DCS and wiring

Art;

Since no one has jumped in yet, I'll offer a few comments.

First, not sure why you want to run DCS in passive mode?

DCS and Legacy will play well together, with the Legacy base cabled to the TIU using the MTH cable.

Each TIU channel will handle 10 amps or 180 watts of power. This is enough to run several trains on each, or at least two long passenger train with incandescent lighting.

You could easily run your lower level on one fixed TIU channel and the upper on the second fixed channel. Observe the 250 feet track limit per channel.

You should observe the star wiring guidelines using a distribution block for each channel. You should also observe the 10-12 track pieces per power feed, and no more than 10-12 feeds per channel/distribution block.

On my layout the common outside rail returns are all buss wired, and it seems to work just fine.

I power 400 feet of track total using all 4 TIU channels and magic bulbs with an older rev I3a TIU. The two variable channels are set to fixed. I get mostly 10's, with an occasional 8-9. I use lighted bumpers at the end of several sidings where the signal previously was poor.

 

Hope that helps,

Rod

Art,

 

While DCS is a lot more tolerant of most wiring schemes than is advertised by MTH, star or home run wiring is almost guaranteed to give you the best possible DCS experience.

 

Your proposed wiring scheme may work just fine, however, if you're building a new layout the absolute best way to go is star wiring. The additional increment in effoft is offset by the fact that you'll know that your DCS signal strength will be maximized if you follow the wiring guidelines in The DCS O Gauge Companion.

 

As regards passive TIU mode, what do you believe that you'll gain? On the negative side, it's more work, you lose the use of variable channels in variable mode and you lose the E-Stop function.

My question is this- can I get away with 4 blocks and bus wiring otherwise I would need 12 -14 blocks and a whole lot more complicted wiring?

 

I 'd bet you can. Use a tiu channel for each block, Only one way to find out though... hook up one block, then another etc.

thanks to all for your reply and to Barry your book is an excellent source - glad I have it. The reason for using DCS passive is as follows: I run conventional, Legacy and DCS. The starting voltage is to high for modern can motors but with Legacy or TMCC I can start from zero. In the passsive  mode I can overlay DCS signal on any channel. At least this is what I am trying to accomplish. I did not realize that I can set variable to fixed. Would appreciate more info. If not, I am sure it is in Barry's book.  Art

Art,

The starting voltage is to high for modern can motors but with Legacy or TMCC I can start from zero

To do that you'll need a TPC or Power Master for each channel. Regardless, why run modern stuff conventionally at all?

I use common bus with no problem, with frequent out rail drops and the ground for accessories.  I have much larger "loops" than you do, and use one TIU channel for each  It's not hard to break up into 3-4 blocks per loop, and gives considerable opertaing benefits.  Just run a length of 14-gauge (what I'd recommend on a layout of your size) to each blocks and sever the center rail between blocks.

 

Can motors are not current hogs.  I don't know why you say they need high starting voltage,  You should have had the old Lionel Pulmor motors.

Iwill not be running modern in conventional. I do like to run conventional locos from time to time and when I run a modern conventional engine such as williams ( with can motor) the 5 volt start speed is about 25 scale mph. i AM NOW SPOILED WITH DCS and want to start with 0 volts so these engine won't jack rabbit.  Therefore, using my second gen ZW with built in power master- 180 bricks and legacy or tmcc I can start at a more realistic speed.  I have TMCC,Proto 2, Proto 1 engines as well as conventional.  The Proto 2 are my favorite therefore DCS is a must.  Thus my decision to set TIU in passive. Have not tried it yet but will try to hook it up this weekend. M any of you have much more experience in this than I do so I am asking do you see any problems with this arrangement?  again thanks for the input.

I should have said I will not be running command engines in conventional.

Art

One way to solve that issue is to stick a diode string in series with a motor lead in the locomotive.  A couple of 6A bridge rectifiers will drop that 5V starting voltage to around 2V and give you the slow start you desire.

ARt, if you have a Williams that starts that fast at 5 volts, you are inputting too much voltage to the TIU variable, resulting in a highly deformed AC wave.  Cut the input voltage wat down and I think you'll be happy.

I have a medium sized layout that is already setup with Lionel Legacy and TMCC.  I used bus wiring as recommended by Lionel, with drops about every six feet.  Power is feed to the track by the use of Powermasters and Direct LockOns.  I would like to add MTH DCS to the layout, but with the current setup is that possible, as I really don’t want to give up the Powermasters and Direct LockOns.  I was thinking that with my current setup the DCS system could be added but hooked up in passive mode as no track power needs to feed thru it.  Am I right in my thinking?

 

Thanks,

 

Ed Kempf

You can hook up in passive and it will generally work, but with limits:

You have a Buss wired Layout and no blocks.

All those drops send the DCS signal to the track, and the signals go both ways from each power drop. Now they are all crossing one another, this results in the tower of babel for electronic communication.

So you need to break the center rail between each power drop. That will keep the signals from crossing and all will be well. Isolate your sidings as well.

Be aware that passive does not put quite as much signal on the layout, but with a Rev L TIU (Current Version) you should be fine.

So I am a little confused. Where is star or home run wiring any different, the signal will still go both ways with each drop that goes back to the terminal block. Unless you want or need separate power districts there is no need for blocks, isn't that the beauty of command control; be it DCS, TMCC or DCC.

Get rid of the TMCC Direct Lockons, they will kill the DCS signal!  I had the problem, and then at our modular club, we had the same issue.  Yes, both locations had a Rev. L TIU.

YThe difference is that a blocked plan ensures that a DCS engine never sees more than 2 occurrences of a DCS command, and it only sees 2 when straddling 2 blocks.

This question is mostly directed to Barry.  I am confused about the various wiring methods. I want to understand why a star or home run wiring scheme is really any different that a bus wire with drops to it. A terminal block (such as the MTH 12 or 24 port version) is nothing more than a very short bus. Potentially all the drops that feed back to the terminal block can be equal if not greater than the length of a single bus with short drops to it. Please take block wiring as you describe out of the equation.  I am not trying to start an argument, just trying to understand the basic difference. My experience with train wiring comes with HO DCC, and we use bus wiring in that scale as a common rule.

 

Thanks,

Ed Kempf

I agree with you, Ed, on your comparison of the terminal block to a bus.  My 70 blocks are fed from a control panel, with each having a toggle switch.  Power from the TIU outputs (I use 6) goes from toggle switch to toggle switch, and from the other terminal of each 14-gauge wire runs out to the block.  Common is fed by a #12 bare wire that loops around the layout.  It works.

I posted the above question on Jan 18th, and only one reply. I am wondering why no responses, did I ask a question that was a hard one to respond to.

Perhaps because you jumped in the middle of another thread.  Try a new topic and see if you can get better replies.  The short answer is it's not the length of the wires, but rather the difference in the propagation of TMCC or DCC signals vs. DCS signals.  The frequency of DCS is 3.27mhz, and the wavelength is much shorter than either TMCC or even DCC signals.  This means that loops created by the track and multiple bus drops can create standing waves and garble the DCS commands.

 

 

Ed,

 

I've just returned from out of the USA.

 

The answer is that you have to think data packets sand not just electricity. That's where is the fundamental difference between using a terminal block or a wired buss.

Barry, welcome home.  Hope trip went well and the ship was illness-free.

 

I think Ed's point is that physically the only difference between a terminal block and a buss is the length between the drops--in a terminal block, it's an inch or less, but in a box it can be much longer.  The terminal blocks we use are not single point, as where many wires are twisted together at a single point.  Ed is not addressing what flows through the cicuits, just the physical layout.

 

But now that you raise it, just what is a data packet composed of?  Is it not a flow of electrons transmitted in bytes of 8 bits each? 

 

 

Robert,

 

It's data, 0's and 1's, just like an Ethernet or 802.11 wireless network, at a specific frequency. Each string of DCS data packets has an engine address and a command.

 

The data packet replicates each time it's split and each resulting data packed is somewhat diminished/degraded by being split. The following is in somewhat non-technical terms.

 

On a blocked, star- or home run-wired layout, an engine typically "sees" one set of packets at a time, whereas on a buss-wired, unblocked layout an engine would typically see multiples of the same data packets.

 

The number of splits and the lack of isolation can cause DCS signal degradation, from the engine's perspective.

OK, I'm not trying to open a can of worms, but explain this in a little more detail. I am a retired IS/IT guy with lots of complicated networking under my belt. A terminal block is no better then a hub, plenty of data collisions.  That is why we went to switches, less data collision and higher speed. In my mind a bus is a bus. A terminal block, like I said is a short bus, the runs to the various drops can be greater than a standard bus with short drops.  Factors such as resistance and impedance can all effect data loss.   

Ed,

 

See my post above yours.

OK, based on this your argument, it really isn't star vs. bus, it really is blocked vs. unblocked. A blocked, bus wired layout will do exactly the same thing.  Before Ethernet became the dominate force, there was IBM's Token Ring, which was a bus system, in many ways better than Ethernet. I am not arguing that the blocked idea is a bad idea, I agree with you on that, it is just the delivery method that I see no difference in.

Ed,

 

I agree that the isolation is what makes the larger difference.

 

However, in order to isolate the blocks, one must wire the Hot wires as star, not buss. A Common buss is just fine for DCS, a Hot buss is, by definition, not isolated and will work less well than if wired using a star topology.

Ed, I think the key to what Barry is saying is his use of the term "blocked."  By this he means that the center rail of the layout is gapped in places, to create separate blocks, and each block should have only one feed, so that data packets don't hit the loco multiple times. 

 

Barry, I note your use of the term "frequency."  This indicates that the data is being sent via electrons at "high" (i.e., greater than 60Hzfrequency).  I note that this is indicated in the discussion of packets at www.inetdaemon.com.  Especially at:

http://www.inetdaemon.com/tuto...ignals/signals.shtml

 

(This could lead to some interesting discussions, possibly beyond the scope of this thread, about the inter-effect of a 60 Hz train power current and a DCS signal, both going over the same conductor.  For example, suppose I'm powering a loop of O-gauge track through, to use an extreme example, 22-gauge wire.  I will have a lot of voltage loss.  Now I add a DCS signal, microscopic by comparison in both voltage and current.  What is the effect on the strength of the DCS signal of its electrons also having to squeeze through the wire??????)

 

As you know from prior discussions and you having seen my layout, I use a common U-terminal bus all around the layout into which everything--track, accessories, lights--is connected.  Hot comes from the TIU into the control panel, where it splits as it passes each toggle switch.  In effect, the control panel is a giant terminal block, branching every 2" to 10" depending on the distance between toggle switches, from which one drop per block goes to each block.  ED is asking what is the difference between this and a terminal strip which branches every 1" or so.

 

As I think about your discussion of splitting, I'm inclined to agree, but I don't see that the distance between splits would matter.  Splitting would have negligible effect on a 60 Hz circuit due to lack of continuity on an unused branch.  But on a higher frequency circuit, like a DCS signal, it could do so because of capacitor or charging effect effect.  Maybe I have the term wrong but I seem to recall something that a high frequency circuit has some flow even through not a continuous circuit.

 

Edit: Barry, this was drafted before reading your immediately preceding posts.

.

A MTH terminal block is a bus. the hot side and common side are not isolated from each run that goes out from the hot side or the common side. In essence the same signal (data packet) goes out every connection to the terminal block. the only thing that would cut down data collisions would be a blocked system as you suggest, but a bus system with blocks is exactly the same thing. The topology, or delivery method is not the issue. The track system unblocked is a bus, by blocking you are isolating the data packets to a given block, but the reality is that all blocks still see the same data packet, since the terminal block does nothing to prevent this. What I don't know is how the transmit and receive work in DCS. In Ethernet it is two separate wires, if DCS worked the same way then a truly blocked system would be both the hot and the common.

I'm guessing that there is no collision detection (like in an Ethernet network) in a DCS "network". True?

 

Following up on this, I have a different perspective on the whole DCS signal thing.  It often gets discussed as star vs. hub, when in reality the important part is segmented, isolated blocks vs. continuous loop.

 

It seems that if you set up a star but don't cut the center rail between drops, you're going to have issues.  This is an AHA! moment for me.  Don Masso and I have spent lots of time experimenting on his large layout and have had issues with DCS signal all over the place.  But, I believe he has drops, but a continuous center rail.

 

I'm planning a small DCS tinplate layout.  I set up a loop of K-Line SuperSnap track and put a DCS engine on it.  The engine works well, but sometimes I get "Engine not on track" or "Check Track" messages.  On a 6' x 4' loop.  Brand new track.  Almost new engine.

 

I'm wondering if a cut in the loop opposite the single feed would be better.

 

 

Yes?

 

Ed

Worth a try, Eddie.  Let us know the result.

Seems like a packet-routing switch between the TIU and track segments would be a great new product for MTH to offer. 

 

Ed

Eddie, there is really no point in isolating the outside rail, since trains will cross it, and as they cross, depending on how the individual cars are constructed, you may get repeated making and breaking very low voltage ground circuits (i.e., making minute sparks).

 

As to a packet routing switch, I don't understand what it would accomplish, especially in my case where everything works now.

Ed, I can't see this being an issue for a 6x4 layout.  Barely bigger than starter set oval.  I would think pin connections and clean track are more important. I use Super track, but I added pins for a more permanent connection and better conductivity than just the spring clips.  G

Scott, I would suggest severing the center rail between drops.  Don't compromise on proper power distribution.  I have many short blocks, each separately fed & severed form the rest of the layout.

 

GGG, overkill yes, but a good experiment.  Related is the question of the effect of varying lengths of track between the gap (different size loop)

Ed,

Yes?

Yes, that way the engine only sees each command once rather than  twice.

New improved diagram...

 

 

Interesting Barry,  that might explain why the signal seems to be better when the engine is equidistant from the power drop.  When I tested, it seemed like the signal strengths were lower when I was real close to the power drop, which would make sense because the signal difference delay would be greatest to each side of the power drop.

 

G, this is brand new track ,with pins.   I didn't expect and error messages from DCS on a simple, clean loop.  The engine works OK around the loop but I may open up a gap to see if it eliminates the error messages.

 

Ed

You would think a perfect dcs track signal on a 6 x 4 loop. Somethings not quite right. 

If you're testing  try joining the outside rails with alligator clips. Did you try a track signal test on the loop?

 

I suppose  remote /tiu/ communications could also come into play,

 

Anyway  we  figured if we  can get a good signal on one block. Hook up  another and then another . (test as you go) until you end up  as  in our case with 4 tiu in super mode with about 6 or seven scale miles of track.. We have a great dcs signal  over most of the layout.. 

 

 

 

Gregg, I don't think it's that simple.  Also, we don't know the actual readings, cleanliness of track & rollers, gauge of wiring, etc.  Size is not necessarily a determinant.

You're right we don't know. Looking forward to the test results. One loose connection can throw a curve into the whole works.