Putting in saftey relays

As I see it, you need to define the qualifying event(s).  Then, we can determine how to detect that event(s).  And then what to do about it.

For example, if a pusher drives cars into a swan-dive from a lifted bridge, the cars are certainly derailing!  You'd want to immediately cut power to the engine...but it is electrically difficult to detect that cars are derailing since the power to the engine does not change much.

If you (your eyes) are the detector in all cases, then I'd think pressing the DCS red E-STOP button is the simplest way to remove power to blocks.

So to use your example, the qualifying event is a derailment in Block 1 which should instantly remove power from Block 2.  What if the derailment is in Block 2?  Should this remove power from Block 3?  Or if operating CW instead of CCW should this remove power from Block 1?

The relay part (removing power) is relatively easy.  And relays are inexpensive...maybe a couple bucks per relay.  We can provide exact wiring instructions.  It's defining exactly what you want to do that's the issue...in my opinion.

I had the same problem. Long trains with engines at both ends and across multiable blocks. I had a friend build this for me, this is above my pay grade. Works great, same system as you,  DCS, 4 blocks and four bricks.  All blocks shut down. So it is doable, just need someone to build this for you. This was built off site then installed.

PS when I asked this same question a few years ago on this forum I was told it can't be done. 

Clem 

Clem, did you get a wiring diagram with that setup that you could share? Looks very interesting to me and just out of curiosity I'd like to see how it is wired, if that's possible. It might help the OP in duplicating the setup if it could also work for him. 

I found these two OGR threads which I believe are what Clem K is talking about:

"can't be done" thread from 2015: https://ogrforum.com/topic/track-blocks

"yes we can" thread from 2017: https://ogrforum.com/...ong-trains-installed

Apparently, for Clem's system, the qualifying event is the breaker in ANY of the 4 bricks/blocks trips.  This event is detected by the loss of voltage coming from that brick.  The resulting action is power to all 4 blocks is removed.  There are additional switches to reset/override this behavior.

I was thinking the problem might require coming up with a method to detect a derailment that does NOT cause a breaker to trip.  That is, you want to quickly detect the derailment BEFORE it causes a problem (like a breaker to trip).

I do not have a wiring diagram, the gentleman that built it for me, just like some of you, likes to tinker around with electronic stuff.  

clem 

Need to add to posting I want all 4 blocks to shut down I would like something like Clem has

Here is a rudimentary circuit.  There seems to more going on in the photo.  The power for each block passes through a control relay.  The set of control relays are driven by an interlock chain made up of a sense relay for each brick.

As shown, a bypass switch could be added to each sense relay to allow you to take that brick out of service.  Also normally closed switches could be added into the interlock circuit to manually kill the blocks.  These could be red mushroom buttons - they would need to be the type that latches where you twist the button to release it.

A small 24 volt ac power source is needed to drive the control relays.  All relays are 24 V ac, assuming the 18 V from the bricks will work for these.  Only Normally Open contacts are used on the relays.

Cam

I was thinking sort of along those lines in the earlier threads Stan linked to, only I think I misunderstood the complete shutdown process Clem was wanting at the time. My thought with the working panel model in the picture was that it didn't seem to have enough relays? I am obviously still missing something here?  

Maybe Clem's electronic friend will see this thread, have mercy on me and post a schematic for us...

Not quite clear why 12 relays are used in the photo.  As Cam suggests, no doubt there's more going on than what's been posted.

Anyway, I like Cam's approach (8 relays).  Should meet the requirements that have been so-far described.

Here's a similar approach using 12V DC (instead of 24V AC) as the "control" voltage.  I don't suggest this unless you're a determined DIY'er since it requires working with and soldering small components.  But it can be done for about $10 out-of-pocket and only uses 4 relays.  Again, this is posted in the spirit of kicking around ideas being that OGR is a discussion forum...

This uses low-cost solid-state optocouplers (vs. electromechanical relays) to detect the presence of brick voltage.  The 4 optocoupler transistor outputs are connected in series so all 4 optocouplers must be triggered to present ~12V DC to the 4-channel relay module.  This relay module is about $1/relay and the relays are 10 Amps.  These are single-pole relays so this assumes the outer-rails of all the blocks are in common; Cam's diagram uses double-pole relays which switch both center and outer-rails.  As Cam suggests, you can bypass/over-ride one or more bricks by placing a toggle switch at the output of the respective optocoupler.

Has anyone thought of a way to sense when a car derails without tripping a breaker in the block?

I started with the premise of controlling both sides of each brick and not making any other connection.  This is the most flexible but depending on how common is handled SPDT relays could be used to break only the hot side.  The control relays must handle the full current output of the bricks.  The sense relays only need to handle the combined current of the control relay coils.  The sense relay coils are driven by the 18V brick outputs.  The control relay coils could be any voltage, even DC.  It would be possible to avoid the extra power supply by using one of the bricks.

Some time ago I sketched out a design to enable track power only if the TMCC signal is present.  The command base was powered first.  After a delay a relay turned on 120V to the bricks and transformers if the command signal was present.   Then if the command signal dropped out the power was shut down.  It used an Amperite time delay relay tube - I inherited a bunch of these in various voltages and delay times from a friend.

CJACK - Detecting a derailment that doesn't trip the breaker is a challenge.  How to differentiate  this from a normal current load.  It might have to be a dynamic check looking for a sudden change in current?  Maybe use a video camera with derailment recognition software.

PS when I asked this same question a few years ago on this forum I was told it can't be done. 

Clem 

Whats that quote, "When THEY say it's impossible or can't be done, someone in their garage is working on the solution"  Or something similar to that effect....

It's pretty easy to detect the command signal, the TMCC Buffer has a detection and monitoring circuit that shows signal strength.  You could just duplicate that capability and add a relay to trigger the shutdown.

penn station posted:

 

CJACK - Detecting a derailment that doesn't trip the breaker is a challenge.  How to differentiate  this from a normal current load.  It might have to be a dynamic check looking for a sudden change in current?  Maybe use a video camera with derailment recognition software.

I was thinking maybe a short piece of rail outside each of the two outside rails for the wheel that's off the outside rail to bump against and trigger the shutdown...

John - I was thinking of hacking an R2LC.  I could either just look for the 455kHz signal or do something with the detected command data.  It would be connected directly to the base as I just wanted to see if the base was up and running.  I am impressed with all your hard work on the buffer.  Dale was a genius and it's hard to believe he is gone.

At one time I proposed making a TMCC receiver using sub-miniature tubes.  I have loads of these in my vacuum tube collection.  There is no valid reason to do this except, why not?

CJACK - that would work in some scenarios.  Inside rails would probably work too.  Perhaps use an accessory track.

I just found Dale's response regarding a tube based TMCC receiver.  We were corresponding about the TMCC signal data encoding. This was from 2004.

You must be crazy!!  Glass bottles in TMCC?  Is that to get the full 'toob'
sound that so many of the recording studio people talk about?  Honestly, it
sounds like fun, but WHY??  Then again, you could try Nuvistors....

penn station posted:

I just found Dale's response regarding a tube based TMCC receiver.  We were corresponding about the TMCC signal data encoding. This was from 2004.

You must be crazy!!  Glass bottles in TMCC?  Is that to get the full 'toob'
sound that so many of the recording studio people talk about?  Honestly, it
sounds like fun, but WHY??  Then again, you could try Nuvistors....

Ah yes RCA Nuvistors, those tiny tubes in a tin can.  Remember them well, especially since old tube testers didn't have those sockets so you needed adapters that plugged into a checker socket.  Then there were the 10 (or was it 12?) pin glass bottles used in portable TVs...

Now I am showing  my age.

Anybody remember the Raytheon CK722 Germanium transistor...?

cjack posted:

Anybody remember the Raytheon CK722 Germanium transistor...?

Me, Me, ME!!!

My dad managed to locate one of them and gave it as a present  (cost something like $4 in 1950s money).  I then built a one stage audio amplifier attached to my Quaker Oats cylinder coil-cat whiskered crystal radio (later upgraded to a 1N34A germanium diode) to drive headphones.  I then found plans to put in a transformer to drive a small speaker but the results were so so compared to  headphones.  I still have that little square can transistor somewhere as a memento to my electronics youth.  Good old days.

cjack posted:

Anybody remember the Raytheon CK722 Germanium transistor...?

My very first transistor, I bought one mail-order, as I recall it was pretty expensive, something like $6-7, a kings ransom in the 1950's!  I was super careful with that part and built a few different circuits all the while following all the handling precautions.  I finally broke a lead off right at the base and killed it, but by then more transistors were available at lower prices.

Rick R posted:

…. I would like something like Clem has

The thread has meandered somewhat.  Did you get the info you need? 

Seems to me if you want what Clem K has, I'd think the individual who built it for him kept some notes and might be available to build another.

The first transistors I had were similar to GE 2N107.  I now have two GE 2N107 and three Raytheon CK722 transistors in my collection.  They come up on eBay occasionally.  The blue CK722s are highly sought after.

I started with tube circuits.  I built a Geiger counter using the Raytheon CK1026 Geiger-Muller tube, and a capacity operated relay using the Miller 695 coil.  The CK1026 tubes are very hard to find.  The 695 coils are near impossible.  I've got several of each in my tube collection.  Thank you eBay.

My first job out of college was with Raytheon, new CK series parts were still being released.

Ah yes RCA Nuvistors, those tiny tubes in a tin can!

yes I remember them they were used in the tv tuner when there was a mechanical tuner for television! that sure brings back memories!

Alan

The relay board in Clem's picture is very neatly laid out.  The power flows from bottom to top with four similar circuits.  A close up of the relays in each of the three rows and a shot of the back would probably be sufficient to work out the circuit.  My first question is what to the toggle switches do?

Alan Mancus posted:

Ah yes RCA Nuvistors, those tiny tubes in a tin can!

yes I remember them they were used in the tv tuner when there was a mechanical tuner for television! that sure brings back memories!

Alan

Tektronix used them in the input of the 454 oscilloscope. They drifted for about a half hour causing me to have to chase the DC vertical position of the trace. The 454A used fets and was much more stable.

penn station posted:

The relay board in Clem's picture is very neatly laid out.  The power flows from bottom to top with four similar circuits.  A close up of the relays in each of the three rows and a shot of the back would probably be sufficient to work out the circuit.  My first question is what to the toggle switches do?

Toggles turn the system off

clem k posted:

penn station posted:

The relay board in Clem's picture is very neatly laid out.  The power flows from bottom to top with four similar circuits.  A close up of the relays in each of the three rows and a shot of the back would probably be sufficient to work out the circuit.  My first question is what to the toggle switches do?

Toggles turn the system off

There are 4 switches at the bottom of photo.

So to be clear, does each switch turn off (i.e., force OFF) 1 of the 4 blocks irrespective of the shutdown feature of your system?

And those are the ONLY switch/configuration controls to your "Safety Relay" system?

We're trying to help the OP here!  In absence of diagrams/schematics/parts-lists from your custom built system since the OP wants what you have, I think we can get there if you let us know exactly what your system does.  In my opinion that is...

What's the p/n for the optocoupler?

Just about any AC-input optocoupler will work.  I just checked on eBay and for about $1 (free shipping from Asia), you can get 5 single-channel PS2505-1 optocouplers.  You'd use 4 of them.  Note that these devices come in 1, 2, and 4 channel packages and can be "stacked" so that 4 1-channel devices looks like 1 4-channel if that makes sense.

The PS2505-1, PS2505-2, and PS2505-4 are shown above from DigiKey though minimum shipping from DigiKey is over $4 these days.

As I type this out, I now recall and dug up a previous discussion about AC-input optocouplers:

https://ogrforum.com/...89#69285198810032689

So I see there I recommended if only going to "stock" one optocoupler type in your DIY parts stash, use the AC-input LTV-8141.  This part would work too.  And it too comes in 1, 2, and 4 channel versions though their numbering scheme is LTV-8141, LTV-8241, and LTV-8441 respectively.

--------------

But as I stated earlier, you have to be a determined DIY'er.  Some comments that may help or confuse depending on your comfort zone.  I measured the input trigger current to the 4-channel relay module.  Each trigger measured about 1 mA with 12V applied.  So it takes 4 mA to trigger all 4 relays.

Each optocoupler is essentially acting as a bridge rectifier to convert AC to DC.  Drive each optocoupler at, say, 10 mA.  In other words an input resistor of 1K-5K.  You need a capacitor to smooth the full-wave rectified output otherwise the relays might chatter/buzz at the line frequency.  Something like 1uF - 10uF should work. 

If someone/anyone is actually contemplating wiring one up I can peel additional layers of the onion; it may bring tears to your eyes .

If it had universal appeal, it might even suddenly pop up with a circuit board design.

Here is a possible circuit explaining the 12 relays.  From the photo each column covers one power channel.  The brick power comes in at the bottom then flows through the switch and three relays in series.  All the coil wires sneak through holes to the back.  Clearly the magic happens on the other side of the board. My surmise is the relays are driven by the three "other" bricks in each channel.

I am still trying to sort out the terminal strip at the top.  Also the red wire on the right side may mean the relay coils are connected to the track sections instead of locally on the back of the board.

Cam

I agree that explains how 12 relays could be deployed...but I don't see how it performs any different/better than your 8-relay method.  The force-OFF switch in your method could be placed on the output relays (red X) and these switches would not need to be heavy-duty 10-Amp (or whatever) switches.

Plus, with the presumed 12-relay method, the nominal condition is all 12-relays are "active".  And for heavy-duty relays which might draw 1 Watt per coil... that's 12 Watts of overhead.  OK, 12 Watts does not break the bank.  But all 12 relays carry the full load current per block...whereas with your 8-relay method only half of the relays carry full block power so lower capacity relays can be used for the sense function so less expensive relays could be used.

In the absence of further information, I vote for your 8-relay method which AFAIK meets the known requirements.

OK here are more photos, the toggles take that particular block off line, four toggles makes the safety feature total off. some of the fine wires on top of board go to micro pilot lights to tell which block tripped the circuit. the digital readouts for each block are...." I forgot".

This is hard to photograph. holding a flashlight ,camera and the board in position.and I don't like taking photographs because they always turnout poor.

Hope this helps the cause

Stan - in my diagram the control relays need high current contacts.  All the other relays and switches including the ones you added only need to handle to the current to the control relay coils.  This was a quick sketch,  I would probably select an lower coil voltage for the control relays like 12VAC, or perhaps DC - with the relay supply selected accordingly.

I agree with your comment on the power consumed by the always-on relay coils in both designs.

Clem - thanks for the additional photos.  I love reverse engineering.

So for the record, could you confirm that all 12 relays are marked "LY2N-J" as it appears to be above?

And, if so, could you read the marking on the coil wrapping.  For example, in above the coil is repeatedly marked "24 VAC".  This tells us the relay coil voltage...which for the LY2N-J type relay might be 12V DC, 12V AC, 24V DC, 24V AC, etc.

clem k posted:

... the digital readouts for each block are...." I forgot". 

Yes, this is a head scratcher.    Those digital readouts are part of a DC voltage regulator module.  It's not clear to me why DC voltage is being used - much less 4 separate DC voltages around 9V DC!  Makes me think the relay coils are operating at 9V DC.  Interesting!

Thx everybody giving me all the great information

I have to think this is killing gnats with a sledgehammer.  Am I missing something?  Why not just remove the 120V power from the layout with a trip, why all the high current relays?  One big control relay on the input power and then any sensing you desire for the power districts.

stan2004 posted:

I agree that explains how 12 relays could be deployed...but I don't see how it performs any different/better than your 8-relay method.  The force-OFF switch in your method could be placed on the output relays (red X) and these switches would not need to be heavy-duty 10-Amp (or whatever) switches.

Plus, with the presumed 12-relay method, the nominal condition is all 12-relays are "active".  And for heavy-duty relays which might draw 1 Watt per coil... that's 12 Watts of overhead.  OK, 12 Watts does not break the bank.  But all 12 relays carry the full load current per block...whereas with your 8-relay method only half of the relays carry full block power so lower capacity relays can be used for the sense function so less expensive relays could be used.

In the absence of further information, I vote for your 8-relay method which AFAIK meets the known requirements.

 

A few random thoughts-

Why couldn't you reverse the logic of the whole thing; use the NC contacts on the control relays and put the sense relay contacts in parallel? That way the control relays would normally be de-energized, no heating and no power consumption. Yes, if the control supply fails you would have no protection.

I'm guessing that the DC supplies were intended to permit the use of 12 VDC sense relays, but the filter caps in the supplies caused the sense relays to hold in for a while after the power was lost, so the 8.x volt setting was chosen to put the relays right on the edge of dropout to speed up the response to a power loss.

Because there are no inexpensive 18 VAC relays, the sense relay function is going to require a bit of fudge. I think the original design is a good attempt to use plain old stock parts.

I understand that the builder of the original unit may well have used what he had on hand. No criticism of the original design is intended.

Just talking........

gunrunnerjohn posted:

I have to think this is killing gnats with a sledgehammer.  Am I missing something?  Why not just remove the 120V power from the layout with a trip, why all the high current relays?  One big control relay on the input power and then any sensing you desire for the power districts.

That's a good thought too.