TMCC Force Field

Have you tried to fix the problem spots with an additional hot and common feed pair?

Are the transformers powered from the same properly grounded receptacle? or another view, on the same circuit from the electrical panel.

Also, the signal is in the house wiring. Fluorescent lighting will create issues.

"Force field shields closely spaced track with small wire between"-Yoda

"Mike Reagan published a video (http://www.lionel.com/Customer....cfm?documentID=6355)
that describes the TMCC zero-crossing sync events, and the inter-track doubling of synch events as a source of interference.  This led to an AHAH! moment for me as I have dealt with drifting sync signals degrading synchronous communications in a past career, and sync interference can explain why parallel tracks work perfectly in one part of the layout and are problematic in another.  I conclude there are two factors for reliable TMCC reception by locomotives: signal strength sensed by the R2LC board measuring the RF signal between track/locomotive signal and the earth ground field; and the presence of a single 60 Htz zero crossing sync as sensed by the locomotive"

What is the wavelength of 60 Hz and of 455 KHz?

Wavelength of 60 Hz is 3100 miles, 455 KHz is 2160 ft.

If you're determined to get to the bottom of it and can work at the small-component level, how about metering the signal-strength of the TMCC signal at the R2LC receiver.  Some R2LC (and R4LC) modules utilize the 3372 chip to detect/demodulate the 455 kHz signal.  The 3372 has a pin called "meter out" which is a voltage proportional to the received signal strength.  Yes, you could drag around a flat-car with a digital meter attached to this pin, but you can also get compact digital meters for less than $2 on eBay (free shipping).  So this would be like the voltmeter car that measures track voltage but instead it measures TMCC signal strength.

I'd think there's value to mapping out the signal strength around the layout.  And when making adjustments to the layout wiring, grounding, etc. I'd think there's value to watching the signal-strength change in real-time telling you if you're getting warmer or colder so to speak.

What would be cool is to have a 10-LED stack that starts red, then yellow, then a couple of greens, they would be driven by the voltage.  It would tell you the signal strength at a glance.  Probably much easier to see at a glance.

So if the time/distance between "sync" pulses is 16.67 ms, and 3106 miles, how much of a difference in path length would it have to be to be out of "sync" 1 ms?

I like both ideas...some of these meters require an isolated power supply. No problem since a 9 vdc battery lasts a long time to run them. The LED stack is interesting. There are chips that do that...maybe a display ready made from China exists.

What is the range of the output on the R2LC for "full" scale?

Above techie plot from the 3372 datasheet.  The 3372 belongs to a "generation" of receiver IC chips that put out a current proportional to signal strength.  It was linear-in-dB which allowed a tremendous dynamic range of measurement.  In practice if measuring with a voltmeter, you simply place a resistor load to convert the current-to-voltage.  So in this case, with a full-scale output of, say, 55 microamps, a 51k resistor would yield a full-scale of 2.5V.  If using a 10-step LED bar graph you'd need choose you scale so each LED represents, say, 5 dB steps which would yield 50 dB range.  I don't know the range of signal level TMCC is meant to operate at but since the modulation is FSK, it probably operates in limiting or near the top of the scale.  So if using a bar-graph (vs. a digital meter) someone would have to experiment how to allocate the 10 steps.

Also, I'm not sure I understand this 60 Hz side discussion, but the meter-output signal is real-time and fast-responding.  In fact, a capacitor is typically placed on it to smooth out the measurement...but in principle the meter output signal can be used to crudely demodulate amplitude-modulation (AM).  Again, TMCC is an FM/FSK signal but if the issue is one of nulling or periodic dead-spots somehow tied in with the line-frequency, you can hook up an audio-amplifier to the meter-out signal and 'listen' for the amplitude variations in the received signal strength.  Thus, you'd hear 60 Hz "hum" if the signal strength was varying with the line-frequency.

Stan and John, sounds like a project to do real soon.  I have a digital meter in the parts bins and will look for the 3372 chip on my R2LC boards.

Chuck, its not the wavelength but the propagation of the 60 Hz wave (at some velocity less than c) that can shift the zero crossing.  Is 1 ms the "window" in which the R2LC senses zero crossing?  In a typical layout with multiple power feeds per loop, can the zero crossing event ever get out of sync?  Noise in the AC wave is a more likely source of sync loss.  I don't know if what Mike Reagan described is a red herring but he emphasized the sync interference and "force field" in the video, so I want to understand the technical underpinnings of Mike's assertions.

At the World's Greatest Hobby at Chantilly VA in 2010, the National Capitol Trackers put together a 140 ft long layout.  TMCC worked.  The biggest interference problem was from multiple CAB1s and TMCC bases on other layouts, all using the same frequency.

I suspect that TMCC signal strength as picked up by the locomotive antennas is the fundamental factor to address, and extra earth ground wires are counteracting signal swamping rather than creating a force field.

John

 Is there any truth  to finding weak TMCC  signal spots by holding the whistle button down ? If  the whistle quits  there a signal problem at that location?? 

Something like the LM3914 Dot/Bar Display Driver would work for this application.  It can handle full-scale voltages down to 1.2 volts and has high impedance inputs.

There are three versions of the part.

• LM3914 - linear

• LM3915 - log response

• LM3916 - "extended logarithmic" VU meter type display.

After watching Mike's video for the first time, I contacted Jon Z. at Lionel to complain about both the sync problem and the idea of a reflector.  Jon agreed that Mike was wrong and that Mike apparently didn't fully understand TMCC signals.  I suggested that Jon, as Chief Technology Officer, should do something about the video's errors, but years later the video still persists.

As cjack is pointing out, at these relatively low frequencies the wavelengths are huge compared to size of even the biggest layouts, resulting in only insignificantly small phase differences.  The chip-to-chip variation of the comparator that detects the zero crossing creates more "timing error" than any signal phase error. 

Also, the TMCC system timing is designed to be fairly loose.  The sync signal is needed to synchronize the conduction of the motor-drive transistors to create the proper "shark fin" current in the motor(s).  It also opens a window for the signal decoder, but the serial command data is relatively slow and robust.

Yes please John! That would be far more interesting to me than chuffs. I think there would be a decent market for such a device. Without a way to quantify signal strength, trying to make thing run smoothly is like chasing ghosts.

While I think this would be useful, I somehow don't think it would sell like the Super-Chuffer!

Hello, guys--

Seems like yesterday I looked at this during the posting, and now 10 days past.  If anyone is still reading, I have some thoughts on this signal strength issue.  I think in general the "signal strength" is not falling off because of the distance it has to travel out into the layout, and in general it will not be, unlike DCS, noticeably diminished by being divided into say more than 10 different sidings at some point.  This is because the TMCC wavelength is sufficiently longer than most of the numerous DCS frequencies (which range up to above 3 MHz, plus I believe their 3d harmonics, which just clear the FM intermediate frequency at 10.25 MHz).

Rather, I would look at the TMCC signal in the way the locomotive sees it-- as a field of electrostatic potential (ie an ac voltage cycling at 525,000 times per second, perhaps output at as much as 5 volts peak (3v with Legacy?), perhaps not, intermittently at the 60-cycle voltage zero crossing (when a command is to be sent).  Regardless, the radio receiver section (R2L2?) will amplify it to a usable 5v or 3v and clip it for dc in the logic decoder.  That is, if it gets any reasonable signal at all.  Now what could happen to cause no signal?

Well, all sorts of things, as Mike described.  I'll go with a simple example of a main line tunneling under a large passenger terminal on an upper level (say 7" above) with several sidings.  So the metallic wired section of the radio path (the sending antenna which creates the field) is the collection of all the outside rails, and they are all at essentially the same signal voltage, say 5v.  That is because the air path has so much higher a resistance (is it 10^12 times? who can remember.)  No matter, all the radio field voltage drop occurs essentially in the air.

Now this layout is in the cellar, with its concrete floor bonded to the earth power ground in some way (if for no other reason than its resistance not being 10^12 times).  So where do these lines of electrostatic force go from the rails?  More or less to the concrete floor, at 0v.  But these line do not just go straight down, because each line is surrounded by a magnetic field, which doesn't like to share space with the field from the neighboring line.  So some of the lines have to start off going straight up, like a water fountain.

Now where these lines are closest together, their associated magnetic fields will be stronger, and therefore the current flowing in these locations will have to be stronger (ie, the microamps per sq inch).  It follows then that the rate of voltage change along the lines of electric force will be relatively larger closer to the outer rails.

Now the receiving antenna in the roof of a plastic-bodied diesel is about 4" above the rails, or a little less than 3" above the metal floor of the diesel (to the extent the floor takes up the action of the rails).  So you'll see that the antenna is actually in a region of the electrostatic field where the voltage is somewhat different that that of the radio signal at the outside rails.  The length of the antenna foil does not act to change that voltage, but to increase the current drawable from the electrostatic field at the voltage present at that particular elevation.

The magnetic field is everywhere at a right angle to the lines of electrostatic force.  This property can be used to sketch the electrostatic field.  Visualized the fountain of lines like a water fountain, and draw a vertical section from floor to track at a right angle to the track.  Draw the lines, and cross them (with the magnetic lines) in such a way that every block is a square, and each successive row of blocks also squares.  Obviously from track to floor each row of squares is a little larger.  The voltage drop across each block is the same, and from this you can calculate the voltage drop at the locomotive foil receiving antenna.

You can buy a field strength meter, which is a probe consisting of a wire loop of known area, which will charge a capacitor (thru a diode) when it is held flat in an electrostatic field and turned over 180 degrees.  No need, tho-- it is actually possible draw the field of little squares with some practice fairly accurately.

The little squares will show you that the upper yard will reduce the electrical attraction drawing the main track lines of force upward to practically zero.  It is easier just to say that on a shortest line joining two outside rails, there will be no change in the air voltage, and if such a line passes thru the locomotive receiving antenna, there will be no voltage at the antenna.

A yard of some tracks will reduce the concentration of lines above each track and thus the voltage drop at the internal antennas will be lessened from what it might otherwise be with single track.

Mike suggested dealing with this by introducing house system ground wires into these situations in particular ways.  If I may, I'd like to suggest the the ceramic capacitor be reduced to 1/10th the value Mike suggested, when any outlet connection is made.  That is because you don't have a whole layout of these wires, just limited areas if you have them.  In this smaller size, a certified personnel protective ceramic capacitor can be obtained that suitable for 120v protection.  (Last time I wrote this, a pop-up ad informed me I could have 50 for $1 each.)  I found that it was not practical to make this protection in larger capacities (leakage current apparently becomes too large).

Of course, the size that Mike suggested should still be used for bridging gaps in the outer rails and other uses downstream of the Lionel Base devices.  But direct connections to an outlet should have the certified protective ceramic capacitor, and the ground wire extending from it should be insulated at the voltage required for house wiring with the anti-abrasion covering, protected from damage and contact (no exposed ends).  There should be no ground loops, nor connection to anything on the layout.  I cannot recommend grounded foil-backing or grounded wire meshes at all.

I had thought that the NEC did not permit such connections (ie, other than by UL-labelled utilization equipment), but I cannot find that.  I now believe it must be that the outlets may only be used in accordance with their labeling, and no manufacturer has labeled their outlets for such use (EC&M digs these nuggets out).

The postwar ZW transformer I believe enjoys a presumptive position in this matter as an inspector approved device (IMHO), because obviously a lot have been used at public shows-- also there is no blanket NEC prohibition of older equipment provided it is properly maintained and not functionally altered.

But for these ground extension, I would if publicly exhibiting a layout in Minnesota, have such a protective ground device made by a licensed ham operator, and ask the local inspector to approve its use (presuming the use to be necessary).  Minnesota is known for checking public layouts, but I would think any inspector would notice a green wire trailing from an outlet box on a column.

With a cellar layout, in an unfinished cellar you have grounds everywhere, so that it is good practice to upgrade to the ground fault personnel protective outlets, and all the more so if wired, good-conductive grounds enter the layout picture.  In other areas, not only do the arc-fault breakers enter the picture, but any electrician doing work other than very minor work is now required to convert the circuits (some places now say all the circuits that a new house would require) to such protection.  This can lead to new panels and worse (up to rewiring the whole house).  But for the modeler, it gets worse:

The arc-fault breakers are advertised as combination breakers retaining the ground fault protection.  That is very misleading.  All of the available devices (from only 4 manufacturers last I looked) no longer provide the 5 ma ground fault personnel protective trip in their arc fault breakers, but instead only the ground fault equipment protective trip and that at 30 or 60 ma (it used to be 10).  I can tell you that 30 ma is already not good, and I'm not joking.  Word on the street is that property losses are the main concern.

I presume you can still insert the ground fault personnel protective devices in the first outlet even on an arc-fault circuit, but this has not been a matter with much experience yet.  (The modeler would want to do this in a finished basement.)  You see, the arc-fault device is finicky about downstream wiring (which is why some houses may require extensive rewiring, totally regardless of model trains or not.

Also of course, don't get your grounds from any source other than the power ground of the house.  The separately grounded outdoor TV twinlead of yore proved to be a hazard with lightning, as indeed the overhead telephone line.  Today in my area, the cable companies are the new offenders.  Don't get your TMCC ground by burying a wire outside a convenient cellar window.

Sorry-- when it comes to grounds, there are a world of caveats.

--Frank

Just curious, where does this come from?

525,000 times per second

Which begs the bigger question...

Gunrunner--

525,000 cycles is more commonly called 525 kHz these days.  I just used the non-abbreviated form to make the difference from the 60 cycle transformer track power more obviously a very large difference.

The large difference means that it is very easy to continue to isolate sections of one outside rail for signal indications, crossing gates, and non-derailing switches.  Bridging these gaps with the 0.1 microfarad capacitor (Mike's suggestion) easily transmits the TMCC signal carrier frequency (525 kHz) with very little loss, while almost totally blocking 60 cycle current.

The 525 kHz has been commonly used on long-distance power lines to enable a relay system to identify between which of the intermediate switching stations a circuit, that has dropped or broken a wire, is located.  Then only the breakers nearest the fault are opened to isolate it.  I tried to locate a reference for you as to how this works, but as the idea is older than the internet I had no luck.

Can't recall how or when I came to know this history of the frequency, or even how I knew it was used by TMCC.  I assume 20 years ago someone put an oscilloscope on it.  It was also used on school campuses to transmit school radio over the power wires to the dorms, which was a simple license for the school to get.  It is just below the AM broadcast band (540 and up), so ordinary radios could receive it if plugged into the campus power.

Anyone knowing of this would see some advantages in using it, with a standard non-interference license.  It is readily transmitted long distances over wire pairs; easily isolated by grounding the wire path at the house ground; blocked by the local power transformer: circuitry for generating and receiving the particular frequency is documented; protection of the wire-borne Roku and Netflix services in the 10-30 mHz band already in place by others.  Lastly, it is not likely to be a frequency reallocated from its present use, unlike the 27 mHz band of the TMCC Cab-1 controller.  And the new idea of running the train on the antenna turned out to be patentable.

I wish I could remember how I learned the use of this frequency for the TMCC carrier.  I could be mistaken, but it just seems so obvious.

--Frank

The guys can speak for themselves, but I think they are puzzled by 525 kHz when it is well-documented in OGR and elsewhere that TMCC uses a 455 kHz carrier.  Similarly, I'm curious about your use of 10.25 MHz for the FM IF frequency.  10.7 MHz is the most widely used IF frequency for FM.  Agreed these are negligible differences wrt to your thoughts on wavelength issues.  However, it clouds the discussion to see these two discrepancies at the start of your post...in my opinion of course.

Stan--

Thanks for correcting my errors with the 455 kHz TMCC carrier and the 10.7 MHz FM intermediate frequencies.  MTH reported their voice channel at 10.25 MHz and I just assumed that was the FM intermediate frequency, without ever checking that.

--Frank

The problem is, when those kinds of errors appear in a post, it gets you wondering what else is accurate or inaccurate in the post.  I don't think you can blame people for doubting what you're saying when some of the basic facts are incorrect, right?

Gunrunnerjohn--

You're right about that, which is why I usually try to avoid such errors.  I am red-faced .

My main concern is about the use of grounded wires (green wire in house 120v power) in the layout environment.  One thing always to remember is that that green wire is the same size and copper just like the white (return) wire, and it is bolted to the white wires bus in the breaker box.  A solid fault in an outlet will place 60 volts on the green wire in that outlet and every downstream outlet in that circuit.  It will persist until the breaker opens.  That very thing happened to me last spring while assisting an experienced electrician test his repair of amateur work at my son's first house-- there was no GFPP in the circuit and it took the breaker more than 5 seconds (10?) to open (bare copper ground wire in the open outlet box had folded back against the hot terminal on the outlet device).

In houses older than his, the green wire would have been 16 not 14, and the rise would have been to about 75 volts.  You would not believe how many people on this board would not believe the green wire could ever have a voltage other than zero, let alone half or more of  the line voltage.  Some, I believe, were electricians.

The green wire is a two-edged sword.  If anyone here doubts that I'll endeavor to provide an example and some calculations.

But as you say, forgetting the AM intermediate frequency does cast doubt on what I say.  Absolutely so.

--Frank

Speaking on inaccuracies, I made one now that I have checked the electrical code. The amp rating for 12 ga for example is rated as 20 amps for a 60 foot run. 10 ga for a 100 foot run.

Another thing I learned last year is that the code for buildings with a large number of switching power supply loads is under revision. It turns out that the switching supply loads cause a greater heating so wire sizes have to be larger. The issue is a high harmonic content on utility lines and building wire.

Chuck, there are a couple of different ways of rating a wire.  The 60' that you mentioned is the length for a specified percentage voltage drop for 20 amps AT 120 Volts!!

My book shows 36' for a 2% voltage drop with #12 at 120V - 2.4V of drop.  For our 18V train systems, the same 2.4V of drop is a very significant drop - 13%.

Ahh yes, thanks! I actually did do my planning using resistance per foot when I ran the red/black zip wire. I have some that is copper clad aluminum and figured I needed to know what I actually had. But I didn't think about that this time.

Sounds like 20 feet or so is about the limit for 20 amps. That would be 1.33 volts. Of course that would be maybe 6 to 8 engines running at once.

But who runs a single wire from a 20 amp transformer?  Add a parallel feeder and the voltage drop is half.  2 parallel 10amp power sources also halves the voltage drop.

I use a short run from the transformer to the distribution block.  Then multiple feeders.  Isn't that the model folks use for wiring a layout? G

Probably. But just good to know what the limitations are.

But...

Take the example above with #12 wire.  Assume you have an oval and feeders at each end.  As an example, let's say that there it is a 20' wire run to the far end and 0' to the near end.  Indeed, the current will split between the two feeders, but what happens if all the rail joints aren't perfect?

The resistance of the 20' wire loop is only .065 ohms.  If any (or a collection) of the rail joints between the two feeders are of the same order of magnitude resistance, the current sharing of the two feeders will be sorely inhibited.  (This assumes we have "weak" joints on both sides of the oval.)

By the way, 20 amps is two 180W bricks in parallel, and Forum members report that this combination makes a good arc welder when there is a derail.  The surge current during a short is MUCH higher than 20 amps.

I have a passenger car that was the victim of the paralleled PH180 bricks.  We did that once on our modular layout because of all the long passenger trains that day.  However, several derailments, including one of mine, made us realize that was a bad idea.  I personally will not run with that much power on a single power district.

yep, yep yep.  All in track work and wiring.  I have seen a layout with 4 PH in parallel for each block, with several blocks completing a loop.  All because 2 Powered ABA and ABBA pulmore motor engine sets could be on the block at the same time.

Derailments where spectacular

As an aside just because you have a 2-3 volt loss isn't the end of the world either.  14-15 Volts runs TMCC fine.  Of course the cruise guys would not like it.  G

Command mode doesn't care as much about voltage drop, but I see a slowdown with as little as 1V of drop with some PW conventional engines.

Dale-- The surge current, if you mean the current fed into a short circuit, can be up to 77 amps with the small brick (7A nominal, 135W).  I measured this using the standard procedure.  This is the current available at the connector end of the output pigtail.

I did not want to repeat this with the 10A large brick (180 watts).  Instead I rely on the common situation with transformers of similar design-- that if their sizes are not too far apart, properties such as the available short circuit current will be proportional.  On that basis, the current available from the 10A brick will 110A; from two paralleled, 220A.  [A recent study found the tripping currents on residential breakers to be between 100 and 1000 amperes, roughly.  This is determined almost entirely by the resistance of the branch circuits, although the path through the breakers (main and branch) adds somewhat to this.]

As the wired path to the location of the short increases in distance, the actual short circuit current decreases.  The maximum power deliverable to an arc welder located at the point of joining the pigtails of two paralleled 180-watt bricks occurs when the voltage drop across the arc is half the 18-volt supply, or 9-volts.  Thus the energy in the welding operation is 9V x 220A or about 2000 watts.  The power drawn from the wall outlet will be twice that, or 4000 watts.   This will trip the house breaker, but how fast?

Let's assume the derailment is far enough down the track that only 1200 watts can be drawn for welding the 700E to the rails.  Then 2400 watts will be drawn from the wall outlet, or 1.33 times its short time rating.  IIRC, the breaker requirement is that at 1.35 times rating, it should open in 2 minutes.  Of course, it is only required to do this once (the postwar ZW used a commercial 120v breaker, and we know what happens to its opening times...).

Oh, yes... the "fast-acting breaker" internal to the brick.  In the smaller brick this is a Millionspot 10A SPST relay (operated by a supervisory circuit) whose specifications do not list either a closing rating or an opening rating.  In U.S. practice this means it will carry 10A, but not open or close into 10A.  It is reported that the 180w brick has a 15A relay.  I can tell you these 10c devices will have their contacts welded shut by anything near 77 or 110 amperes. The UL standard for toy transformers was actually rewritten at their introduction to specifically state that the requirement that any current that could be produced must be interrupted by their overcurrent devices did not have to be met.

That's the reason for the special plug-- the electronic voltage regulators (the various PowerMasters, new ZW, and Legacy ZW(?)) have self-protecting MOSFET transistors (one cycle max to open), and the Lineside [Shack] Lockon supervised relay has a relatively heavier contact (but still looks light enough to be sub-cycle).  It is a cascaded protection system in two parts, only half of which is in the brick.

There was one permanently open contact [burned] and one permanently closed contact [welded] reported here soon after the advent of the 135 amp brick.  Their owners had assumed "factory defect."

The standard method for short circuit current is slightly approximate.  It is customary simply to mention it, and also to indicate if the transformer magnetizing current was supplied from the line side or the load side.  I supplied it from the line side.

I never describe the test in a public forum as the connections to the transformer are non-standard and can be hazardous if not made correctly.  Also, these transformers have electronic circuits to measure and limit the output power; these circuits are supplied with a non-standard voltage during the test, which could be damaging to this feature.  Actually, I only have the test result because, while distracted by thinking about this problem, I absentmindedly started the test. 

As a frame of reference, the postwar ZW has a maximum short circuit current of about 45 amperes (either of the two models or their variants, from any combination of the four outputs, from one to all).  As its output power is about the same as the 180-watt brick, you could say that brick is two and a half times the welder, if its intended downstream companions are omitted (two in parallel would be 5 times the welder).  Those ZWs were arranged to have deliberate magnetic field leakage to add reactive impedance at high currents.  In the bricks, leakage is deliberately minimized, both to reduce voltage drop under load and to reduce the inductive voltage surge on intermittent opening of the secondary.  This can be seen from the positioning of the coils.  The maximum surge voltage from the postwar ZW was determined by QSI to be 39 volts.

Large commercial transformers generally do not have exactly the same number of turns from one to the next.  When to be paralleled, as in multiple substations within a building, as by connection of the substations to a ring bus, they are often required to be furnished in factory-matched set.  This is often a requirement of the utility particularly where a buried networked distribution system is involved; I assumed this is because of the risk they present to the network.  The ones I've seen use an array of 3 transformers in each substation.  Even so, it is customary to derate the combined group by 10% due to heating which may occur because of circulating currents between the transformers.  I am not sure if toy train transformers are more likely to be more closely identical to each other.

--Frank