Common Rail Wiringgreenspun.com : LUSENET : Wiring for DCC : One Thread
I do not understand the recommendation to avoid common rail wiring. When a locomotive set crosses the boundary between boosters, the track in effect becomes common rail. This is unavoidable, and if DCC has a problem with common rail wiring then it would seem that a loco set would be unable to pass from one booster to another without a special electrical transition track section. One local club installed a special (and expensive) set of optocouplers at each transition to take care of this, as recommended by Digitrax.
Is this a problem peculiar to Digitrax? There are a lot of weaknesses in this system, which is why I went with NCE.
-- Norman Beveridge (NBeveridge@aol.com), April 11, 1999
I recently tried to make a drawing that would make the common rail problem crystal clear. I am finding it easier said than done. Here are two web pages that attempt to make it clear. Maybe they will click with you.
In a sentence, the problem with common rail is that if one booster is reversed, either because it's control lines are reversed or the output of the booster has autoreversed, the two boosters output's will add together. This would apply double the track voltage to a decoder and blow the decoder. This problem is most likely with locomotives that have offset power pick up rather than all wheel pick up. Dirty pick ups or dirt on the track and a good dose of "if it can go wrong it will" usually helps to facilitate the disaster. While I am having trouble making a crystal clear drawing of the failure scenario, there is no doubt this is a problem because numerous people have blown equipment this way.
One thing that can be easily explained about this situation is called ground loops. This is where you have two different paths for ground currents to flow. In this case, one of them is in the Loconet cable. The other would be the layout common rail. If the two grounds don't follow the exact same physical path, then noise from flourescent lights for example can cause erratic behavior or cause something to blow. Since the Loconet cable is likely to be running along your facia, and the layout common rail follows the track where ever it goes, you definitely have this problem.
The problem is basically unique to Digitrax because they do not use opto isolated outputs like most of the other manufacturers.
How do boosters work at all without an expensive transition section? A "virtual connection" within the boosters. This is more electrical engineering smoke & mirrors. I am also trying to think of a crystal clear way to make this point.
You are right that a common of some sort must exist somewhere. That common is the Loconet ground inside the phone cable. The boosters reference their outputs to this ground. Therefore the track power outputs of the boosters work as if they were connected together. However, the currents flowing through the locomotive do not flow through this virtual ground connection. This is where the smoke and mirrors comes in!
Digitrax can install opto isolators in their boosters upon special request. Opto isolators are inexpensive and I didn't think Digitrax charged much for the conversion. I'm curious to know if your local club did something different.
From what I hear NCE is a nice system. It wasn't out when I bought my Chief.
-- Allan Gartner (email@example.com), April 11, 1999.
Don Crano was kind enough to send me more on this topic. Here are his words and links to his pages and others.
It has already been proven, just by the number and fact decoders have already smoked, blown holes in loco shells, etc. And for me blowing up a CTC16 reciever used with CTC and DC on the same layout.
Now couple things to get started on the right foot. Lets drop the name common rail, as it is not proper only descriptive and adds to the confusion on this. Lets call it by the proper name, common return, because it is a common, but is after all power supplies and controls so it is a return.
Next does not matter if boosters are isolated at their buss connection or not, the princable of series and parallel operations still maintain. Object of isolated inputs on boosters, is cheap noise reductions, but almost always forced common return to be used. There needs to be a common for locos to cross multiple power sources, or it has to create parallel supplies as it crosses boundaries.
Now for examples lets combine some of what I have already done on this, and may be make it a little clearer. First look at this drawing, it will show how it happens to the loco. http://pages.sssnet.com/donc/commons.html#CommonRail
I know just realized, I should change the name common rail also. Any way you know see the possition and configuration of the loco, and how the supplies boosters, what ever are opposing phase/polarity. Does not matter if we are talking DC or DCC, or combo, it is the same.
Now one has to ask how does this configurement happen. It is very simple, think of a normal split phase 220 volt house drop, or if easier a center tap transformer, or is still easier two batteries in series. Any time voltage is measured on power sources in series, which all the above are, including common return. The voltage across both ends of the source is that that is added together, whenever the center tap is not used. And this voltage is dependent on the polarity or phase of the two sources versing the common, or center tap.
Now how does it happen on our layouts, look at this for a typical example. http://pages.sssnet.com/donc/controls.html#rail
You will notice it is very common to have added voltage across a single gap, this has happened for ever, any time two cabs of opposing blocks were set to opposite directions with common return. Now we all know this is not a major problem, with this config, the added voltage being across the same rail, either the loco would have to derail and some how have each side of the pickup across the gap of a single rail. Such a possible might be on a dual main with a derailment. Or possible derailment on a cross over. Not much concern, just a possible for Murphy.
But if we move down to the reverse section, this is reason for getting away from the term common rail and using common return, you will notice this same configuration now results in the added voltage being not only across gaps, but on opposite rails. And no matter what you do, on the reverse section one end will be proper or added voltage across a single rail. The other end is either improper, or in the latter case above, added voltage across gaps and rails.
All this is because, granted the common rail does not continue through a reverse section as descriptive, but the common return does, it has to or again it would require the loco to parallel the supplies to get across the gaps. What actuall happens as pictured is the common gets swapped from on rail to the other so proper polarity/phase can be had to get in or out. And it is this fact that always causes the other end to be wrong or opposite.
We can ask how does this happen, series supplies. Ok as pictured in my pages, if the reverse mechanism is before the common, then the supples are switched from parallel or series, based on the action of the reversing mechanism, I say mechanism, becuase it does not matter if we are talking boosters, DC cabs, or even playing with Dcell batteries, it is all the same. When the mechanism is before the common, or back to center tap, Went both say negitives are connected, the supplies are in parallel as related to the common. But if the reverse mechanism is changed before the common on one supply only, now the common is connected + to -, or in series. With out reference to the common, the power sources add together. Again this is typical of DC cabs, DCC, AC, anything that is phased or polarized. This is why for a typical DC cab layout with a reverse section, the AUX reverse mechanism is used, it puts the reverse mechanism after the common. Porblem solved, unless some on using the reverse section, uses the reverse or East/West switch in the DC cab. Now the polarity was changed back, even though the AUX had it correct. Chalked up as bad operating practice, no harm done.
Now lets move to DCC or command control in general. Same thing applies, only problem is, we can now use DC and Command control together. And worse yet, these supplies of isolated input boosters, that again almost force common return, supply autorevrse boosters. Thinking for a second, where is the revrse mechansim here in relation to the common return. Yep as far infront of the common as it gets. And if we add DC operations with DCC it does nothing but get worse.
Ok we now know how it gets setup, I hope
. And going back to the first link above, we see a visual of it. Now we sure can say it can't happen, then the locos wheels cross the gaps it will short the supplies. And this is very true. But we all know what dirt can do, so a short is not a given all the time. But lets add one more major factor to the mix. What happens when a booster shorts, it says opps and trips out. Now when this happens what does the loco do, it stops, but more then likely after a short coast or so. If all wheels are not picking up, either by offset pickup design, or dirt, and is coasts to as stop across the gaps, guess what happens when a short is removed from the boosters output, yep it comes back on. One decoder smokes. This is what happend to Wayne Roderick's I beleive Kato or Atlas RS, it even burn a hole through the shell. This also happened to be the second time it happened to him. Now Waynes problem was compounded by the fact the end of the section happened to be under DC Cab power, and the added voltage were way up there. Bill Rankers happened to be booster to booster, no blown shell, but it melted the shrink rap right off the decoder. Al Silverstein's was in N scale, booster to booster, a little less voltage I guess, it only blew a hole in one area of the shrink rap. My self it happend to a CTC16e receiver crossing over to a DC block, my fault, but the last of the common return was in 1985 for me.
You can look at my software simulation screen clips for actual setup and scope clips of it in action. http://pages.sssnet.com/donc/simulate.html
Or you can read about it Waynes experiance with it on his site at: http://www.ida.net/biz/tetonsl/railroad/DCChome.htm Keep in mind Wayne is holds a P/EE and it took it twice to happen to him before he figured it out.
Keep in mind, it is not that common return will not work safely, we all know there are plenty of layouts that use it. If there are no reverse section, it more then likely will never happen. If there is a reverse section and an auto-reverse booster is used, the odds are it may happen, and if DCC and DC are combined on a layout, the odds just got greater. With DC, it justs used to be called bad operation practice by the old pros, with DCC or onboard electronics we call it expensive.
If common return is used, then one has to make sure that all reverse mechanism/s are after the common return, this includes and reverses in cabs being disabled. And auto-reverse boosters are a no-no. Other wise as Wayne says, put all electronics after the common, and the problem not only goes away, it is impossible for it to even happen. This means the common is at the power supply, obsevering polarity/phase, or like Digitrax and EasyDCC, place it at the - side of the H bridge input.
I really do not think I can do much more in the drawing department, or the explaining. But it is really easy to simulate with a couple batteries or DC cabs and common the outputs, add a revese switch, and place it before and after the common to see the results across output. And on reverse sections, loops, wyes, turntables, check the difference between having the common change polarity, reverse switch before the common, or swap rails, reverse switch after the common.
But I will see if I can come up with something clearer then I already have done??
But the basis and basics are very simple, when using standard modes of operations of parallel and serial. Serial equals added voltage, same current. Parallel equals same voltage, added current. Two power sources, with the same phase/polarity with a common from - to - will equal parallel operations. With common connected to + to - will equal series operations. Any time the reverse switch/mechanism is before the common, the power sources are + to -. How this makes it to a decoder, just think of a reverse section, onside will always match, the other will not. No way around this fact. And if the power sources are in series, any time the pickup can be across the gaps of the wrong end, if all wheel pickup or offset as steamers match and do not contact the common, the pickups will see the sum of both power sources together. If the common is off the layout, and not used as a return, all reverse switches/mechanisms are automatically after the common, and only possible at all is parallel operations, and if all supplies are connected to same common, no voltage can ever be higher then that of the single highest power source.
Linn Wescott in 1950 did a nice job of explaining it in 'How to wire your model railroad'. And this is where I coin the phrase Bad operating practice. And as he notes, it is not a major problem for our locos. But at that time our locos did not have onboard electronics so it was not then, but to me it is today.
Also if you look my pages over then compare it to Stan's At: http://www.tttrains.com/dcc/commonrail.html
After reading through all the mismash, and such on both our pages, you will see we are both trying to show the same thing.
-- Allan Gartner (firstname.lastname@example.org), April 11, 1999.
A common is anything we want to make it and call it. As example, the wording common rail is used all the time, but it is properly called common return, this is because the common is after the controls and supplies the return path to said controls. And the fact is the common rail does not have to be physically continuous at all. Only that the common return comes back to the controls. Such as the case with a reverse section. Here the common rail, must be broken, but the common return still flows through the reverse section. Only difference is such as a DC cab control system, if there is a properly installed AUX reverse switch, it will be placed after the common return, thus the common rail is swapped from side to side. If there is a reverse switch at the control, DC cab, then the common return is not swapped, only the rails polarity is changed. Once this happens, then two power sources are placed in series operations mode, and their combined voltages add.
Also when a loco crosses gaps between boundaries, it really does not great a common, it actually places or forces two boosters into parallel mode of operation. And if there is already a common return, even with a common rail, that happens to be in series operations because of polarity or phase differences, this will show as added voltage across the gapped rail, the force causes a short.
BTW, this is really nothing at all to do with just DCC, it happens with DC, DCC, AC, any type of power source. As example with multiple DC cab control, any time a loco over runs a block of another cab, with common return, and the other cab is set for reverse direction, the loco shorts out the series operations between the two cabs, when trys to force parallel operations crossing into the reversed block from the other cab. Take a volt meter, and measure the voltage across the single gap, with two power sources in opposite phase or polarity and you will see what I mean.
This happening across a single gap is not much of a problem, and only results in a short from the parallel force. But if this same thing happens at one end or the other of a reverse section, now this same added voltage, again reguardless of power source, DC, DCC, AC, anything, the added voltage is now across diagnally the two gaps. Again you can measure the voltage by probing across gaps of the opposite rails, just like a offset pickup loco would if the loco was on one side and the tender, or front and rear trucks, of the double gaps of the reverse section. One end or the other of the reverse section will be in this state, if there is a reverse switch or mechanism before the common return. This could be the reverse switch of a DC cab, or a Auto-Reversing booster. Both are in front of the common return. The answer here is with DC we used a AUX switch to swap the common rail from one side to the other. With DCC, do not use a Auto-Reverse booster with common return, use a reversing module after the common return, this then is the same as the AUX reversing swith of a multi-DC cab setup.
Opto-Isolated boosters do not enter in to this problem. Opto-Isolating the boosters input, is a form of cheap noise protection. The Optos only act to form digital information, and filter out anything else on the communications buss. Look at then a switch debouncer, or just a On/Off signal on their output. This also isolates the common between systems, so it forces another common to be used some were. Usually common return, because it already exsists. But on systems like NCE, you will also find there is a place to by-pass the isolation for commons other then common return. Reason is here, if the common is such as a System common, like Digitrax or EasyDCC, or a power supply common, then the common is not a return common, and all reverse mechanisms/switchs are after the common automatically, and there is no possible way for power sources to get into series mode of operations. And if all power source commons are connected to the same place, then the highest possible voltage availible with be that of the single highest voltage source. There are also other issues of earth gounding for verious reason, such as static discharge paths, etc. But never to this with common return.
As far as Digitrax, and opto-isolated booster, not sure what they had done, but it is not expensive to have from by Digtrax. Have not checked in awhile, but they have done it from free to approx $10.00 per booster, I would not call that expensive. But also the last time I checked, they had had only 4 requests to do it.
As far as the NCE, this a great system, especially the PowerHouse Pro, have used it and like it very much. But as to the weaknesses of the Digitrax, you will have to point them out to me, have been using it for years, and not found any system I like better, or with as much power as the the Chief has.
Remember Always Have Fun and Enjoy!, Don Crano Akron, Oh NMRA #096211 mailto:email@example.com Visit Model Railroading with DCC at: http://pages.sssnet.com/donc/
-- Don Crano (firstname.lastname@example.org), April 11, 1999.
As already explained, common rail wiring is not necessarily the problem, only one of the several contributory factors. Doubled rail voltage can happen with several combinations of haphazard booster and reverser connections. It is not necessarily a problem limited to Digitrax. For another treatise with drawings, charts and pictures, see the 'Using Common Rail Wiring with DCC' selection, by Stan Ames at www.tttrains.com.
One way to protect a decoder is to add a bidirectional rated Transient Voltage Suppressor (TVS), or two zener diodes wired back to back, between the rail to rail pick-ups internal to each loco. Should start clipping at 17-18 volts for HO scale and be able to withstand 5-10 amps long enough to repeatedly trip booster current limit. About $1 each. See Microsemi brand in www.Digikey catalog, p/n P6KE(18 or 20)CAMSCT-ND.
-- Don Vollrath (email@example.com), April 13, 1999.
One of the other contributary causes of a problem is the use of a single ac supply transformer (14-18 Vac) to power several boosters. A booster has only an input rectifier and an output H-bridge in the power current path. Usually only one side of the internal rectified dc feeding the H-bridge switches has overcurrent measuring. So during operation there is a momentary current path through each booster from an AC (or DC) input terminal to one rail that is being switched, but with unmeasured current. When two boosters are connected to the same transformer secondary winding it is possible (and according to Murphy, probable) that the AC inputs to the boosters will be reversed. Now when those two boosters are connected to 'booster district' track rail, and even when their DCC signals are 'in phase' the transformer secondary voltage appears across the rail gaps. When a loco truck with multi-axle pick-up crosses the gap, a short circuit is formed. Current flow is limited only by the resistance of the transformer winding, circuit wiring, and wheel contact. If one attempts to use 'common rail' wiring between the boosters, you create a transformer short circuit directly through the boosters by eliminating one of the rail gaps.
THE RECOMMENDED SOLUTION IS TO ALWAYS USE A SEPARATE TRANSFORMER SECONDARY WINDING TO POWER EACH BOOSTER. No amount of opto-isolation of the DCC control signals will fix this problem.
The probability of this happening is greater when one mixes booster brands as there is no assurance as to the AC input phasing, nor of the location of internal protective current limit measuring. Even if identical boosters are used there is no assurance that the current limit function can shut down the equipment without damage.
-- Don Vollrath (firstname.lastname@example.org), April 15, 1999.
Norman and ALL,
There are three very informative and definitive sites on commons:
Don Crano's site with diagrams and explanations. Dick Bronson's site with good discussion of "Home Wiring Common' and links to some very nice diagrams. TTTrains site with Stan Ames' discussion, diagrams and 'formal' test set up.
These sites in combination offer the best of insight and clarity for the common problems. Note: You can use a common rail with in a power district for signaling activity (you'll have to have the power district boundry also be a singnal block sense boundry) - but you'll find that using the issolated sensing (Dick Bronson or Don Crano) will be better anyway.
-- Ed McCamey (email@example.com), April 19, 1999.