Tutorial on DCS wiring consideration and the down-side of "magic bulbs"

In the video above Adrian advocates for routing the positive and negative track wire together to support the DCS signal. 

A now deceased forum member Dale Manquen did a lot of research and posting on the TMCC system.  He argued (though I cant find a reference post) that the positive track wire should be separately routed from the ground track wire to support the TMCC signal.

If both are true, this seems to create a dilemma for larger layouts that want to operate both TMCC and DCS.

Bill

Still have not found Dale's post, but found this from another poster:

"The outer track signal to the Base U Terminal can get "blocked" by capacitance set up in the wire by running it in a bundle with ground or hot wire.  This can be cured by running the outer track wire to the base away from any other wires."

Also

"What causes problems is when the wires from the Base U terminal are run close to the earth ground or track power wires.  When two wires are close together there is a property called capacitance between the wires.  This capacitance acts as a path for radio frequency signals.  The closer the distance between the wires and the longer the wires run together the more the capacitance.  This capacitance can effectively short out the signal from the Base.

 To prevent this, do not run the Base U terminal wires in the same wire bundle as the ground or run them over a ground plane. "

Again this seems to create a dilemma between ideal TMCC and DCS wiring.

Bill

What is recommended instead of the bulb?

Tom,

Adrian

What are some of the sources of disturbance (inductance ?)  also would the Deats filter be better  than no filter ?

ogaugenut posted:

Still have not found Dale's post, but found this from another poster:

"The outer track signal to the Base U Terminal can get "blocked" by capacitance set up in the wire by running it in a bundle with ground or hot wire.  This can be cured by running the outer track wire to the base away from any other wires."

Also

"What causes problems is when the wires from the Base U terminal are run close to the earth ground or track power wires.  When two wires are close together there is a property called capacitance between the wires.  This capacitance acts as a path for radio frequency signals.  The closer the distance between the wires and the longer the wires run together the more the capacitance.  This capacitance can effectively short out the signal from the Base.

 To prevent this, do not run the Base U terminal wires in the same wire bundle as the ground or run them over a ground plane. "

Again this seems to create a dilemma between ideal TMCC and DCS wiring.

Bill

There is a dilemma. This is pretty unavoidable since TMCC is single ended and the DCS is differential and have opposite transmission objectives.  I guess it comes down to trade offs between how well you want each to run.

Of course DCS is beter because this post is in the DCS part of the forum.

Tom Tee posted:

What is recommended instead of the bulb?

The bulb or filter are a patch job treatment of too much inductance in the wiring from split cables or connections. They compensate for the inductance at the cost of lowering the dcs signal voltage considerably (bad). The best is to go back and try to figure out why your layout is inductive in the first place and correct it.

clem k posted:

Adrian

What are some of the sources of disturbance (inductance ?)  also would the Deats filter be better  than no filter ?

The aadded inductance is from having too much space between the + and - wire in the layout going to the tracks. The filter can fix the resulting ringing is always going to reduce signal strength. If you want optimal performance you need to remove all loads on the TIU (no bulbs, no filters, only trains) then adjust your wiring to remove the ringing 

Adrian.....I use the double wire sheathed, and was thinking of placing the filters at the far point of each block               (4 blocks) and at the end of each spur track. I run lash-ups of five or more locomotives and four trains all on one 300' loop. Lash-ups really like good strong signal. 

Thank you 

Clem

Adrian,

This is what I use on my layout right now. I've learned from you that the TVS diodes I added probably aren't doing any good. Same for the filters. I use a red-black wire pair for distribution.

Each track block has an insulating pin on the center rail. Should I be feeding the power into the block at one end of the block? Should I have insulating pins on the outer rails at the end of each block?

in ham Radio they sell stuff called "ladder Line" with parallel wires spaced apart about 3/4 of an inch. So that is an option for minimizing capacitance over long wire runs.

But the capacitance referenced in this post should not, IMHO, be an issue except for long runs. So layout size, TIU locations and so forth will have a bearing on whether or not you need to pay attention to a potential (pun!)  capacitance problem, right? I think some real engineering investigation, testing and measurement is required to break down , understand and deal with this problem. I feel like I am sensing a lot of "groping in the dark" when I read these posts. Lord knows I have not got the necessary expertise to judge this problem.

Don Merz

Run the wire for the DCS signal and TMCC takes care of itself.  Avoid rats nests and keep the power and ground wires at equal length located together, or in close proximity.  I once solved a pesky DCS yard signal problem by loosely twisting a multitude of track power and ground pairs together back to the terminal block.

Electricity doesn't take the shortest path, it takes all paths (Ohm's law).  Simplify the wiring (routes, terminal blocks, etc.) as much as possible.

My modular club finally achieved decent DCS signal propagation after many years of trying magic light bulbs, filters, insulated blocks, magic incantations, etc.  The solution was to make the wiring consistent and keep each loop 100% isolated from the other loops.  Cross-overs have all three rails insulated.  Each loop has its own power and ground.  Wiring harnesses are standardized with 12G stranded wire and track feeds are 14G.  Filters were removed.  No magic lights.   No accessories drawing track power.  The only place where the ground is common is the output side of the TIU.  Legacy signal is attached at the TIU.

Running twisted pair cable would keep the wires together and would address this portion of the DCS signal loss between the TIU and track (active mode).  Run the TMCC "U" terminal signal connection separate and away from all earth ground connections (water pipes, metal electrical boxes, 120 AC power lines, etc, etc) as the other half of the TMCC signal is carried on the earth ground path.  Need to try to keep the TMCC "U" terminal signal approximately 2+" away from earth ground to minimize the capacitance effect talked about above.  This should allow both DCS & TMCC to work together.

Bob D

Tracker John posted:

My modular club finally achieved decent DCS signal propagation after many years of trying magic light bulbs, filters, insulated blocks, magic incantations, etc.  The solution was to make the wiring consistent and keep each loop 100% isolated from the other loops.  Cross-overs have all three rails insulated.  Each loop has its own power and ground.  Wiring harnesses are standardized with 12G stranded wire and track feeds are 14G.  Filters were removed.  No magic lights.   No accessories drawing track power.  The only place where the ground is common is the output side of the TIU.  Legacy signal is attached at the TIU.

This is like our club. Clean wire, tightly spaced twisted pairs, no added loads of any kind and solid performance.

Given the concurrent post about Rev L issues, and based on experience on my layout, I would not buy a Rev L to avoid bulbs/filters.  If your Rev G, H, or I works satisfactorily, stick with it.

I guess I'm an iconoclast, so you may want to disregard the above.  I don't use star wiring.  I don't run hot and ground close together.  My TIUs feed into a control panel which has 76 toggle switches each feeding one insulated block (through 14-gauge), I use a common bus for ground/neutral that runs all around the layout, with frequent drops from outside rails and serving as ground for all accessories.  I don't feed ground through the TIU.  All I do is run trains with perfect control all over my 36x16 layout.

ogaugenut posted:

 To prevent this, do not run the Base U terminal wires in the same wire bundle as the ground or run them over a ground plane. "

Again this seems to create a dilemma between ideal TMCC and DCS wiring.

This is a non-issue I believe.  They're talking about any earth ground wire you run to help the TMCC signal, NOT the transformer hot and common lines. The transformer common and hot can be run together with no impact on TMCC operations.

The earth ground wire should be kept away from the transformer wires to minimize the capacitive loading of the command base drive.  AAMOF, the TMCC buffer that I'll be cranking out soon is specifically designed to provide additional drive to the signal to overcome a greater capacitance between earth ground and the track common, such is found on larger layouts.

Here's a nice side-by-side comparison.

I took the same track, same engine, and same TIU channel. The only difference is one capture has the + and - feed twisted together over 15 ft while in the other I split the twisted pair about 15 inches apart.

Twisted Together:

Pulled Apart:

Adrian! posted:

Tracker John posted:

My modular club finally achieved decent DCS signal propagation after many years of trying magic light bulbs, filters, insulated blocks, magic incantations, etc.  The solution was to make the wiring consistent and keep each loop 100% isolated from the other loops.  Cross-overs have all three rails insulated.  Each loop has its own power and ground.  Wiring harnesses are standardized with 12G stranded wire and track feeds are 14G.  Filters were removed.  No magic lights.   No accessories drawing track power.  The only place where the ground is common is the output side of the TIU.  Legacy signal is attached at the TIU.

This is like our club. Clean wire, tightly spaced twisted pairs, no added loads of any kind and solid performance.

I got into a discussion about actually (breaking the outside rails in your case) breaking both of my rails into blocks outside on my larger G scale layout. After a certain DCS upgrade I began having signal response issues. This cured it for me. I was told that it should not matter. I don't know why it happened all of a sudden or maybe it snuck up on me and finally showed up strong enough to wreck my consists that grew larger?

 This is on Aristo G scale stainless track (2 rail) outside with my mainlines divided into four blocks. They have performed well ever since. I thought it was the brass track joiners oxidizing so I upgraded to stainless on the inside mainline. The outside main right next to it still uses the same brass joiners to this day. Breaking both rails solved my issue.

 I find these posts very interesting Adrian and I appreciate you doing the work and sharing it here!

I have to be honest, I don't know what I'm looking at in the photos of the scope. What does that show exactly? Is the wires pulled apart method better for sharper signal or something? Which one is better?

Engineer-Joe posted:

 I find these posts very interesting Adrian and I appreciate you doing the work and sharing it here!

I have to be honest, I don't know what I'm looking at in the photos of the scope. What does that show exactly? Is the wires pulled apart method better for sharper signal or something? Which one is better?

With the wires puled apart everything is roundy and ringy and it's hard to tell exactly what is "0" and what is "1". The DCS decoder performs poorly like this.

When the wires are tightly coupled it's easy to tell the "0" and "1" so the DCS decoder performs better.

Thank you again!

I am with E-J above, was going to ask the same thing he did. Thanks for the further explanation! 

I've been following this thread and would like some clarification. The implication is that the DCS TIU track based signal is fed to the locos directly via the pickup rollers as opposed to TMCC where the track propagates the base generated low frequency RF signal to a loco mounted internal antenna. Is this correct?

BOB WALKER posted:

I've been following this thread and would like some clarification. The implication is that the DCS TIU track based signal is fed to the locos directly via the pickup rollers as opposed to TMCC where the track propagates the base generated low frequency RF signal to a loco mounted internal antenna. Is this correct?

 

DCS has a wideband excitation on the center rail and return on the outside rails. So that part is very circuit-like. The pickup roller takes the signal to the decoder and grounds on the wheels.

TMCC excitation is applied between the layout as a whole and earth ground.  People call it an antenna but the wavelength is like 1800ft so im not comfortable with that since it doesn’t meet formal definitions. Let’s say “capacitively coupled”

It should be remembered that Adrian's experiments point to the absolute optimum.  But, if a layout runs fine with good signals all around, it's not necessary to attain the electronic optimum.  There are many layouts (including mine), where hot & ground wires are not run together --- in fact, I use a common buss loop running all around the layout --- and signals run at 10 or so.  Some of my hot leads are a good 25-30 feet long, and all feed through toggle switches on a central switchboard, but all are 14 gauge from TIU to toggle to track center rail

So, Oguagenut, don't sweat the small stuff.

Adrian! posted:

BOB WALKER posted:

I've been following this thread and would like some clarification. The implication is that the DCS TIU track based signal is fed to the locos directly via the pickup rollers as opposed to TMCC where the track propagates the base generated low frequency RF signal to a loco mounted internal antenna. Is this correct?

 

DCS has a wideband excitation on the center rail and return on the outside rails. So that part is very circuit-like. The pickup roller takes the signal to the decoder and grounds on the wheels.

TMCC excitation is applied between the layout as a whole and earth ground.  People call it an antenna but the wavelength is like 1800ft so im not comfortable with that since it doesn’t meet formal definitions. Let’s say “capacitively coupled”

 

Capacitive coupled!  Now that makes sense to me.  I have DCS only, but have read and heard about TMCC/Legacy and how they do it.  In my "first career" 1976 through 1993, they used capacitive coupling in the power substations for putting low frequency signals right on the high voltage transmission lines to send information to trip breakers at remote stations.  This makes a bit more sense now.

Thanks for the clarification on DCS. What is the frequency of the RF carrier that is applied to the track in DCS? Can I assume that DCS locos have no need for a loco mounted antenna?  Since the track is used to radiate the TMCC signal, it can loosely be considered an antenna.

DCS has a 3.27mhz carrier.

The track really isn't an antenna since it is wired directly to the TMCC receiver, it's more like a "ground".  The locomotive TMCC antenna is picking up the other side of the voltage.

There is no antenna in a DCS loco.

BOB WALKER posted:

Thanks for the clarification on DCS. What is the frequency of the RF carrier that is applied to the track in DCS? Can I assume that DCS locos have no need for a loco mounted antenna?  Since the track is used to radiate the TMCC signal, it can loosely be considered an antenna.

DCS isn't RF so it has no carrier. It's a broadband ds/ss or CDMA signal. The symbol rate is about 3.3 MHz. Here's a detailed explanation posted awhile back if you want the numbers and such.

Adrian,

Back in the 1990's I was one of the first people to run dcc with O scale trolleys. This brings back many memories as I had to plow through a lot to get it running the way I wanted it to. 

Thank you for sharing all of this, I just bought an old zw and some quick circuit breakers for my new layout, you are making me think dcs or tmcc would be a better way to go.

I was planning on keeping the outer rails separated so I could run relayless 2 color signalling. Maybe I won't.