Automatic Reversing Loop - Will this low cost method work?

Discussion in 'FAQs' started by BigJim, Nov 29, 2006.

  1. BigJim

    BigJim Member

    I want to have an automatic turnout & polarity switching of two hidden reversing loops. I don't care which direction the train goes when it enters the loop. The layout will use DCC.

    I know there are some commercial items for this purpose but they seem complicated, expensive, require more "cuts" or sections and besides the DIY part of this hobby is what I am looking forward to.

    N = North rail of main line
    S = South rail
    O = Outside Rail of loop
    I = Inside rail
    S1 & S2 are Photosensor locations (~$1)
    T = Tortoise Switch motor

    Two CD power sources - Both with S as common - CD1 = South to Inside & CD2 = South to Outside. Both can be very slow charge off of the DCC power as they will only be needed once per full round trip. (One capacitor, one diode & a resistor each ~ $4)

    One DPDT two coil latching relay. ($3.50 on E-Bay)

    Two SCR's (about $1.50)

    I know "polarity" is not the right term for the AC used by DCC but the concept is easier to describe this way. Assume the main line polarity is fixed east bound.

    When the loop polarity is clockwise CD1 will charge. (Outside rail opposite of South rail)
    When the loop polarity is counter clockwise CD2 will charge. (Inside rail opposite of South rail)

    I should be fairly easy and cheap to create a circuit to "fire" an SCR once on a light to dark transistion using a Cadmium-Sulfide photoresistor (CS), one transistor, one resistor and an SCR. (Assuming a 12K/300K CS detector) Transistor & SCR connected "darlington" style, CS placed emitter to ground, about a 1M resistor up to the positive and the relay coil between the positive and the SCR. If I can find a very low current SCR I might not even need the transistor:) When the CD discharges through the relay the pulse will change the relay position and the SCR will drop out when the CD discharges. Not sure but, depending on the holding current, I might need a resistor across the relay coil to have enough current to fire the SCR.

    The relay output is used to switch the polarity of the power to the switch motor.
    The switch motor DPDT contacts are used to match the "polarity" of the loop to the direction of the turnout.

    Anytime a train crosses a sensor indicating it is near one of the legs of the wye AND the CD is charged (indicating a change is needed) the SCR will fire to pulse the proper coil of the latching relay. The latching relay will change the polarity of the switch engine. The switch engine will then change the direction of the wye and the "polarity" of the loop to match the correct side of the wye.

    With a latching relay and loop polarity controlled by the switch motor there should never be any mismatched states on power-up. Since the CD's are only powered when there is a mis-match with the associated sensor it will only fire once even if the light between the cars pulses the CS.

    The only potential problem I see is the proper placement of the sensors relative to the cuts to make sure the entire process is complete before the train gets to the cut at the highest speed and having the loop long enough to make sure the longest train is shorter than the distance between the sensors.

    I don't care which way the train goes around the loop on entry but I guess you could use a push button across the SCR to manually change the direction of the wye.

    Do the DPDT contacts on the Tortoise Switch motor switch fast enough to not have a bad hesitation if used to switch the loop polarity?

    Has anybody tried this method or something similar?

    Anyone see any reason why it won't work before I buy the parts?

  2. Russ Bellinis

    Russ Bellinis Active Member

    Actually polarity is the correct term, in dcc the carrier signal that carries the instructions to the train locomotive or whatever else you are operating on dcc is high frequency ac, but the trains are still powered by dc. When you go into the reverse loop, your reversing device needs to maintain the same polarity in the loop as the mainline the train came off of in order to keep the train from suddenly trying to reverse as it crosses the gap. Once the train is fully in the loop between both gaps, an automatic reverser will throw the switch to allow the train to come out. I think your circuit will do this. Someone who knows more about the workings of dcc will have to chime in here. Does the decoder instantly recognise the different polarity of the mainline when the train comes out and correct accordingly to keep the train going forward? How does it avoid having a short circuit as one set of wheels is on the loop with one polarity and the other is on the mainline with opposite polarity?
  3. EngineerKyle

    EngineerKyle Member

    Over my head, Big Jim...

    I will say an MRC autoreversing loop module will do that SO easy for 25 bucks....

    It's all you need. Two wires to the loop, the other two to the mainline... and everthing is Automatic!
  4. 60103

    60103 Pooh Bah

    Bigjim: I suggest that you use a switch machine with reversing switch built in. This can be used to flip the reversing loop polarity when the turnout is changed. If it causes an unwelcome hesitaion in the train movement, you haven't lost anything when you replace it with the DCC reversing module.
    The revesing module works on the basis of sensing the momentary short circuit faster than the master control does and flipping the polarity.
    The decoder chip in the loco is indifferent to track polarity -- remember that you can turn a loco around on the track and it still goes "forward". (I don't know how it works, but when I learned computers they had just abandoned relays for arrays of magnetic rings.)
  5. BigJim

    BigJim Member


    I was planning on using the switch motor to change the track "polarity". That part is easy and comes free with a DPDT equiped switch machine like the Tortoise. Changing the direction of the wye and the polarity of the track before the gets to the "cut" with the minimum amount of electronics is what I am trying to accomplish.

    I really want to keep it "ohm meter" simple if I can.

    BTW I also go back close to that era. In the 60's I worked at a "core house" where we pressed, fired and tested those little magnetic rings called "computer cores". One bit of information per physical core or about 9,000 cores for 1K of memory (8+1 per byte). Probably never known by a lot of people but that is why the main memory on computers is often referred to as "core memory" - not because it is in the core of the machine but because it was made out of magnetic rings.

  6. BigJim

    BigJim Member


    According to what I found about the MRC item it "makes the operation of reverse loops transparent to the operator, they require no intervention other than aligning turnouts."

    I also want to align the turnout when a train approaches the back side and it is in the wrong direction. I think the Hare will do this but not sure about he polarity issues as the drawings I have seen do not show any power reversing connections to the loop.

    However, the very short section used by the hare for detection might work easier than the CS dectector to "fire" the SCR. I will look at that. Could be even easier.

  7. EngineerKyle

    EngineerKyle Member

    Look into it big Jim... tell us what you find out.

    You might have a great idea there....
  8. 60103

    60103 Pooh Bah

    Jim: the Tortoise has a DPDT switch built in. I trust adding the hare to it doesn't do anything to them. The only problem I would worry about there is the time that it takes the slo-mo machines to change over.
    Have you seen the December Model Railroader? page 94 covers a lot of what we're talking about.
  9. shortliner

    shortliner Member

    Big Jim - I think that this may be what you want. I should tell you that the SOLID turnout DOES work - but it IS weird to watch it do so - almost like watching a magician do a magic trick :D HIH
    Shortliner(Jack)away up here in the Highlands
  10. BigJim

    BigJim Member

    I think I could use the Hare and swith the tracks with the Tortoise but since I want to have the train cause the switch a DCC control seems like a waste. I did some more looking and I think a wire in the middle of a double cut in one side of the track (loop around the cuts for power) or even a wisker wire to the inside of the wheel could trigger the SCR directly.

    As above the Tortoise would switch the polarity of the loop to match the switch position.
    Relay - DPDT magnetic latching relay - singlecoil version shown.
    For dual coil versions one coil goes to each SCR. From what I could determine the dual coil versions are still polarity sensitive with a common connection. Many have two diods internal for use with AC coil current.

    R1 sets the CD charge time. Could be quite large (1M ?) as it only has to fire once per round trip.
    R2 is very small - 10 to 47 ohms. Only used as protection in case both SCRs fire at once. Not needed for double coil relays. Probably only needed on one side.
    R3 just keeps the gate from floating.4.7K > 10K should be fine
    R4 should not be more than 3 times the size of R3 so the gate voltage reaches at least 1/4 the CD voltage.
    Capacitors size is determined by the amount of power needed by the relay. I would guess that most small relays would need less than the average coil switch so anything over 500 uf should be fine.
    Larger holding current SCRs might need some additional load across a high resistance relay coil to make sure the holding current level is reached.

    Since the SCR fires very quickly I would think a small wisker wire mounted on the inside of the track would also work but I don't have any experience in this area or know how reliable it would be. I guess I will try it first rather than cutting the track.

    60MPH = 88 fps or about 1 fps in HO. I don't think I will try and run this speed in the reversing loops but still would only require about 3 feet from the sensor to the switch for a 3 second delay.

    I will let you know if it works.

  11. baldwinjl

    baldwinjl Member

    Maybe I am missing something, but I think there is overthinking going on here. The only way a short is going to happen in this type of reverse loop is when you are on the turnout. And with the tortoise you know which way the turnout is thrown. So gap the loop close to the turnout, and use the Tortoise to match the loop to the selected side of the turnout. There is no need to detect anything else, no need for any relays, or any more confusion. You throw the turnout, polarity matches, enter the loop, throw the turnout, polarity changes (remember, with DCC the loco direction doesn't change), you leave the loop.

    Forget this whole thing, I wasn't seeing that you wanted to throw the turnout automatically as well. Too much skimming, not enough reading........
  12. BigJim

    BigJim Member

    Sorry, but I do want to detect the train and switch the turnout. My reversing loops are going to be in a hidden area so I just want the train to leave the main room and come back without intervention.

    From what I I have read so far DCC is a square wave with matching positive & negative pulses and the length of each pair determining a "1" or a "0". The signal is also used to provide power with adjustment times on the last "0" top or bottom to balance the DC value to the desired level.

    The signal is the power - the power is the signal.

    Think of being miles away from me with just a single pair of wire connecting us and I have to send you power and signal "1"s & "0"s to you. I have a power sources that I can switch back and forth between 50hz and 60hz each cycle (same voltage). All of your power equipment works fine on either 50hz or 60hz. If I send one 50hz cycle you read the time between peaks and see "0". I send one 60hz cycle and you see "1". All of your AC equipment continues to run fine during either.

    The DCC signal is made up of a string of single cycle square waves of 58 or 116 microsec in duration (~8.6 Khz & 17.2 Khz). The short ones are "1"s & the long ones are "0"s, If I read it corectly this is where DCC departs from the computer technology I am familar with. It takes 1/2 the time to transmit "1"s as it does to transmit "0"s. Computer disk drives record data in a similar fashion but since they don't have to worry about the DC component and need a constant data rate they use either one full cycle of the short pulse or 1/2 cycle of the long pulse to keep the "1"s & "0"s the same length.

    DCC looks more like morse code. dit-dit-dit (111) takes less time than dash-dash-dash (000). Since we are not sending that much information and even the dashes are only 116 microseconds long there is pleanty of bandwidth.

    DCC can also "stretch" the top or bottom of the last "0" pulse up to about 10 miliseconds to add a DC component to run one DC train at the same time. The DC train might "sing" when it receives the signal but DC motors don't turn with the AC part and their natural inductance limits the amount of power they draw with a AC signal between 8.6 khz and 16 khz.

    The "SOLID" would just clip 1/2 of the DCC signal and add the voltage drop of two diods. I think DCC trains might not stop although they probably will slow down - not sure.

    I will try my circuit and let everyone know how it works. As I said before I like things "Ohm Meter simple". I even prefer analog Ohm Meters:) . Anybody have an old Simpson 260 for sale?

  13. baldwinjl

    baldwinjl Member

    OK, now I have done a little more thinking. But not too much! Because the contact in the Tortoise don't throw instantly, you will need a relay or an autoreverser to make polarity happy. What would be handy would be an autoreverser that told you when it was thrown and which way. You could put the gaps far enough ahead of the switch to detect the train and throw the turnout. I wonder if such a thing exists? (I think that is not far from what you are designing, when you get down to it.)

    The Digitrax PM42 manual claims that when used as an autoreverser there is an led that is on or off depending on the polarity, so that combined with knowledge of the switch position from the Tortoise would give you enough to know whether the turnout needed to be thrown or not (might need a flip-flop to keep track of things, but I haven't thought that much, yet.)

  14. steamhead

    steamhead Active Member


    There's an article in this month's (December) Model Railroader with a circuit that will do exactly what you're looking to do.

    Good luck...!!!
  15. cidchase

    cidchase Active Member

    Hi Jim,
    I think what I'm reading from your sequence of operation is that you want the polarity of the loop to change automatically while the train is in the loop. Prolly not what you mean to do.

    If i'm reading you right, what you really want to do is change the polarity of the mainline when the train on the loop approaches the turnout. It sounds very doable with the detector and relay like you describe.

    When you enter the loop and pulse a detector, the relay and turnout are already set to that side. When you continue around the loop to the second detector, the relay and turnout will switch, and the polarity of the mainline will be set to allow the train to return in the opposite direction from whence it came.

    Maybe that's just what you already said. If so, just ignore me. :D :D What this setup implies is that the loop can be connected direct to the power pack, and the loop direction corresponds to the power pack polarity. The main line polarity (and direction) are set by the turnout position.

    If you use the same setup on the other end of the main, then both turnouts need to be switched simultaneously, or you'll get a short. :curse:

    Am I all wet, or what???:D :D The only thing I don't like is the whisker switch. How about an optical detector or a magnetic reed switch responding to a magnet on the loco?
  16. BigJim

    BigJim Member

    I do want to change the polarity of the loop while the train is in the loop. Since I am running DCC this should not be problem. Main line and the rest of the layout will never change. I do need to change the polarity of the end of the loop to match the wye. It is not a problem to change the polarity of a DCC engine while it is on the track but you can't have reverse polarity at a "cut" where the wheel sets could be on both sides of the cut at the same time or a single wheel could touch both sides.

  17. cidchase

    cidchase Active Member

    I don't know why I can't read, sometimes! Had it in my feeble mind you were DC.:oops:

    :curse: on these Wal-Mart reading glasses, anyhow:D :D
  18. Russ Bellinis

    Russ Bellinis Active Member

    I did not go back and read the entire post, so this may be something that has been suggested already. I think if my memory is correct, you mentioned using a Tortoise machine to operate the switch in the reverse loop. Just wire a double throw double pole relay coil to the auxiliary terminals on the Tortoise so that in one position the coil is energised and the other is deenergised. Then wire the relay n.c. & n.o. points to flip the polarity when the Tortoise is operated.
  19. trainman4

    trainman4 Member

    auto reverse unit

    Check this out.I have eight of them in use they sell for $34.95 and can be had at your hobby shop or look up Pocono Mountain Lines.If i can install them anybody can.sign1
  20. BigJim

    BigJim Member

    Thanks for your input.

    I really want a cheap, simple, automatic solution. I don't need any DCC control. I don't care which way the train enters the loop since you won't be able to see it. I am using DCC so I can switch the "polarity" of the loop (so the exit matches the main line) while the train is in the loop and don't have to worry about the rest of the track. I don't want

    Looking at it a little more I came up with the following using a CMOS RS Flip-Flop to "latch" the position on the first sensor pulse and four transistors in a "voltage following" configuration to buffer the outputs of the flip-flop yo drive the Tortoise. Since the loop will only have one train at a time the DPDT switch in the Tortoise should be fine to switch the loop. If it was visible I might have the Tortoise drive a relay to avoid the dead spot and hesitation of the slide switch in the Tortoise.

    With the 3K > 300K light dark spec for the photo-cell I think 33K should work fine but I will experiment when the parts get here next week.
    Total Cost:

    Two CS Photocells @ 0.95 = $1.90
    RS flip-flop (1/2 of CD4013) @ 0.25 = $0.25
    2- PNP & 2 NPN power transistors (BD135 &136) @ 0.20 = $0.80
    2-33K Ohm resistors @ 0.10 = $0.20
    (Prices from

    Total cost = $3.15

    Assuming the $34.95 solution would still need the photocells the cost of this part is $1.25. - My kind of solution.

Share This Page