Recommendations for block detection and signal control?

[jagged ben]

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

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.

Remember, it's Loconet based, so if the supply gets cut off mid-project you still have the option of mixing in other components.  Everything's compatible through JMRI.

...

We have lagged on doing wheelset resistors.  Instead we have just programmed 'double-block' protection, which basically means an additional red signal behind each train.  Much of that is entailed by 'tumble-down' and 'overlap' in any case.  And of course if you have rear helpers then it's not so much of an issue.   Our current programming is basically APB.  The dispatcher can 'fake' CTC by holding signals manually, if he really knows what he's doing and pays attention. 

In lieu of putting resistors on every car, you could make a policy of having one on the rear care of every train and/or every certain number of axles.  Of course if you're really making up trains in a yard according to realistic operating rules, that might go out the window, but for unit trains it works good.  I've done wheelsets for all of my cabooses, which works great.. except of course when I'm modeling post mid '80s.     

The resistor wheelsets I've done so far are with nothing more than a mix of shaved pencil lead and acrylic paint.  (It's tedious to get this right, but with a little patience I'm able to get 2K-5K ohms ).   I don't recommend doing this for your whole fleet, but it will work if you can't wait to get a few test cases going.  I expect I'll eventually do something like Mike is suggesting, with the smallest resistor I can actually work with.  (Heck, I still have a lot of cars to convert to metal wheels.)

One thing worth noting:  You must scratch some the blackening off the wheels to get conductivity.

[John]

Remember, it's Loconet based, so if the supply gets cut off mid-project you still have the option of mixing in other components.  Everything's compatible through JMRI.

...

We have lagged on doing wheelset resistors.  Instead we have just programmed 'double-block' protection, which basically means an additional red signal behind each train.  Much of that is entailed by 'tumble-down' and 'overlap' in any case.  And of course if you have rear helpers then it's not so much of an issue.   Our current programming is basically APB.  The dispatcher can 'fake' CTC by holding signals manually, if he really knows what he's doing and pays attention. 

have you done any work on programming CTC ?

[John]

The resistor wheelsets I've done so far are with nothing more than a mix of shaved pencil lead and acrylic paint.  (It's tedious to get this right, but with a little patience I'm able to get 2K-5K ohms ).   I don't recommend doing this for your whole fleet, but it will work if you can't wait to get a few test cases going.  I expect I'll eventually do something like Mike is suggesting, with the smallest resistor I can actually work with.  (Heck, I still have a lot of cars to convert to metal wheels.)

One thing worth noting:  You must scratch some the blackening off the wheels to get conductivity.

I am going to do SMD resistors, and add some silver conducting paint .. they sell those in paint pens ..

[jagged ben]

have you done any work on programming CTC ?

No, not yet.   It will probably take a few more years to finish the entire signalling project on the club layout and have working APB throughout.  Probably we won't do much CTC programming until that's done.

The nice thing is that with JMRI we could eventually have multiple versions of our panel.  We could turn off CTC and run without a dispatcher (our default mode!), or have a US&S type CTC panel for the dispatcher, or have a modern CTC panel...

[John]

I'm doing apb for now as well .. and will try it during an ops session .. maybe combine with dispatcher warrant control until i get the dtc in place ..

[GaryHinshaw]

In order to move ahead with my bus re-work, I need to think through block sizes and intermediate signal locations.   (BTW, this signal page has a really nice tutorial on signal operations, aspects, and indications.)  So here is my next question: is there any harm in having intermediate signals that are less than a train length apart?  All of my single track sections are about one train length, so I don't really need them, but I'd like to have intermediates at locations where the prototype had them.  Take this diagram as an example:



In my application, intermediates 1203 and 1227 would be ~half a train length apart.  As long as the signal logic is correct, it seems fine, but it could happen that a train stopping at 1227 would foul interlocking 13.  But that should be fine as long as the interlocking logic is correct, right?   (Half a train length is plenty of stopping distance.)

Of course I don't want to go crazy with intermediates, just enough to keep things interesting and quasi-prototypical.  Also, I think I should be separately detecting interlocking switches in their own blocks, so I know when the routes are clear, right?

Thanks again,
-gfh

[John]

I don't see where it would be a problem .. it will increase cost of course .. but if you are building a faithful prototype, you should at least add those ..  the JMRI logic will take care of the signal status for you ..

[jagged ben]

Gary, just consider your prototype...

http://redoveryellow.com/signals/tehachapi/index.html

Most locations on the pass have one or two sets of intermediate signals between sidings.  The locations with only one set are typically less than two train lengths apart.    e.g. as the typical train enters Walong from Woodford (or vice versa), it will still be occupying both blocks that are in-between.  Same is true (I believe) for Walong-Marcel, Bealville-Cliff,  and Woodford-Rowen.  The issues here really aren't train length.  They are a) to give a train one extra chance to slow down or stop in case it misses a signal or the status of the next signal changes, and b) to indicate whether a train is to take the main or siding at the next signal.

If I were you, I'd model the prototype signals in the locations you are actually modeling.   As for hidden areas, I'd add intermediate signals one train length apart, but only if there's a way for an operator to view the track and stop the train at that location.   Your hidden area signals could just be fascia board representations or similar, as long as this is the case.

[GaryHinshaw]

Thx jb - I put some signal planning notes in my TBC thread.  Feedback encouraged!

[ChrisKLAS]

Hi Gary,

I know I'm late to the party, but I just found this thread.

I couldn't be happier with the Digitrax BDL168 and SE8C boards for my implementation. I've had exactly zero problems using a wiring scheme identical to the second graphic in your original post. I didn't get into master/sub-busses as suggested earlier, instead opting to run a separate dedicated bus wire from the central cluster of six BDL168s to each detected block. I went this route for two reasons 1. I didn't want to further complicate things 2. given that BDL168s detect up to 16 blocks, running 16 sub-busses from a central point would consume nearly as much wire anyway. I also avoided the remote detection diodes as the only thing they save you is a few gauges on your wire (you still have to run a wire from the BDL168 to the diode, it can just be smaller wire) and, at least in my opinion, have all the makings of an unnecessary additional link in the chain with the potential to fail.

Having all the BDL168s clustered together in a central location made wiring and installation a breeze (no loconet cables, grounds, etc. running all over the place) though I won't be able to maintain this in light of the layout's imminent expansion to nearly double its current footprint (which will involve an additional 8amp booster + 2x BDL168s and 1x SE8c).

One last suggestion that will hopefully save you the head-bashing I experienced for a couple days: make sure you run your undetected bus through an unused, unmonitored slot in one of the detector boards. Prior to doing that, every time a locomotive would straddle the border between a detected and undetected block on my layout, the detected block would clear as the lesser resistance on the undetected wire compelled the current draw to flow through that side of things.

[GaryHinshaw]

I know I'm late to the party, but I just found this thread.

Thank you for your input - it will be most valuable!  And sorry for the late response - I just now noticed it.... 

... though I won't be able to maintain this in light of the layout's imminent expansion to nearly double its current footprint

Are you moving to a 6-car garage?