The Viability of Solid-State Components for Signal Systems

This forum is dedicated to Riding Scale Railroading with propulsion using other than steam (Hydraulics, diesel engines, gas engines, electric motors, hybrid etc.)

Moderators: Harold_V, WJH

rkcarguy
Posts: 346
Joined: Tue Aug 22, 2017 10:33 am

Re: The Viability of Solid-State Components for Signal Systems

Post by rkcarguy » Mon Oct 30, 2017 11:41 am

The relays I'm using are overkill as I wanted to be able to support incandescent bulbs, and have a 6 volt 1 amp rating on the "trigger side". I'm looking for 1-2 seconds of delay, not much. I found a post on a guy wanting to run a camera mounted on a car using the running lights/turn signal lights which blinked when the turn signals were activated. It was a similar voltage and amperage and he used a pair of 4700uf caps to smooth out the blinker so the camera saw uninterrupted power. Calc's online look like maybe I can get away with one though, I'll have to get one and play with it.

jabsteam
Posts: 47
Joined: Sun Feb 19, 2012 10:31 pm
Location: Colorado Springs

Re: The Viability of Solid-State Components for Signal Systems

Post by jabsteam » Mon Dec 11, 2017 6:42 pm

Chuck, could you explain a little more on the track requirements for an automated system please.
I read about track bonds - how and where?
Insulated joints - how? Teflon sheets between fishplates and rail? What about the bolts holding it together shorting out the bond?
How about turnout frogs? Just isolate, like HO DC or DCC?

Thanks!
Eaton Huskie electric
Invicta coal burner
Electric Goose (being built)

rkcarguy
Posts: 346
Joined: Tue Aug 22, 2017 10:33 am

Re: The Viability of Solid-State Components for Signal Systems

Post by rkcarguy » Mon Dec 11, 2017 9:21 pm

I'm planning on using phenolic sheet for my isolating joints. It's expensive stuff, but it's strong and doesn't flow electricity. I only need 14 joiners so it's not a big cost in the scheme of things.

John Hasler
Posts: 411
Joined: Tue Dec 06, 2016 4:05 pm
Location: Elmwood, Wisconsin

Re: The Viability of Solid-State Components for Signal Systems

Post by John Hasler » Mon Dec 11, 2017 10:19 pm

Don't use teflon sheet. It'll cold flow.

User avatar
ChuckHackett-844
Posts: 59
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: The Viability of Solid-State Components for Signal Systems

Post by ChuckHackett-844 » Tue Dec 12, 2017 12:35 pm

jabsteam wrote:Chuck, could you explain a little more on the track requirements for an automated system please.
I read about track bonds - how and where?
Insulated joints - how? Teflon sheets between fishplates and rail? What about the bolts holding it together shorting out the bond?
How about turnout frogs? Just isolate, like HO DC or DCC?

Thanks!
For bond wires we use the short green "ground bond" wires found at Lowes, Home Depot, etc. They come with terminals crimped on but we solder them before installation and use self tapping screws to the foot of the rail with anti-corrosion compound. Every joint (except insulated joints obviously) within a section of detected track needs to be bonded.

We have experimented with several different types of plastic for insulated joints. Ideal would be 1/2" x 3/16" fiberglass strip but I have not been able to find any.

We currently use 1/8" x 1/2" lexan that is punched on the same punch we use for the aluminum fishplates. We bolt them on with the same bolts used for the aluminum fishplates.

You also need to place an insulator between the ends of the rail so that rail expansion does not cause the ends of the rail to touch. I CNC cut these in west coast rail profile from plastic. I sell these for $9 per dozen + shipping.

I am in the process of building a mold to make plastic injection-molded insulated fishplates using ABS plastic. These will have 1/2 of the rail profile end insulator molded into them so only two of these will do the whole joint - no separate end insulator needed.

Where your insulated joints are and whether you insulate both the common and detected rails is determined by the type of system you are installing.

For turnouts I use the following placement of insulated joints:
Image

This results in no 'dead spots' where a train is not detected.

Hope this helps ...
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

rkcarguy
Posts: 346
Joined: Tue Aug 22, 2017 10:33 am

Re: The Viability of Solid-State Components for Signal Systems

Post by rkcarguy » Tue Dec 12, 2017 1:55 pm

Chuck, why break the block in the middle of the turnout? I was under the impression that the turnout should be included in the "main line" and the isolators placed such that when the train on either route was clear of the fouling point the signal system would show clear.

jabsteam
Posts: 47
Joined: Sun Feb 19, 2012 10:31 pm
Location: Colorado Springs

Re: The Viability of Solid-State Components for Signal Systems

Post by jabsteam » Tue Dec 12, 2017 1:57 pm

Thanks for the track info, Chuck!
EVERY rail joint must be bonded (or insulated), and EVERY turnout must be insulated/bonded? Wow! That's a LOT of work for converting an existing track… Especially for the Lone Wolf track owners.

Looks like a Capture/Release system would be easier to install and less track work needed, for small home type tracks.
Eaton Huskie electric
Invicta coal burner
Electric Goose (being built)

User avatar
ChuckHackett-844
Posts: 59
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: The Viability of Solid-State Components for Signal Systems

Post by ChuckHackett-844 » Tue Dec 12, 2017 3:11 pm

rkcarguy wrote:Chuck, why break the block in the middle of the turnout? I was under the impression that the turnout should be included in the "main line" and the isolators placed such that when the train on either route was clear of the fouling point the signal system would show clear.
The purpose of the diagram is to show how to insulate a turnout itself so that there are no 'dead spots' where a train is not connected.

Some people avoid the issue by just insulating the turnout from all three adjacent tracks but this is bad because a short train (box cab electric, or a car that becomes disconnected from a train for example) can stop on the turnout.

If you don't insert insulators in the turnout it forms a short-circuit in the frog and points.

There are other methods of doing it but I use this method because there is zero dead area (assuming the insulators are opposite each other). The insulators could also be in the curved section but having them in the straight section reduces the stress on the insulators.

Typically there are insulators at the clearance points on the two tracks adjacent to the turnout (off to the right in my picture) or, sometimes just in the diverging route - depends on your signal system design.
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

User avatar
BigDumbDinosaur
Posts: 554
Joined: Tue Jun 28, 2011 9:19 pm
Location: Midwestern United States

Re: The Viability of Solid-State Components for Signal Systems

Post by BigDumbDinosaur » Tue Dec 12, 2017 3:46 pm

ChuckHackett-844 wrote:For turnouts I use the following placement of insulated joints:
Image

This results in no 'dead spots' where a train is not detected.
Back when we first undertook to signal our railroad, (occasionally-heated) discussion of what to do about turnouts came up. I had worked out a scheme based upon full-sized practice (which is similar to that in Chuck's illustration), but the track committee chairman wanted no cuts or other alterations done to either the stock or closure rails due to concerns about integrity, despite a mechanically-sound insulated joint having been developed. Also, an insulated throw bar ("bridal" in Chuck's illustration) was out of the question.

Hence we do not have block occupancy detection within turnouts. The adjustable delay-on-clear function in the block occupancy detector attached to a block that includes a turnout is used to "mask" the detection gap. As noted by Chuck, if a short train goes very slowly through such a turnout, or stops entirely within said turnout, it will "disappear" from the ABS. In practice, that hasn't been as much an issue for us as was originally anticipated.

Speaking of track circuitry, below is a photo of one of our insulated joints installed on the mainline. At the time of the photo (2013), the joint had been in service for about 7 years.
abs_insulatedjoint.jpg
Insulated Rail Joint
Below is the exploded view of an insulated joint.
abs_insulatedjoint_exploded.gif
Exploded Insulated Rail Joint
Not shown are the stainless steel socket head capscrews, washers and self-locking nuts used to assemble the joint.

Next is a photo of the bonding jumper we use, this photo taken before the (machine-crimped) ring tongue terminals were soldered to the wire. The wire itself is 12 AWG THHN stranded. Incidentally, while it was still available, 20/80 rosin-core solder was used for soldering jumpers and other track circuit wiring. Nowadays, the most lead-rich rosin-core solder available through general distribution is 40/60.
bonding_jumper.gif
Rail Joint Bonding Jumper
Below is a photo of a bonding jumper that had been installed for approximately 10 years at the time the photo was taken (2013). It is still in service as of this writing.
abs_bondingjumper01.jpg
Installed Bonding Jumper
Note how the jumper is installed in a "bow tie" fashion. This arrangement causes the jumper to flex in the middle rather than at the ends as the rails expand and contract with temperature changes. One of the ty-raps had broken and fallen off, a problem that was noted with some other relatively old jumpers. We have since switched brands of ty-raps to eliminate this problem.

A test procedure used to verify continued block electrical integrity is that of comparing the track circuit voltage at the output terminals of the block occupancy detector (BOD) with the return voltage seen by the BOD's track relay, this being when the block is not occupied and the track is dry. In a Robinson failsafe circuit, the track circuit is a loop that depends on continuity through both rails in order for the track relay to pick (actuate) and indicate "block clear." Hence one can determine if bonding jumpers and other trackside wiring are electrically sound by seeing how much voltage drop occurs over the block.

In our ABS, older BODs operate with a 5 volt (nominal) track circuit; newer BODs operate at 3 volts nominal. A track circuit is deemed to be satisfactory if the return voltage is 4.5 or higher in a 5 volt track circuit, or 2.7 or higher in a 3 volt circuit. These minimum voltages allow for the increased leakage seen when the track and ballast are wet. Reliable operation is assured as long as the return voltage is at least 80 percent of nominal. Below that, the block may fail to clear, producing a false-occupied status.
Science makes it known. Engineering makes it work.

User avatar
ChuckHackett-844
Posts: 59
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: The Viability of Solid-State Components for Signal Systems

Post by ChuckHackett-844 » Tue Dec 12, 2017 4:20 pm

jabsteam wrote:Thanks for the track info, Chuck!
EVERY rail joint must be bonded (or insulated), and EVERY turnout must be insulated/bonded? Wow! That's a LOT of work for converting an existing track… Especially for the Lone Wolf track owners.

Looks like a Capture/Release system would be easier to install and less track work needed, for small home type tracks.
Yes, it's a lot of work but, like laying rail, only needs to be done once. Our track crews add the bond wires at the same time they bolt the rail together. We (the signal crew) add insulators when we have determined how the section will be signaled. I think we have something over 3,500 feet of track bonded.

Capture/Release works well IF the engineers adhere to the rules and there are not a lot of trains on the railroad at a given time.

The reason I started designing the automatic signal system is that several tracks I ran at would fail miserably when there was a large meet with a lot of visiting engineers. People forgot to hit capture buttons (leaving the block occupied), conductors were hitting buttons not knowing that the engineer had done it, on and on.

Another factor: Sometimes a coupler will come undone dropping one or more cars and the engineer will not notice. With Capture/Release the next guy may come around a corner and run into the cars sitting there. With track-circuit detection this won't happen.

A place where I have seen Capture/Release get complicated is where you have two routes that merge. For one thing there will be 4 buttons in the area (two for each route) and it can get confusing, especially at night.

Track-circuit becomes more desirable as the complexity of a railroad increases to enhance traffic flow and safety.

Another example of what can happen even with one simple siding: Suppose you have a passing siding between two "blocks" that are capture/release and the siding is long enough for one train. Assume that there is an east-bound train (train 1) sitting in the siding but the east block is captured by a west-bound train (train 2) so that block is red. There is another east-bound train (train 3) in the west block heading for the east-bound siding where train 1 is sitting (remember, train 1 released it when he got to the siding). Train 2 arrives at the west-bound siding and releases the eastern block but before train 1 hits the capture button another west-bound train (train 4) captures the eastern block. After train 4 arrives at the west end of the eastern block we have:
  • Train 1: in the east-bound siding but can't move due to the block being occupied by train 4
    Train 4: cannot get into the west-bound siding because he won't fit with train 2
    Train 2: in the west-bound siding but can't move due to the west block being occupied by train 3
    Train 3: cannot get into the east-bound siding because he won't fit with train 1
You now have a "deadlock" and it will have to be sorted out manually. This happened because the signals allowed the train to capture the block because it did not anticipate the deadlock (i.e.: the train had nowhere to go).

Can't happen? I have seen it personally many times.

BTW: This exact same scenario can happen with ABS/APB signals ...

My solid-state can avoid this kind of deadlock because it knows the status of tracks other than just those within the block itself.

But: Capture/Release is much simpler to install and will work within its capabilities. As I said at the end of the first message when I opened this thread:
ChuckHackett-844 wrote:Does this mean that this kind of system is the best fit for everyone? No, the needs and desires of a given railroad should dictate the approach to be taken.
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

User avatar
ChuckHackett-844
Posts: 59
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: The Viability of Solid-State Components for Signal Systems

Post by ChuckHackett-844 » Tue Dec 12, 2017 4:57 pm

BigDumbDinosaur wrote:In our ABS, older BODs operate with a 5 volt (nominal) track circuit; newer BODs operate at 3 volts nominal. A track circuit is deemed to be satisfactory if the return voltage is 4.5 or higher in a 5 volt track circuit, or 2.7 or higher in a 3 volt circuit.
I reliably detect trains down to 1 volt (unoccupied track voltage). That level is arbitrary, set by me. I set it at 1 volt to keep the noise-immunity higher.

In practice, the dry track voltage is usually around 8 volts. If the track voltage falls below 1 volt the system will mark the segment as being in an "unknown" state and raise an alert. If due to track that is unusually 'leaky', the railroad can adjust the track voltage higher to compensate.

I also keep the track current as low as possible to prevent rail corrosion induced by the track current. Typical current is around 0.8 milli-amps per track segment.

The exact track voltage level used to determine when the segment is 'occupied' is calculated by the adaptive train detection algorithm and is adjusted as the track condition changes (i.e.: wet/dry).

Typical track strip-chart from the signal system:
Image

The red = 1 is where the system considered the track 'occupied'.

These charts serve another purpose also - they can indicate bonded joints going bad because the trace will rise/fall slightly as the train passes through the segment. By communicating to a MOW locomotive the dispatcher can identify the exact joint where the fault is located.
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

User avatar
ChuckHackett-844
Posts: 59
Joined: Wed May 03, 2017 3:54 pm
Location: Tampa, Florida

Re: The Viability of Solid-State Components for Signal Systems

Post by ChuckHackett-844 » Tue Dec 12, 2017 5:08 pm

BigDumbDinosaur wrote:Hence we do not have block occupancy detection within turnouts. The adjustable delay-on-clear function in the block occupancy detector attached to a block that includes a turnout is used to "mask" the detection gap. As noted by Chuck, if a short train goes very slowly through such a turnout, or stops entirely within said turnout, it will "disappear" from the ABS. In practice, that hasn't been as much an issue for us as was originally anticipated.
So, on the turnout, diverging route to the siding, do you have a short section of detected track between the insulated joint at the signal and the turnout to detect that a train has passed the signal and thus drop the block to occupied? If not, does this mean that the train departing the siding is not detected until he reaches the other end of the switch?

Also: I looked on the ILS website but I couldn't find a track diagram. Is there one available?
Regards,

Chuck Hackett, UP Northern 844, Mich-Cal Shay #2
"By the work, One knows the workman"

Post Reply