BLI SW7 Current Pulse Investigation

by Larry Maier, 03/05/05

Buy Broadway Ltd Locos Here
Switchers Reversing "Load" Issue
More About Switchers

I received the GN 163 and C&O 5236 BLI SW7 engines. Out of the box, both appeared to work on my layout. As I ran 5236 more extensively, I had frequent problems tripping the On Guard breaker that feeds my layout. This happened particularly when I reversed the engine while running. I was able to cure this problem by using Deoxit on the metal bars that serve as the power pickups from the main axles. My guess is that poor electrical contact may have caused enough charging/discharging of the power supply capacitors that too much current was drawn. After treating 5236, I was unable to get either engine to trip the On Guard breaker by either normal operation or by reversing the unit while running (not normally done).

I moved both engines to my workbench where I connected a test track to a Digitrax DCS 100 with a DT100 throttle. I measured the current drawn by inserting a 0.1 ohm resistor between the DCS100 and the test track. Both engines showed very high current draws when reversed while running. The operating speed made little difference. It appeared that as long as a speed step above 0 was used, reversing the engine caused a large current inrush. The figures below show typical results for each engine:

Note that the background “fuzz” in the trace is actually the varying current drawn by the sound system.

These views show the the worst case observed for each engine, but it is easy to see that the two engines are about equivalent. I did the rest of my testing with GN 163. Reversing an engine while it is running is a worst case scenario since the motor rotating in the “wrong” direction acts as a negative impedance and draws more current than it takes to simply start the motor turning from a stop. I tried a normal stop (i.e. Speed step = 0), reverse, start sequence and recorded the current when the engine started to move. In many instances, there was no excess current drawn. In some instances, however, I observed a significant current draw at the start. Again note that for this test, I made sure the speed step was 0 before reversing and starting the engine. I never observed a current inrush during the reversing operation, but sometimes, there was a current inrush when I started the engine moving after reversing it. I did not observe the current inrush after the first instance. In other words, I could start the engine any number of times after the reversal, and only the first time after the reversal showed the inrush, all other starts showed negligible current increases. I should also note that I did not have a single DCS 100 shutdown during this phase of the testing. The oscilloscope trace below shows typical results from these tests. Note that the time scale is changed from 10ms/div (above figures) to 25ms/div (figures below). The trace below represents the worst case (i.e. the longest) transient that I observed during these tests. Again, I was unable to cause a DCS100 shutdown at any time during these tests.

In an attempt to validate field reports of DCS100 shutdown, I added a fixed load of approximately 2 amps to the test track. I then repeated the stop, reverse, start sequence. Under these conditions, I was unable to demonstrate a shutdown of the DCS100. I then tried a “live” reverse (reverse while the engine is moving), and I was able to get the DCS100 to shutdown and then re-start. Strangely, the shutdown occurred approximately 3 seconds after the start-up current surge from the reversed engine and appeared to last for only a very short period of time. After the shutdown, power automatically returned to the track, the engine regained power, and the engine began operating in the new direction at the commanded speed.

The oscilloscope trace below shows a typical DCS100 shutdown transient, although I was unable to capture the actual shutdown.

Conclusion: It appears that there is a shutdown problem only if the reversal occurs when the speed step is other than 0 (assuming good contact between the wheel axles and pickup shoes). I suspect that there may be a software glitch for the QSI decoder causing the problem?

A secondary contributing factor may be poor contact between the pickup shoes and wheel axles as mentioned above. This problem is easily cured by removing the truck side frames and placing a drop or two of Deoxit D5 Power Booster (I know, pretty much a marketing name) at the juncture of the wheel axles and pickup shoes.

We welcome comments or suggestions from readers; please write or call.