Technical Corner

DCC GROUND FAULT CIRCUIT INTERRUPTER

Many of us have Ground Fault Circuit Interrupter (GFCI) devices in our homes and workplaces.  Bathrooms, kitchens, outdoors, garages, and wet bar sinks, laundry sinks, and utility sinks, Crawl spaces and unfinished basements must have GFCIs, according to The National Electric Code.   A ground-fault circuit interrupter, or GFCI, is a fast-acting circuit breaker designed to shut off electric power in the event of a ‘ground fault.’  A GFCI compares the amount of current going to and from the connected equipment.  When the amount going differs from the amount returning by a predetermined amount, a ‘ground fault’ is said to exist and the GFCI device quickly turns off the power.  

DCC boosters and command stations that include a booster can be thought of somewhat like power outlets.  A booster has two output terminals and the amount of current going to and from the connected layout should be equal in them, but of opposite direction.  When the amount of DCC booster current going to the layout differs from the amount returning from the layout a ‘DCC ground fault’ could be said to exist.  But how could that be?  What goes out of one terminal must go in the other!  A DCC booster actually has three terminals; the third is the zero volt side of the booster’s power supply and is “booster common.”  A DCC ground fault could be said to exist if part of the current from one of the booster output terminals somehow finds its way back to the booster common terminal.  To see how a DCC ground fault might come about it is necessary to examine how boosters are connected to a model railroad layout. 

SINGLE BOOSTER LAYOUT

The drawing titled ‘SINGLE BOOSTER LAYOUT’ includes a simplified diagram of a booster showing only a power supply and a diagram of the four transistor ‘switches’ that make up what is called an “H” amplifier.  The transistor switches are controlled so that ‘A’ & ‘C’ are on or ‘B’ & ‘D’ are on.  If ‘A’ & ‘C’ are on, the Red rail is positive and the Black rail is zero volts.  When ‘B’ & ‘D’ are on, the Black rail is positive and the Red rail is zero volts.  The polarity of the voltage on the rails changes thousands of times per second, according to the data being sent to mobile and stationary decoders.  The decoders rectify the rapidly reversing voltage and use the DC power obtained to drive loco motors, lights, sound, and, for stationary decoders, signals and other stuff, according to the instructions in the DCC power bus.

The potential for wiring errors that might allow booster current to get back into the booster common is slight in a single booster layout.  The booster common connection is one or more of the wires in the throttle bus from the command station.  The first as well as the last rule of a DCC layout builder should be never connect anything to the cab/throttle bus or throttle panels unless absolutely directed to by the DCC system manufacturer’s instructions.  If this rule is followed, there is little chance for DCC ground faults in the single booster layout.  This is not to say that they cannot happen, even single booster layouts may include multiple DCC circuit breakers, block detection, signaling system, stationary decoders, computer interface, and more,  Each added bit of complexity increases the potential for an unintended connection, what is sometimes called a ‘sneak path.’  It is typically one of the unintended connections that otherwise causes no obvious symptom but may cause a DCC ground fault.  The most insidious ‘sneak path’ is a rail gap that opens and closes with the changes of season and relative humidity.  One symptom of a layout wiring problem is a booster that runs hot, even with no trains on the layout.  If a booster runs hot for no apparent reason, and comes back from the manufacturer with ‘no trouble found’ on the repair ticket, then one should suspect something is wrong with the layout wiring.

Multiple booster layouts, by virtue of their greater complexity, are far more likely to have problems of DCC ground faults.  Boosters are connected to the layout rails through a plethora of wires and other equipment.  The more complex the layout the more ancillary stuff there is and the greater the chance for an unintended connection or ‘sneak path.’

Multiple boosters are required when the layout size and/or the number of trains in operation exceeds the capability of a single booster.  Most problems occur at the electrical & mechanical boundaries between boosters.  If the wheels of any motive power or rolling stock spark at the rail gaps between power districts, there is something wrong.  If any motive power stutters, stalls, or otherwise hiccups at the rail gaps between power districts, there is something wrong.  There should be absolutely nothing about the operation of a train to indicate the crossing of a power district boundary.   

MULTIPLE BOOSTER LAYOUT CONFIGURATIONS

The most common configurations of multiple booster layouts are:

Common wired boosters with both rails gapped between districts.

Isolated boosters with common rail between districts.

Isolated boosters and no common rail.  

“Isolated boosters” means more than just boosters without their ‘booster common’ terminals connected together.  The boosters must be completely free of any metallic connection between one and another.  This means no electrical connection from booster to booster by way of the signal cable to and between the boosters.  The isolation is typically referred to as ‘Optical Isolation’ and means that the booster electronics is not connected to any wire in the signal cable to the boosters.  The signal from the command station controls a light emitting diode (LED) in an optical coupler integrated circuit and a light sensitive transistor in the optical coupler connects the signal to the booster electronics.

Other combinations are possible, but no effort will be made to define them.  Auto-reversing boosters for reversing sections also add to the complexity, but will not be discussed.  

The descriptive drawings make no effort to illustrate the mountain of wire, DCC circuit breakers, block detection, signaling system, stationary decoders, computer interface, and other stuff that may be hiding beneath a layout.   They are purposefully simplified to get to the essence of the connections between boosters and rail. 

CONFIGURATION #1

This configuration wires the booster common of all boosters together, and does not have a common rail between power districts.  The boundary between districts has both rails gapped and the only thing in common from one district to another is the wire that connects the booster common terminals together. 

COMMON WIRED BOOSTERS WITH OFFSET POWER PICKUP

The drawing titled “COMMON WIRED BOOSTERS WITH OFFSET POWER PICKUP” illustrates the issue of an older “brass” steam locomotive with offset power pickup and this booster configuration.  

When the locomotive is across the gap, the electrical path for the locomotive is from output terminal 1 of booster #1 to the Red rail in power district #1, through the decoder to the Black rail in power district #2, to output terminal #2 of Booster #2, through the booster common connection of Booster #2, and back to the booster common terminal of Booster #1.  Note that if the booster common connection was not in place, the locomotive would stall at the power district boundary because there would not be a complete circuit path.  Note also that the full track current will flow in the wire that connects the booster common terminals together.  Still further note that while the locomotive wheels ‘bridge the gap’ that the #1 output of the boosters are connected together.   Similarly note that while the tender wheels ‘bridge the gap’ that the #2 output of the boosters are connected together.   

Furthermore, note also that when the locomotive is in the position shown that there is no current path for output terminal #2 of Booster #1 and also there is no current path for output terminal #1 of Booster #2.  Booster #1, output #1 is putting out current to the Red rail and Booster #2, output #2 is receiving current from the Black rail.  So long as the locomotive is across the boundary, both Boosters are working under DCC ground fault conditions.  Knowing that this condition will exist every time this type of motive locomotive crosses a power district boundary makes it clear that track power cannot be immediately shut off when a DCC ground fault condition is detected.

COMMON WIRED BOOSTERS WITH ALL WHEEL POWER PICKUP

The drawing titled “COMMON WIRED BOOSTERS WITH ALL WHEEL POWER PICKUP” illustrates a locomotive with all wheel power pickup crossing the boundary between power districts wired with this booster configuration.  It is immediately seen that, although necessary for offset power pickup, the booster common connection is not required for motive power with all wheel pickup.  The all wheel pickup actually ‘bridges the gap’ between districts and makes its own connection between boosters.  While the locomotive is across the boundary, the boosters are actually connected in parallel.  

CONFIGURATION #2

This configuration has optically isolated boosters and common rail between power districts.  This configuration is essentially the opposite of configuration #1.  The boundary between districts has just one rail gapped and the only thing in common from one district to another is the common rail.  Some modelers may have physical gaps in both rails, with wired in jumpers that make the rails common between power districts.  

COMMON RAIL WITH OFFSET POWER PICKUP

The drawing titled “COMMON RAIL WITH OFFSET POWER PICKUP” illustrates the issue of an older “brass” steam locomotive with offset power pickup crossing the boundary between boosters in a common rail configuration.   

When the locomotive is across the gap, the electrical path for the locomotive is from output terminal #1 of booster #2 to the Red rail, through the common rail to the locomotive’s front wheels, through the decoder to the Black rail in power district #2, to output terminal #2 of Booster #2.  If the common rail connection was not in place, the locomotive would stall at the power district boundary because there would not be a complete circuit path.  Note that booster #1 cannot carry any current to or from the locomotive in power district #1 because there is not a complete circuit across its output terminals.  Note also that while the front tender wheels are in power district #1 and the rear tender wheels are in power district #2, the boosters are connected in parallel through the tender wiring.  

COMMON RAIL WITH ALL WHEEL POWER PICKUP

The drawing titled “COMMON RAIL WITH ALL WHEEL POWER PICKUP” illustrates a locomotive with all wheel power pickup crossing the boundary between power districts in a ‘no commons’ booster configuration.  It is obvious that although necessary for offset power pickup, the common rail connection is not required for motive power with all wheel pickup.  The all wheel pickup literally ‘bridges the gap’ between districts and makes its own connection between boosters.  While the locomotive is across the boundary, the boosters are actually connected in parallel.  

CONFIGURATION #3

This configuration has optically isolated boosters and no common rail between power districts.  The boundary between districts has both rails gapped and there is nothing in common (i.e. no metallic connection) from one district to another.  

NO COMMONS WITH OFFSET POWER PICKUP

The drawing titled “NO COMMONS WITH OFFSET POWER PICKUP” illustrates the issue of an older “brass” steam locomotive with offset power pickup and a ‘no commons’ booster configuration.  When the locomotive is across the gap, there is no electrical path through the locomotive and it will stall at the boundary between power districts.  This configuration cannot be used for motive power that has offset power pickup.  Note also that lighted cars that have offset power pickup will go dark while the car moves across the gaps between power districts.  

NO COMMONS WITH ALL WHEEL POWER PICKUP

The drawing titled “NO COMMONS WITH ALL WHEEL POWER PICKUP” illustrates a locomotive with all wheel power pickup crossing the boundary between power districts in a ‘no commons’ booster configuration.  It is immediately obvious that the all-wheel pickup ‘bridges the gap’ between districts and makes its own connection between boosters.  While the locomotive is across the boundary, the boosters are actually connected in parallel.  

OK, WHERE TO FROM HERE?

Each of the several configurations has its own set of peculiarities and they change as the type of motive power or lighted car crosses the boundary between boosters.  This review of the several configurations is purposefully limited in its scope, discussing just one loco crossing one boundary.  The list of peculiarities and questions grows as various operating situations are considered for a given configuration.  

When boosters are connected in parallel by the passing train, how likely is it that the boosters will equally share the current load of the locomotive or locomotives?  Many writers say that having two types of “commons” is terrible.  What happens to the common wired boosters in configuration #1 when multiple locos ‘bridge the gaps’ and connect the boosters in parallel and make a temporary ‘common rail’ to go with the ‘booster common’ being tied together?  Older brass trucks with conductive side frames would bridge the gap between power districts twice for every car.  Modern lighted cars with all wheel pickup would do the same.  How can there be ‘DCC Ground Fault Current’ in configurations #2 & 3, the boosters are optically isolated?  What happens when two 5 ampere boosters are connected in parallel by an ‘all wheel pickup’ locomotive or lighted car and something derails?  Is there a potential for 10 amperes or more to flow in the derailed equipment?  There are so many variables, what to do?

I read in the Greeley Freight Station Museum (now the Colorado Model Railroad Museum) newsletter, December 2013 (Volume 4 Issue 12) www.gfsm.org -

“Several problems have been there since the beginning; for instance, older “brass” steam engines which have the engine picking up power from one rail, while the tender is wired to pick up power from the other, won't run between booster power sectors. There is also the occasional problem that a short circuit in one power sector might trip the booster in another for no apparent reason; this makes the process of finding the cause of the problem very painful.”

The newsletter from Greeley was describing the replacement of 14 CVP Products “Booster 5“ track boosters with NCE 5 ampere boosters so that all 18 boosters used on the layout were from the same manufacturer.  The museum says “Both vendors are noted for meeting the [NMRA] standard, but we thought using equipment from the same vendor might work slightly better.”  Reading the newsletter reveals that their DCC installation uses a “DCC Booster Common Wire” as in configuration #1.  Although not explicit in the newsletter, it is presumed that both rails are gapped between power districts.

I have no idea if the Colorado Model Railroad Museum layout has had, does have, or ever will have, what I define as a DCC Ground Fault Current issue.  I know that I have read many Internet discussions regarding seemingly inexplicable DCC system behavior symptoms, with all three of the principle DCC systems.  It seems that these discussions simply fade away, with never an explanation or description of just what was changed that caused the undesirable symptoms to go away.  This is to be understood, if a seemingly inexplicable symptom does away after removing and reinstalling great gobs of layout wiring, what is there to be explained?  The modeler who has ‘fixed’ a serious problem and has no idea why, may prefer to remain quiet rather than speak out and be embarrassed by not having the slightest idea of what changed, other than that the undesirable symptoms of something being amiss have gone away.  Perhaps the modeler did nothing at all other than exist through a frustrating change of seasons whereby the change in relative humidity caused bench work to move and open a gap that had closed months before.  Exploring the possible consequences of DCC Ground Fault Current conditions may lead to a better understanding of the various ways a DCC layout can go awry.

A DCC GROUND FAULT DETECTOR

A DCC Ground Fault may exist in certain conditions, but those conditions are short-lived, lasting only while locomotives or current consuming rolling stock are crossing power district boundaries.  If a long term DCC Ground Fault condition is a bad thing, how can it be detected?  A current sensing transformer will do the job quite simply.  The transformer senses the magnetic field generated by a current carrying wire.  The correct operation of a DCC booster will have equal but opposite current in the two output wires going to/from the layout wiring.  If the current in two wires is equal and opposite, and the two wires are close together, the magnetic field surrounding the two of them will be zero.  If the two currents are opposite but not equal, then there will be a finite magnetic field surrounding them and the transformer will have an output that is proportional to the difference In the two currents.  It is convenient that such a transformer and some essential electronics as well, are packaged as the NCE BD20A block occupancy detector.

PHOTO 1

PHOTO 2

Photos 1 & 2 show a BD20A block occupancy detector packaged with a relay and a 4N35 optical isolator ready to be used as a DCC Ground Fault Detector and Indicator.  A potentiometer is connected to the BD20A, according to the NCE instructions, for sensitivity adjustment.  The assembly operates from a regulated 12 Volt DC power supply.  I used a Jameco P/N 2151005 plug in power supply.  The optical isolator output can be connected to other electronics where an isolated logic level signal is required.  The two sets of relay contacts can be used to operate an indicator or alarm.

DCC GFCI ASSEMBLY

A plan view of the assembly is shown in the drawing titled ‘DCC GFCI ASSEMBLY.’  The euro style terminal blocks, cut to length as required, are attached to the plywood base with #2 wood screws.  The sensitivity potentiometer and power connector are mounted on a bracket made for the purpose.  The relay is plugged into a DIN rail type socket that is held with a couple #6 screws.  The power connector is Jameco P/N 2151136, a plastic bodied connector that isolates power from the metal bracket.  The bracket was made from a piece of 1/16th inch thick by 1 inch wide aluminum bar from ACE Hardware. The 4N35 optical isolator is plugged into a DIP socket.  The socket pins are inserted directly into the terminal block while the input connections are made simply by soldering wires to the socket terminals.

DCC GFCI WIRING

The wiring is defined in a schematic and pictorial combination in the drawing titled ‘DCC GFCI WIRING.’  I used a few pairs of AWG 24 wire from some CAT3 cable for all the wiring except the twisted pair that goes through the transformer on the BD20A.  The wires passing through the transformer are AWG 20, twisted together as is my general practice.  The current limiting resistor for the optical isolator input is simply soldered right in series with the input.  The stiff DIP socket pins and the resistor lead will hold the resistor in place.

Operation was confirmed with an NCE Power Cab system, version 1.65, a DCC Specialties RRampMeter, REV 2009, an adjustable primary load, an adjustable “sneak path” load, and a moving coil meter used to measure the average “sneak path” current.  None of the instruments were calibrated by a nationally recognized testing laboratory.  

The DCC test load was set to 1.0 amperes as indicated by the RRampMeter. The Ground Fault load was varied to determine the sensitivity at minimum and maximum setting of the sensitivity control.  With the detector sensitivity control set to maximum sensitivity, a Ground Fault current of approximately 18 milliamperes or more was detected.  With the detector sensitivity control set to minimum sensitivity, a Ground Fault current of approximately 120 milliamperes or more was detected.

The ways that the output of the DCC Ground Fault Current detector may be used is essentially limitless.  The simplest would be just to drive an indicator, either visual or aural.  The other end of the application spectrum might be as an input to some sophisticated data acquisition and control system.  Perhaps a more reasonable use would simply be as a temporary tool to check the operation of a DCC layout, even if no particular symptoms seemed to warrant.

This article is not intended to be a detailed ‘how to build & use’ a DCC Ground Fault Detector and Indicator device.  The great variety of DCC layouts ranges from an Inglenook shunting puzzle on a plank to something like The Colorado Model Railroad Museum layout or those marvelous exhibits in Germany and The Netherlands.  The electronics ranges from a starter set by Digitrax or NCE to something that can barely be imagined.  

My intention is to provide a tool that might aid in the investigation of seemingly inexplicable symptom or symptoms of something not understood.

Any errors, technical or typographical, in this document are solely my own.

Rex G. Beistle

Submitted for publication by the Rocky Mountain Region of the National Model Railroad Association.

This original work is donated to the NMRA without expecting compensation of any sort.

If this is published by any other NMRA body, just give me credit.