco in the block).  Does this affect train control at all?  At the very least, I think this would nearly double my power requirements for boosters, no?

-gfh

[C855B]

Power consumption from wheelset resistors - even those parked in sidings - should not affect train control unless you have so many locos running that you're already on the edge of power supply capacity. But, yes, you do need to consider increasing booster capacity, but it shouldn't be double. Maybe 20%.

I am still "fooling with" ideas for putting resistors on wheelsets. At this point it looks like an 0603 SMD resistor (0402 is too small to handle easily) tacked across the insulator with a dot of epoxy... maybe CA is OK, still prefer epoxy... and then "soldered" to the axle and wheel with conductive paint used for circuit board repair. Easy peasy.

FWIW, my grand plan is to mitigate some of the need for wheelset resistors in the overall system by also deploying optical detection of either method, IR pairs or ambient, as a secondary system. Occupancy is occupancy.

[John]

Makes perfect sense, John.  Thx.  If you were starting fresh, would you go with Digitrax or RRCirKits components for the detection task?

I would probably go with rrcirkits .. its well engineered and Dick Bronson and his wife are super people to work with ..  the only drawback .. i think its a two person shop ..

[GaryHinshaw]

Thanks Mike.  Keep me posted if you do come up with an smd solution that you like.

the only drawback .. i think its a two person shop ..

That was my concern, especially since they are not especially young.  But if things are installed and working properly, supply is not really a problem.