G Scale Model Train Forum banner

1 - 4 of 4 Posts

·
Registered
Joined
·
2 Posts
Discussion Starter #1
I've had this problem for quite awhile. I've not read anything about this problem.

Here is some info

NCE power house pro radio setup. with 10amp boosters.

Stationary decoders from 2 different manufactures. NCE and DCC Specialties.

QSI decoders work great no problems.

DG583AR Decoders cause the problem.. When the Train is moving the stationary
decoders will not respond to any commands. If I stop the DG583AR Train I can
then operate the Stationary decoders with no problem. It works fine with
QSI decoders, only the DG583AR decoders bother.

If the DG583AR is on a differnt block from the stationary decoder then those stationary
decoders will work fine. But if on the same block and the Train is moving the
stationary decoders will not work until you stop the Train.

Anyone else have this problem? The DG583AR seems to be the problem. Interference maybe?
Motor noise?

I'm slowly changing over to all QSI deocoders but would love to find a fix to this is possible

Thanks.

Jeff T.
 

·
Super Modulator
Joined
·
20,521 Posts
You need to find Bob Grosh... I think he's on this forum.

He turns up on chat quite often in the evenings. He's pretty much my "reference guru" on Digitrax stuff.

Regards, Greg
 

·
Premium Member
Joined
·
411 Posts
Jeff,

Is the DG583AR in a Aristo locomotive?? If it is I would check the wiring & make sure that every thing is isolated.. You could have a cross connection.. This could put a signal on the track that you do not want.. If the wiring is good, I would look @ getting the decoder checked out.. Could be a short in the decoder that is putting noise on the track by backfeeding.. Hope this helps..

BulletBob
 

·
Premium Member
Joined
·
254 Posts
OK, here is the scoop.
(AS BEST AS I CAN TELL)
The Digitrax Decoder, as well as SOME other decoders, use Schottky diodes in their bridge rectifiers.
The bridge rectifier is the first thing the DCC signal sees when it enters the decoder. It changes the DCC signal into a DC voltage to provide the power for the lights, motor and decoder electronics.
Unlike regular diodes in bridge rectifiers, a Schottky diode does not drop 0.6 volts per diode. A bridge rectifier will drop 1.2 volts off your track voltage, leaving that much less voltage to run the motor. Schottky diodes drop less voltage, (that is, the have a much lower forward resistance) and therefor deliver more voltage to the motor.
Also, Schottky diodes switch on and off much faster. As the DCC signal changes from one polarity to the next, the Schottky diodes switch on and off WAY faster, so,,, the entire pulse gets converted to DC, not just part of each pulse.
Here is the benefits you will see if you use a decoder with Schottky diodes:
There is less resistance, so, there is less variation in the speed of the loco due to grades, curves or length of train.(Even with back EMF off.)
A higher voltage pulse, with less impedance will produce smoother slow speeds. (even with BEMF off)
Because Schottky diodes switch on and off faster, the also deliver a less distorted DCC signal to the decoder.
Now the problems....
Many DCC manufacturers have warned years against putting capacitors in locos. Capacitors cause decoders to mot be able to read back CV's, Write to CV's, distort the DCC signal, cause booster shutdowns, prevent systems from starting up, cause locos to draw more current, and in some cases can blow decoders.
Remember, a capacitor, when it sees a square wave acts like a dead short until it charges up.
Even a small capacitor for RF noise suppression across a motor can cause a motor seem to draw a lot more current, in some cases several times what the motor is actually drawing.
A large capacitor, (say 22,000 micro-farads) can put a huge strain on a decoder's bridge rectifier when power is first applied. Not just when you flip on the main supply, but every time the loco regains contact after hitting a spot of dirt.
-------------------------------------------------------------
OK, so what does Schottky diodes have to do with capacitors?
-------------------------------------------------------------
Let's see if I can find an analogy.You booster is a big tow truck
Your loco is a VW bug stuck in the mud.
A regular diode is a old rope. You can easily snatch the bug out of the mud with your old rope and big tow truck.
A Schottky diode is like a new steel tow chain.
Putting a 22,000 micro-farad capacitor in you loco is like adding 22 tons of gold bars to the VW beetle.
Now if you snatch the VW, something is going to break.
Some people will be quick to point out that the output of the decoder's bridge rectifier is DC, not a square wave DCC signal and therefor the capacitors will not cause any problems other than when DCC is first applied.
The problem is, the DCC signal is NOT a perfect square wave. Each time the DCC signal flips polarity, it does NOT do it INSTANTLY. The output of the bridge rectifier slopes down to zero and then slopes back up to the track voltage. If the capacitor discharges, even a little, and it will if the decoder is driving a load, then the diodes in the bridge rectifier have to charge the capacitor back up. That happens at every transition of the DCC signal.
IF the diodes are regular diodes (like the old rope) they limit the rate the capacitor charges. (They sort of stretch,,, well, actually, they have a resistance that does that.
IF the diodes are Schottky diodes, they try to charge the capacitor back up instantly. That is to say, they will instantly draw all the available current from the rails that they possibly can. Normally, this is not a problem. The time that it takes to "flip" the DCC signal is so short, compared to a "1" or a "0" that the average current increase caused by this fast charging is only a tiny fraction of the total time required to send a one or a zero. However, the resistance of the rails, impedance of the booster and power supply, algorithm the booster firmware uses and several other factors can cause booster shutdowns and failures. (I won't do into it here, but hook a current measuring scope up to a Bachmann K27 with its TWO huge capacitors and you can see why it shuts down boosters when a Schottky diode equipped decoder is installed.)
So, what happens when there are big electrolytic capacitors in an engine that are driven off of the decoder's bridge rectifiers?
Well, If we look at it with a current measuring scope, we would see that while running a typical loco it might draw 1 amp during the ones and zeros of the DCC signal. But, during the transitions, the decoder can draw short spikes of current that even exceed the "AVERAGE" current capacity of the booster. In fact, it will do that thousands of times per second anytime the diodes in the bridge rectifier of the decoder are FASTER than the tracks DCC signal transition time.
Now, are your stationary decoders edge detecting, or pulse width detecting? The spikes can have different effects depending on the algorithm the decoders use to decode the DCC data.
Since you did not say what loco you put these decoders in, it is hard to guess what to correct, However, if it is an Aristo loco, you will need to AT LEAST cut or bend one pin (either 6 or 7, depending on the production run) going into the decoder. For newer RS3's you will need to remove the switch board and wire up all the lights to the decoder and install an old 18 volt smoke unit element (use the ones sold for the original Aristo Pacific). As to other Aristo locos, the fixes differ.
Basically, the loco is garbaging up the DCC signal. Since stationary decoders are only sent 1 to 4 times (depends on system) they are the most effected. You may also see problems with missed functions (above F4) doing to the same or other locos in the same power district. This is only noticeable with functions above F4 and really a problem with functions above F8.


OH!
By the way, I have 5 Aristo locos with DG583 decoders. I gutted all the Aristo electronics in then and did a normal wired install. They cause absolutely no problems for my 30 or so stationary decoders. And, they program and read back CV's just fine on four different programmers I have used. All told, I have close to 70 Digitrax decoders on the ALLY if I count all my cars and locos. Even with all of them on the track, I have not had a problem with sending commands to turnouts.
 
1 - 4 of 4 Posts
Top