Question for Jim Banner (refreshed)

[Yampa Bob]

For the benefit of all the new members, I retrieved this from the archives for a fresh exposure.

Jim,
Had I known years ago that I would one day be working with GPS and digital command for trains, I would never have sold my oscilloscope.  Since you have one, I'm hoping you can settle my mind on a couple of issues.

I'm interested in how the EZ Command controller modifies the digital pulses on address 10 to allow operation of a DC loco with throttle control.  My theory is that the alteration must be one that is recognized by the DC loco, and rejected by any DCC locos also on the track, even though (or because) they are on different addresses.

Would it be possible for you to take pictures of two wave forms?  One would be of the usual DCC wave form sent to a DCC equipped loco, another picture of the address 10 form sent to the DC loco, if indeed there is such a modification.

Also, is there a current restriction that limits the use of only one DC loco on address 10, as stated in the manual?  You know what I'm thinking, running 2 low current draw DC locos in consist on address 10.

I have set up and used special encoders/decoders to send packets of GPS information via commercial VHF and UHF radio signals, to track search and rescue teams and plot their locations on a computer topo map.  Of course this is a different application, but when I run my DCC locos, my mind starts spinning nonetheless. (another reason I run my trains sitting down) 

Thanks in advance for your help. (when you get time) 

[Atlantic Central]

Bob,

I don't know for sure, but since DCC uses an AC signal, than rectifies it in the decoder, it would make sense that the analog address simply imposes a DC signal on the line.

I know from the few times I have operated my analog locos on DCC layouts, they still hum like crazy when you bring them to a stop. The DC signal may be linar or some sort of pluse, but it must be DC, which is not normally present with DCC.

I would suspect that the decoders of the other locos simply reject the DC signal.

Sheldon

[Jim Banner]

Bob, I would love to be able to do what you ask.  Unfortunately, it would require a memory 'scope to give a nice, clear photo of a single trace.  But my old 'scope, like my old head, suffers from a lack of memory.  But maybe I can explain anyway.

The DCC signal consists of square wares that typically swing from about -14 volts to about +14 volts.  It can be higher or lower, but is always symmetrical in terms of voltage.  This means that when it is rectified, it will give a constant output that does not need any form of regulation.  If the DCC signal were not symmetrical, or in other words, if dc were added to the DCC signal, then after it is rectified, it would not give a constant output but would vary as larger and smaller voltage half cycles were rectified.  Figure 1 below shows a symmetrical DCC waveform and what the full wave rectified output looks like.  Figure 2 shows a similar wave form with 6 volts of dc added to it and what that looks like after rectification.

 

The asymmetric waveform in Figure 2 gives a rectified output that varies with time.  The average voltage output is the same, but remember that for a fixed load resistance, the power output varies as the square of the voltage.  So the rectified output in Figure 2 is almost 20% higher in average power than the rectified output in Figure 1.  As we increase the added dc voltage up to 14 volts, things get even worse.  At 14 volts, the average power of the rectified output rises to 200% of that in Figure 1.  Bottom line, adding dc to the DCC signal is not a great way of running analogue locomotives on a DCC track.

To look at what does work, we first need to understand a little more about the DCC signal.  As we know, it if digital, broken up into zeros and ones.  But what exactly is a zero and what exactly is a one?  In Figure 3, below, we see firstly a string of ones.



A valid "one" is defined as a single square wave having a positive duration of 60 microseconds and a negative duration of 60 microseconds.  There are some tolerance qualifiers.  And if you stand on your head so that the single square wave goes negative first and then positive, that's okay too.

Next to the valid "one" in figure 3 is a valid "zero."  Here the positive and negative durations must be at least 95 microseconds long.  This difference in length or duration allows a decoder to distinguish between a "zero" and a "one."  But hold on.  Once you have defined a "zero" to be longer than 95 microseconds, then it could be longer than 950 microseconds and still be a valid "zero."  Or even longer than 9500 microseconds.  In fact, a valid zero can have either a positive duration or a negative duration of up to 9900 microseconds, which is about 1/100 of a seconds.

Now we can use these extra long "zeros" (also known as "stretched zeros") to operate a dc motor.  Basically what happens is that a stretched positive will make the motor go forward a little bit.  And the non-stretched negative will make it go back a little bit.  But as long as the positives make it go farther forward than the negatives make it go backward, the net effect will be motion in the forward direction.  We are all familiar with the idea of two steps forward, one step backwards.  But think of this as 10,000 steps forward, 100 steps backward, or as 100 steps forward for every step backwards.

With 100 steps forward for every step backward, we are cruising along at full speed.  Now stretch those "zeros" a little less.  We won't be going quite as fast, maybe only 10 steps forward, one step back.  As we stretch the "zeros" even less, we will finally get down to where we are going one step forward, one step back, in other words, standing still.  Incidentally, that is the noise you hear when you run a dc locomotive on a DCC track.  It is the motor singing
     "One step forward,"
     "One step back."
     "Please Mr. Owner,"
     "Get me off'n this track!!"

In Figure 4, we can see what happens when we rectify a DCC signal that contains stretched zeros.  Basically, nothing happens.  The dc output is still nice and constant.  So by using zero stretching, we can operate a dc locomotive along with our DCC locomotives and not have it affect their operation.  We are not necessarily being nice to our dc locomotive.  After all we are making its motor march on the spot, working hard without the benefit of a nice cooling air flow which a spinning armature would create.  No wonder dc locomotives left sitting still on DCC tracks overheat.

Just for the record, you can operate one, two, or more dc locomotives simultaneously using zero stretching, right up to the power limit of your DCC booster.  It is just that you will not have separate control over them, just as you wouldn't  if you were using a regular power pack.  So if you have some favorite dc locomotives that you always enjoyed double heading, by all means continue to do so - just don't leave them sitting stationary on the tracks for too long.

There are four other common problems with running dc locomotives on a DCC layout.  These include
(1) half wave lighting in DCC locomotives.  This is where you return the lights to a pickup (either left or right) instead of to the decoder blue wire.  The lights get brighter or dimmer as you vary the speed of the dc locomotive.  Solution - avoid half wave lighting.
(2) slowing down throttle response time.  You will rarely see this on a small layout and probably not notice it even if it does occur.  But on large layouts with many throttles, for example, on club layouts, the command station may well be spending more time stretching zeros for the benefit of one locomotive than it does sending data to dozen of other locomotives.  Solution - no dc locomotives on the layout.
(3) shorting the tracks causes loss of programming in decoders.  I suppose breaking up dc (during zero stretching) is more likely to produce valid but unwanted commands than breaking up short pulses into even shorter pulses.  Solution - no dc locomotive on the layout.
(4) mystery reprogramming of multiple locomotives - occurs occasionally  when the operator of a dc locomotive tries to program his locomotive in operational mode.  Or when anybody tries Ops mode programming and forgets to enter an address.   An ops mode command sent to address 00 (Bachmann 10) gets interpreted as a general command to all locomotives on the track.  I don't know much about this one - I came across it only the other day.  But the solution is obvious - hang, draw and quarter anyone who uses OPs mode programming without having a masters degree in it.


 

[Yampa Bob]

Thanks Jim, for the great explanation.  I suspected a regular scope might not be able to lock a single trace, but your diagrams are much better anyway.  Your  poem says it all,  for my "memory recall".  Hey, that's "poetic" also. 

It reminds me of a "Texas Two Step". Then you get a klutz on a crowded floor whose only "Ops programming", choreographically speaking, is "one step forward, one step back". 

This reinforces my policy of no dc locos on my dcc layout.  I figured one could run 2 DC locomotives in consist, but didn't want to endanger my remaining DC locos to try it out. I interpret a caution note as: Don't do it! 

If this question ever comes up again, I will definitely remember to post a link to this thread, and I copied it to my logbook.

Your efforts are very much appreciated.

[pdlethbridge]

Texas 2 step, is that anything like the Rochester quick step? Thats my dance to the john!

[Yampa Bob]

I think this thread deserves a "bump" for a fresh exposure.  Lots of new members since it was written. (Membership recently topped 4,000)

Congratulations to Bachmann Industries for a very popular board.

[pdlethbridge]

With members like you and Jim, it will even get larger.