Yes, the traditional way to handle reverse loops on DC is to have a separate reverse switch for the main, and a separate one for each loop.
The loop reverse switch must be set to the correct polarity for the train to enter the loop.
Once in the loop, the loop reverse switch controls the direction should you need to back up, as long as you stay in the loop.
While in the loop, the polarity of the main needs to be reversed and the turnout aligned for the train to re-enter the main in the oposite direction.
But there are other ways to handle DC reverse loops, too complex to explain without drawings of examples…
On my new layout using the Aristo wireless throttles there will be a wye and a loop that can reverse trains. Both require the train to come to a stop and for the operator to throw a turnout and change the direction on his throttle before proceeding.
Because the direction controls on the throttle relate to east and west movement on the layout, not forward and reverse movement of the locomotive.
The decoder alwyas tells the loco which direction to move. Which rail is connected to which terminal on the DCC booster makes no differents. Assuming the motor wires in the loco are correctly connected to the decoder, forward is forward, aka, to the front of the loco, and reverse is reverse.
With two rails to carry current, they must always be at opposite polairy. The square wave AC nature of DCC does not change that - connect the two terminals on your house wiring, which is AC, and you get a short just as surely as connecting the two terminals of a battery, or the two DC terminals of a DC power pack causes a short.
With DC< you typically reverse the direction of the main line while the loco is in the loop. this makes the polarity match on the exit route of the turnout, so no short, and the polarity is such that the loco continues moving forward. If ther were another loco on the main track at the same time, it would suddenly change direction.
With DCC, the square wave AC is only carrying data addressed to each decoder which tells it which direction to go, how fast, and what functions (like the lights, horn, or bell) are on. The ‘polarity’ of this signal carries no direction information. So to handle a reverse loop, with DCC you typically change the polarity under the moving loco. Since this has no effect on direction, the loco keeps right on moving just as it was before the polarity flip. But now the polarity across the gaps (DCC reverse sections still need gaps in the rails, just like DC) match between the main and the exit route the loco is taking, and it can continue on as if nothing special happened.
It is perfectly possible to flip the mainline polarity with DCC, same as with DC, except that any other trains on the main with DCC would just keep right on going the same way they were, no sudden direction change. The main reason this is not usually done is that the automatic reversing units will last a lot longer switching the current o
Randy, thanks for bringing up that issue. It is an interesting concept, and one that I have not really thought about, and I am sure that it is true for others as well.
the short circuit current doesn’t depend on the number of locos drawing track current. the metal wheels of rolling stock will cause just as much of a short as the metal wheels of a loco
when not switching, more current wouldl pass thru the auto reverser if there are multiple locos and lighted cars on the mainline track. But why should any electronic circuit fail if it’s operated within it’s limits. Is your 10A booster more likely to fail if it supplies 9A much of the time instead of 6A?
if you have multiple reversing sections, aren’t you limited to having a single train enter/exiting each reversing section if a single auto-reverser is used for both reversing sections or just the mainline track?
if you have multiple power districts, would you need separate auto-reversers for the mainline track in each distrcit?
I’m saying the reason to switch the loop in DCC, when you can just as easily switch the main automatically, is to keep the reverser from having to switch under a high load. Relay or solid state, there’s a limit to the switching current, usually much less than steady state current.
If you flip the reverse section, you have one train in it. Whereas if you flip the main, you might have 6 trains out on the main, with one going through a reversing section. Flipping either will accomplish the same thing, but flipping the reverse section means switching a lot less current (potentially) than the main.
Anyone with a PSX-AR want to read the part number off the switching MOSFETs on one, so we can find the data sheet and see what the switching current is?
FQ, you are getting a lot of technical talk thrown your way, so it is not surprising that you are confused about the issue of reverse polarity and the function of an auto-reverser.
We don’t know much about your layout, such as the track plan, the types and lengths of trains you are running, metal wheels versus plastic wheels, the brand of auto-reverser, the types of turnouts, etc. More information would help to give you more specific advice.
But, let me give you an example.
Say that your track plan includes a mainline with a loop of track that folds back onto the mainline via a turnout. Let’s further assume that the straight through route of the turnout heads directly into the loop and that the feeder wires inside the loop match the polarity of the feeder wires on the straight through route of the turnout. That means that when the train exits the loop via the divergent route of the turnout, there will be a short because the polarities will be mismatched.
The solution is to gap both ends of the turnout (opposite the tail end of the turnout) and route the feeder wires inside the loop to the output side of the auto-reverser. The result will be that a train entering the loop from the straight through route of the turnout will not trigger the auto-reverser since the polarities already match. But when the train exits the loop via the divergent route of the turnout, the auto-reverser will flip the polarity inside the loop to match the polarity on the divergent side of the turnout.
If you think about it that way, the next time that the train enters the loop from the straight through route of the turnout, the auto-reverser will flip the polarity inside the loop to match the polarity of the straight through route. On the other hand, if the train were to enter the loop through the divergent route of the turnout, the auto-re
With DC reversing - the main line polarity is changed, agreed?
With DCC reversing - you can change either the main or the reverse loop, as long as ONE of the two, and only one, gets flipped, agreed?
Now, our club layout has no reverse sections because it is basically a giant oval donut, but typically there may be 10 trains running. Say we did have a reversing section. Since we run DCC only, an autoreverser would most likely be used. If one train goes into the reverse loop, and the autoreverser flips the phase in the loop section, it’s switching the load of one train. The other 9 don’t see so much as a microsecond of interruption.
If instead the autoreverser was connected to the main, it would flip the phase as the train on the reverse loop crossed the gaps, same as before. Only this time, it’s switching phase with the load of the 9 trains on the main.
The mosfets in a circuit breaker, frog juicer or auto-reverser are constantly turning on and off each time the DCC signal changes phase.
i assume even an auto-reverser using a relay can act as a circuit breaker, blocking power to the reversing section if there is a short with both polarities. I further assume it blocks power when toggling the relay to avoid contact arcing (which avoids the problem you’re suggesting)
the schematic below shows circuit breaker design using pairs of mosfets to conduct the alternating DCC current to each rail. the mosfets are connected common drain. A mosfet conducts current in a single direction. There is an inherent diode in the mosfet that conducts when the voltage across the mosfet drain/source is reversed. the opto provides a bias voltage to turn on one mosfet or the other as the voltage alternates. the mosfets are constantly turning on/off with each DCC phase.
the circuit detects a short when the current thru the small 0.22 Ohm resistor is close to 0.7V, causing the bipolar transistor to conduct.
an auto reverser has another two pairs of mosfets (8 total) connecting the rails to the opposite inputs. The auto-reverser biases one set, 4 mosfets for each polarity, or biases none to act as a circuit breaker.
when an auto-reverser “reverses” polarity to the reversing section rails, it simply biases the other set of mosfets. there is no “switching current” that occurs when reversing polarity. the mosfets must be capable of handling the short circuit current.
I’m still trying to figure out how that circuit can even work, you have two MOSFETs in series, on each leg of the DCC, with the gates controlled by the same source. So both will turn on and off at the same time, and since they are in series, both will pass current only one way.
i believe these are p-chan mosfets with their drains connected. source-to-drain, drain-to-source. when properly biased, only one mosfet is turned on during each phase of DCC and current flows thru the body diode of the other mosfet.
The parts list at the bottom says all 4 are the same N-channel part. Perhaps that is an error. If they were opposite types, it would work like you say, though selection of the proper part is kind of important than. My servo controller uses a MOSFET as the reverse polarity protection on the DC input. Less loss than even a schottky diode - with the input polarity correct, the MOSFET is turned on and with a really low Rds, even if the circuit draws a couple of amps (which it doesn’t), the loss is less than the forward voltage of any diode.
NCEs new AR10 auto-reverser illustrating 4 pairs of mosfets used to connect the output terminals to either input terminal and not a whole lot more (see DCC specialties Frog-AR)
i see an axial leaded 100 (??) (brown black brown gold) resistor set above the board for better cooling presumably used for short detection, but think even 1 Ohm might be too high for that purpose.
the orange things on left look like pulse transformers often used for detection. maybe they detect a short based on the amplitude of the coil voltage.
