A weird problem...and a lesson learned

[grande man]

Hi everyone.

We recently had an unusual DCC issue that might be of interest to members here. For several years we have run two Lenz LS150 stationary decoders to control the turnouts in our three hidden staging yards. The layout runs on a Digitrax DCS100 Super Chief system. Switch routes programmed into the system control turnout interlockings to get trains safely out of the staging yards and onto the main layout.

The LS150s have always been on a separate power supply from the DCS100 and the entire system has been very reliable. The layout has a few twin coil switch machines in view blocked areas that have fascia board manual switches (Atlas) that are powered from the same power supply the LS150 decoders are on.

Yesterday one of the manual switch buttons stuck and toasted a switch motor and the associated wiring (since the power supply didn't have overload protection). It also nuked the power supply. We replaced the switch motor, wiring and control button but didn't have a spare power supply. I came up with the bright idea to run the LS150 decoders and manual control switch motors on the same power supply with the DCS100 DCC controller...big mistake. Trains still ran ok for the most part (we did have some erratic behavior out of our lone QSI equipped unit), but the switch routes for the staging yards became very intermittent in operation. What was interesting is that the system could be powered down and when it was brought back online the switch routes worked fine for a short period of time and then began erratic operation again. It was completely erratic as to which routes would or wouldn't work...

Troubleshooting logic always tells us to go back and look at anything that was changed to isolate a problem, but the changes made didn't seem like they would have caused an issue. After searching the Yahoo Digitrax users group I came across a reference to Loconet corruption caused by, you guessed it, some stationary decoders being powered by the same power supply as the main DCC system.

Today we built up a 12 VAC power supply from Radio Shack parts and put the switch machines / LS150s back on their own power supply. It fixed the problem! I'm not sure why the network data was being corrupted by this, but it obviously was an issue. The new power supply is a verified fix.

Live and learn. Hopefully our experience will save someone some aggravation in the future.

 

[picsntrains]

I'm guessing it might have been a ground-loop. Depending on where the power supplies were plugged in and how they were grounded, you can actually end up with different voltages on the grounds! This difference in potential can induce noise on signal lines. The new AC power supply is isolated from the house ground so no ground-loop.

Ground loops are commonly seen in audio systems as hum. The solution is to run the same electrical circut to both locations or isolate one end from ground.

To check, put an AC voltmeter between the two grounds and see if there is maybe a volt or two difference.

Chuck

 

[hamltnblue]

Looks like what happened is that you were getting kickback voltage.
When a coil charges it stores energy. When the voltage is removed the coil field collapses and the coil releases it's energy back into the wiring. Sometimes the voltage can be much higher. If DC voltage is used to supply the coil all you have to do is place a diode across it in reverse direction. When the voltage is removed the diode shorts the kickback voltage. This does not work with AC.

By the way, you shouldn't power the switch machines with both decoders AND manual switch control directly. That in itself is looking for trouble.