DCC Block Detection

Our club layout has a few block detectors scattered around. They're made by Integrated Signal Systems. Does any one have a instruction booklet for
these? I want to know how they're supposed to be installed. It looks like the fellow who installed them put them in one feeder to the block, but left the bus alone.
I've observed an interesting behavior where the signal in front of an approaching train changes before the train crosses in to the next block. There's a good 15-18 inches before the signal when it changes.
The parts of the signal system I'm referring to are here: http://integratedsignalsystems.com/electronics/index.htm
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper wrote:

That's because the 'Electrical' block starts just before there. Look at the gaps in the rail.
Chuck D.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The electrical block starts immediately under the signal bridge and continues through a turnout to single track before stopping at a second turnout continuing on to double track. (The double track reduces to single track to go through a tunnel.)
Where the signal changes, there's still 12-18" to the gap.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper <puckdropper(at)yahoo(dot)com> writes:

The behavior you're describing points pretty clearly to either shorted blocks, a sneaky current path that closes based on turnout position, or something conductive making intermittent contact with the rails.
It's rather irritating that they don't have the full manuals for the detectors on their site. Not even a part number so you could find it somewhere else. I was able to find in their catalog that they are dcc compatible - some of these kind of systems aren't, so you're in good shape there.
From the picture, it looks like these go in-line to one rail of the track - they're not induction based like the team digital or digitrax detectors. That means you need to have a very solid break on both ends - make sure one end or the other hasn't expanded together against the adjacent blocks. You also cannot share feeders - check if someone maybe patched in a power feed hooked to your detector's rail input for some other block. Sloppy stuff happens, especially when 'it doesn't matter'.
Another thing to check for is any sneakage from your turnouts - make sure the power routing is working properly and there's no chance that the non-switched leg of the turnout is supplying power to your detection block (or the block you're leaving). That one rail needs to be 100% isolated from everything else, 100% of the time. Sometimes accomplishing this can be brain-seizure inducingly difficult.
Good luck with this - it can be a pain to solve, especially when turnouts are involved. I have a lot of turnout and signal automation on my layout, and it tooks months to work all the kinks out.
--
* PV something like badgers--something like lizards--and something
like corkscrews.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
pv+ snipped-for-privacy@pobox.com (PV) wrote in

The detected feeder is the last feeder before the signal bridge gaps, so it seems like the next detector is detecting something before it should. (Some sort of sneaky current path or capacitive effect from having the wires so close?)

I'm going to give them a call tomorrow, and see if I can get anyone to answer the phone. It looks like after their move they just gave up doing business.

Please expand on the part about not being able to share feeders. Is it ok to go from the bus, through the detector to the rail, and then have the next block go from the same bus through another detector to another block?
Do all feeders for the block need to pass through the detector (essentially splitting the bus--series), or will everything work ok with the detector in parallel to the other feeders?

Wouldn't double gapping the divergent rail side and feeding the turnout from the bus solve the isolation problem?

Thanks. I powered up the signals tonight and everyone was impressed. Next week, they'll be hooked up to the 12V bus permanently.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper wrote:

Yes!
YES!
or will everything work ok with

NO!
Yes! Not a bad way to do things as a matter of course, but because that would take more time, effort, and materials, during initial construction, it often doesn't get done.
Chuck D.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Thanks Chuck. That might explain a little of the odd behavior. One detector is in parallel with the other feeders to the block.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper <puckdropper(at)yahoo(dot)com> writes:

I was talking about someone patching into the wire leading from the detector to your detection rail. I've never used these units, so I can't tell you whether a voltage drop (which shouldn't be dramatic anyway unless you're using too-thin wire) on the booster side could cause you trouble. If you can do a home-run back to your booster's distribution panel, by all means try it! It lets you eliminate another possibility.

You really should connect all feeders to the block to the rail side of the detector. But, not doing that would probably cause false negatives, not positives. Generally, detection blocks are short enough that you can pull the other feeds anyway, at least temporarily. Try it and see if the problem goes away.

Unless something has fallen into the turnout and is making intermittent contact, I'd say yes. It's good practice to isolate any turnout involved with detection for this reason. Note however that different manufacturers (or even different lines from the same manufacturer - Peco, I'm glaring at YOU) do things differently, and sometimes you have to fiddle with the rail breaks to get everything to behave. When I'm doing a detection block involving turnouts, I try to end the block at the turnout itself, briding power as needed. Having the block travel through the turnout can drive you crazy, and you might also get chatter effects on a detector without good debouncing.

My "signaling system" is mostly just showing turnout positions and is there for visual interest - I haven't done any fancy CTC stuff with it yet. But even then, it really adds something. *
--
* PV something like badgers--something like lizards--and something
like corkscrews.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
pv+ snipped-for-privacy@pobox.com (PV) wrote in

That makes quite a bit of sense. If something else is between the detector and track, it'll be detected too. A home-run would be difficult, but splitting the bus wouldn't. It'd be a great place to add a diagnostic toggle, too.

We've gotten a few false negatives. Low current draw locomotives didn't seem to trip the detector. (It doesn't trip when I've got the headlight on but the locomotive isn't moving. It would be great for approach-based grade crossing circuits.)

It looks like we'll have one block that ends after a turnout, with another one inside the block. The turnouts are Tortoise powered, so I'd think any boucing would be limited to when the turnout was moving.

Most of what we need to do is ABS, and that's really easy to do. Some parts of the main line are single tracked and blind, so it requires the operator to walk ahead of their train to make sure the way's clear.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
There have been a lot of block detector makers over the years. The circuit is public knowledge. Normally with this type of detector, the detector is in series with the South Rail (which is the common rail in common rail systems) and the North Rail is the rail that is fed the power for the train for speed and direction. If you're running seperate rails with seperate power supplies, you need to isolate the power for the detectors also as they WILL "trigger" or even destroy themselves if you start tying things together. YOU HAVE TO BE RUNNING COMMON RAIL POWER FOR THE DETECTOR TO WORK!!! It also helps if the whole trackage is setup with detectors as this will keep the train speeds constant over the whole layout. Anyplace that you have a detector will be seeing 0.6 to 1V lower voltage.
-- Bob May
rmay at nethere.com http: slash /nav.to slash bobmay http: slash /bobmay dot astronomy.net
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

No you don't. A double-isolated block is fine - just make sure you connect the ground part of the detector correctly. That said, you never really NEED to double-isolate unless you're in a reversing section anyway.

The detectors Puckdropper is using have a very low voltage drop (.06v) if I read the specs right. These aren't your father's T-detectors.
On my layout, I use team digital's current sensing detectors - you just pass a feeder wire through a hole in the detector, and it works. The downside is you need to power them seperately, but I run that alongside the detector's signal wire to a controller.
Block detector: http://teamdigital1.com/prod_catalogue/dbd22_product/dbd22.html
Signal controller: http://teamdigital1.com/prod_catalogue/sic24ad_product/sic24ad.html
With this stuff and JMRI, you can do all sorts of automation and sensing on your layout. *
--
* PV something like badgers--something like lizards--and something
like corkscrews.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Nice to be able to detect small voltages. One thing that the site doesn't mention is what is used to detect the current going through the detector. I will note that the use of Schottkey diodes will drop the voltage across the detector and also decrease the current needed to get enough voltage to trip the detector. Your detector, if, as you say, is just running a wire through a hole, is probably a current detector (one turn transformer in correct terms) will be a lot more sensitive to DCC power as it will be looking at the current that is developed by having to charge the capacitance of the track. DCC is using a high frequency AC voltage to send the signals to the locos so you have to deal with the capacitor problem! The detector (of which there are two on that design you mention, probably for ease of wiring them at the place where the modules are placed) is still wired the same as any Twin-T design and the same as Puckdropper's detectors. He's got something that will work but, if his detectors detect that little of a voltage, the current needed is quite low and he could be having problems with leaky ballast on his track. I basically doubt that because that would cause the next block to detect the train at any point along the previous block. I work in electronics so I understand what is going on! -- Bob May
rmay at nethere.com http: slash /nav.to slash bobmay http: slash /bobmay dot astronomy.net
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I wish I could tell you more about what the detector is doing itself. All I've bothered to look at is there's a couple diodes and a LM324 IC (I'm fairly certain it's a LM324). Sure sounds like the Twin-T style as described here.
I've found a schematic for a detector that uses a coil and wire loop as PV described above. It doesn't look like it'd be more than about $10 each to build (probably around $5 in quantity), and senses down to a few mA. http://home.cogeco.ca/~rpaisley4/DccBODvt5.html
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Yep, you've got the standard Twin-D (d for diode version of the Twin-T detector) design. Two of the op-amps in the chip (the LM324 which is a quad op-amp - 4 amps in one package) are used to detect in either direction the voltage across the diodes, one to combine the signal and the 4th to output the signal. When the current thhourgh the diodes gets high enough to produce some voltage, the circuit sees it and puts out a signal. I'd have to actually play with the particular design to find out exactly what is going on but the first thing is to verify exactly where the module will trip on the track. The thing that is really strange is that the engine approaching the end of the block trips the next signal and I can't really think of anything that would cause that unless the engine rolls onto a last part of the track which is electrically isolated from the rest and is fed power directly from the module and the DCC HF signal is feeding across the wires going up to the track.
-- Bob May
rmay at nethere.com http: slash /nav.to slash bobmay http: slash /bobmay dot astronomy.net
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

It's time to do some snooping under the layout. I'm supposed to be documenting the layout electrical system anyway...
The two detectors are wired in parallel with other feeders (which now I know not to do), so perhaps when the locomotive gets close enough to draw most of its current through the detected circuit there's enough of a draw that the other detector sees it. Without actually measuring it, though, this is just a guess.
Thanks for all the help,
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper <puckdropper(at)yahoo(dot)com> writes:

I may have to give that a try someday. It's a nice circuit that I could easily think of other uses for. Thanks for the link.
The team digital ones go for $17-18 each depending on where you buy them. If sensitivity is an issue, there's a spot on the board where you can wire in a resistor to lower it. I haven't had to do that though. *
--
* PV something like badgers--something like lizards--and something
like corkscrews.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper <puckdropper(at)yahoo(dot)com> wrote in

I gave them a call today, and got a response. ISS is still around, and the fellow I talked to (probably the owner) is going to send me the manuals.
Apparently the block detectors don't always work well with DCC. They were designed for a different command control system, using a common rail and signal rail.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Puckdropper <puckdropper(at)yahoo(dot)com> writes:

They need to rewrite their ad copy. You don't assume DCC to mean "some systems, but probably not the one you're using". It definitely says DCC compatible.
I put up some links to an alternative system in a seperate message. *
--
* PV something like badgers--something like lizards--and something
like corkscrews.
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
pv+ snipped-for-privacy@pobox.com (PV) wrote in

Well, from the looks of it their website has been neglected since 2007. IMO, it would be worth half a day's work to update it and finish the electronics page descriptions. Uploading manuals would be great too.
Puckdropper
--
"The potential difference between the top and bottom of a tree is the
reason why all trees have to be grounded..." -- Bored Borg on
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
The detectors are compatible with DCC. One wire from the DCC station needs to be attached to the ground power wire for the power of the signal detector power supply. That is the common side of the power. It is, in the common rail system called the South Rail. Multiple DCC power stations means that you need a seperate detector power supply for each of hte DCC power stations. Either that or you need to connect one side of the DCC power stations to each other.
-- Bob May
rmay at nethere.com http: slash /nav.to slash bobmay http: slash /bobmay dot astronomy.net
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.