DCC Snubber

[bcjim]

My friend who uses NCE was showing me a snubber schematic. Then I see it in the Sept. M.R.
The cap is mis labled should be 50-150 volts not ohms.

I am using Digitrax radio but the brand is irrelevant.
I have now installed this on my layout.
Anyone else bothered to assemble and install one?
Any point?
Jim in Kelowna B.C.
   

[Jim Banner]

Some systems ring worse than others.  I checked the ends of my 30 foot buses and the spikes were less than 30 volts, so I didn't worry about snubbers.  I don't know if this was because my buses are all twisted pair with 3 or 4 turns per foot, but I have never had a problems that could be related to ringing.  I also use Digitrax with radio throttles with DCS200 command station and PM42 power manager.

Jim (in Saskatoon)

[richhotrain]

My friend who uses NCE was showing me a snubber schematic. Then I see it in the Sept. M.R.
The cap is mis labled should be 50-150 volts not ohms.

I am using Digitrax radio but the brand is irrelevant.
I have now installed this on my layout.
Anyone else bothered to assemble and install one?
Any point?
Jim in Kelowna B.C.
   

Jim,

I operate an NCE PH-Pro 5 amp DCC system to run a continuous loop 160 mainline.  On the Wiring for DCC forum, the subject of snubbers recently came up.  I decided to add snubbers every 30 feet around my layout, and I have noticed an improvement where before there was a noticeable slow down in the speed of the locomotives at the more "remote" areas of my layout.

Following the recommendations on the forum, I went to Radio Shack and picked up some 100 ohm, 1/2 watt resistors and some 0.1uF, 50 WVDC capacitors.  The whole lot cost just under $10.

Rich

[Jim Banner]

PD, you might want to skip this one!!

Rich,
I am glad the snubbers have solved the problem for you.  As I said before, I do not have a problem so have not had to use them.

I am wondering what "a continuous loop 160 mainline" is.  I am also wondering if your track and power bus form a closed loop, and how you run your bus.  Let me explain.

Two wires run side by side form a transmission line.  An example was the old, flat, twin lead television cable.  It had a 300 ohm characteristic impedance which matched both the impedance of the source, typically a folded dipole antenna, and the impedance of the load, typically the input of your tv set.  With everything matched up, there were no standing waves on the line and the line appeared neither inductive nor capacitive.

Now compare that to our DCC.  The load is far from constant and the source (our command  station) does not match the impedance of our transmission line.  The bus conductors that make up our transmission line often have some space between them, and we may have several sets of transmission lines running parallel to one another, particularly if we use centrally located power management.  The result is that our lines tend to be inductive.  This means that when we have fast voltage transitions on them, there will be inductive surges or spikes, which show up as overshoots on the voltage transmissions.  Another way of saying that is that the inductive transmission lines emphasize the high frequencies.  The result, as far as your DCC is concerned, is voltage spikes that make it harder for your decoders to read the digital information and may even exceed its voltage rating.

So what can we do about it?  Adding some fixed load at the end of each power bus is a first step.  This helps absorb some of that high frequency energy, reducing the overshoots.  A next step up is to add capacitors along the line to try to balance out the inductance (i.e. to "tune" the line.)  However, as we well know, capacitors can look like a short circuit at high frequencies (that is why we have to cut them out when we use ultra sonic decoders in B-mann locomotives.)  So we have to add a resistor in series to limit that high frequency current.  Thus the "snubber."

There are other things we can do as well.  Since transmission line inductance increases as the area between the conductors increases, and we cannot reduce the area by reducing the lengths of the conductors, we can only reduce the inductance by keeping the conductors closer together.  This is one of the reasons for twisting the conductors together - it both reduces the area between the conductors and increases the capacitance between them without having to add physical capacitors.  The second reason for twisting the wires together is that it reduces mutual inductance (cross talk if you prefer) between sets of conductors.  This becomes important when there are multiple transmission lines from a single source running together.  There is yet a third way of reducing inductance and that is to avoid large, closed loops of track fed by a large, closed loop power buses.  These large loops can act as large, single turn coils, or inductors, and cause problems.  We can reduce this inductance by twisting the bus wires together but twisting the track may interfer with the trains running on it.  So we need to do it another way.  One way is to run half of the bus half way around the circle in a clockwise direction and the second half of the bus half way around the circle in a counter clockwise direction, but not connect the ends.  The track should be similarly gapped at the break in the bus.  This may work or it may not.  The problem is that when a train bridges the gap in the track, it connects the track into a loop and also connects the ends of the bus together, making another loop.  A better way is to divide the layout into four or more sections, using gaps between sections, and run two parallel buses each way from the command station, one for each section of track.  Multiple drops can and should still be used in each section.  That way, bridging the gaps will not create a closed loop, either in the track or the buses, even if two trains happen to bridge two gaps at the same time.

My H0 layout, with over 200 feet of mainline in a large loop is wired this way, with each of the four buses connected through a centrally located power manager.  The power manager limits current current in each of the four divisions ("subdistricts") of the layout to 5 amps and also prevents a short circuit in one area from causing a power loss in the other three.  It was originally wired as 25 blocks, then the blocks were consolidated and connected to a single bus.  This occasionally gave problems, particularly with MRC decoders, and particularly in one area.  Unfortunately, I did not have a high enough frequency oscilloscope at the time to check for overshoot.  With the present setup - twisted buses, four subdistricts - the overshoots are less than 5 volts on top of a 15 volts p-p signal and cause no problems whatsoever, even with MRC decoders.

I don't know if .1uf every 30 feet is right or not.  My gut instinct is that it might be enough capacitance to cause undershoot (rounding of the rectangular wave corners) but my brain never believes my guts unless they have hard evidence, like a 'scope trace, to back them up.  In any event, a little undershoot is less likely to give you problems than a little overshoot, unless you are using Lenz Gold decoders.  These, if I understand correctly, look at transitions rather than levels when decoding the digital signals.  (where is Stan Ames when we need him?)

Jim

[Mike]

Oh my! Where 's mr. p lethridge when we need a comment?

[Jim Banner]

Thanks, Mike.  I meant to add a warning for him.

Wave guides, transmission lines and Smith Charts are part of the Black Arts of Electronics and are usually practiced in dank cellars after midnight.  I hope nobody gets Slurpee Head from reading it too fast.

Jim

[richhotrain]

PD, you might want to skip this one!!

Rich,
I am glad the snubbers have solved the problem for you.  As I said before, I do not have a problem so have not had to use them.

I am wondering what "a continuous loop 160 mainline" is.  I am also wondering if your track and power bus form a closed loop, and how you run your bus.  Let me explain.

Jim

Jim.

I have a 25 x 40 foot P-shaped layout with a double main line.  Although there is a yard, an engine servicing  facility and other types of sidings, the double main line forms a continous loop around the perimeter of the layout.

My bus wires are 14 gauge and run, untwisted, completely around the layout and connect back at the command station.  So, I call it a 160 foot continuous loop.  

As engines get farther away from the command station and booster, I have always noticed a slight slow down in speed.  So when I learned about snubbers, I tried them and the slow down in speed has ceased.  Hope that all makes sense.

Rich

[richhotrain]

Oops, I just noticed that in my original reply, I inadvertently left out the word "foot" when I referred to my continuous loop 160 foot mainline as a "continuous loop 160 mainline".  Sorry.