Hi guys,

I've been chasing a sporadic ground fault issue with our GP38. This is a straight 38 without Dash 2 electronics. Has a D32E main generator. At low speed it pulls heavy and has no issues. Once it gets out on the road is when we run into issues. At around 13-15 mph when you start to notch out, the ground relay trips out. This happen either running lite, or with a train. This engine has had a history of getting an Excitation Limit alarm once in a while when running at sustained speeds of 20mph, but it has never had ground faults until recently. I suspect its either a main generator issue or a traction motor problem. Possibly triggered by back EMF? This engine is not equipped with motor cutouts so thus far I have not been able to try cutting out traction motors to try and pinpoint the issue.

Are there any comprehensive guides to properly performing megger tests on main generators and traction motors? I'd love to read up on the process and educate myself before attempting anything.

Any help would be appreciated!

Disclaimer: I am not a diesel guy.

Are you able to inspect the traction motors' commutators? The ground fault relay is not tripping because of "back EMF" but at higher speeds you might be getting flashover in one of the traction motors. If so, there would be evidence of this on the commutator and/or around the brush holders. Check all your brush holder springs for proper pressure, brushes that are bottomed out or stuck in the holder.

Does this GP-38 have a main generator with a commutator or a main alternator with a rectifier?

Megger testing might not indicate the problem if it is an intermittent fault. Flashovers resulting from poor brush or commutator condition are not insulation weakness, and very often a flashed-over armature will read just fine on a megger test, unless and until the flashover fault blows a hole in the winding somewhere.

There is also the possibility that the ground fault circuitry is defective and these are false trips.

Another problem that can cause commutator flashing is if the commutator is out of round or there is excess height bar to bar. If you have some way to move the loco slowly, put a dial indicator on top the brush and see how mush variation you get. On higher speed motors than you have, the max. allowance was 0.003” TIR or 0.0003” bar to bar. See if you can get actual limits for your loco.

If there is a high bar that is causing the brush to lift, you should see evidence in the form of pitting, burning, or erosion of the bar edges on the bar in question or the adjacent bars.

Out of round is more difficult to see but if it is getting to the point where a flashover is commencing that faults over to ground, you will see some copper flash marks on the brush holders, or a ring of molten copper drops on the mica/string banding, or in the motor case.

I had a D3F compressor armature in the lathe recently for comm truing since it was visibly worn into a "U" shape (what you'd call "hollow worn" if it were a wheel). After mounting it between centers on the lathe, I indicated the surface of the comm at several points along the axis. It was more than 50 thousandths out. I thought surely, that isn't possible, right? I managed to get an indicator on the armature shaft beyond the commutator, and that is running dead nuts on. So screw it, I went ahead and turned it. As it was, I needed about .200 in diameter just to get the "U" shape out of it, and then I could see there was a big flat spot on the comm that took another 50-60 to get out. Of course had to undercut each slot again, lots of stoning and cleaning, the comm finally looked pretty good.

So we put it back in the compressor and ran it on low voltage. The brushes were dancing up and down. Turns out, the comm end armature bearing is worn in an oval shape, to the extent that the commutator is gyrating as it revolves in the bearing. That would explain the big flat spot. OOS pending a babbit job.

Ground relay is caused by one thing...Voltage finding a path to the carbody. What you need to keep in mind that voltage only rises as a result of commutation. You can hold the brake on and get high amperage and not kick the ground relay. When you start to roll is when voltage begins to rise and eventually will find the path to carbody.

It's been years since I worked on straight GP38's but toward the end of their service lives they were troublesome primarily on account of deteriorated control wiring causing loading and ground relay problems.

First thing I would do is visually inspect the main generator and look for signs of arcing across the string band to the frame. If the creepage area is dirty clean it. Make sure the brushes are not short and move freely in the brush holders. I would strongly caution against using any kind of insulating varnish on the generator frame unless you are absolutely sure the are in question is clean and free of carbon tracks. If not you will entomb the pathway to ground under the varnish.

The traction motors are easy to meg. Block the movable tips with a paper cup or popsicle stick and shoot away. If you find a motor with low resistance the process is the same as for the main generator.

I think you'll find the most likely culprit is going to be the high voltage connections to either the transition or wheel slip resistors. The neoprene wiring in these areas was beyond problematic.

One thing that you didn't indicate is if it kicked when the unit was trying to make transition. If that was the case the issue could be the unit is holding onto load when S drops out and P's pick up with the resulting arc.

EBL

Adding to what Eric posted earlier, check that the generator field dissipation resistor is intact and has electrical continuity. This is RE2 on many EMD locomotive types. If this resistor is failed open or has the circuit open, and you go through switchgear cycles with the generator partially excited, it can shoot high voltage arcs all through the traction and control circuits.

Also note that there is a much better writeup of circuit testing in EMD locomotive service manuals written after 1972, a couple years after the documentation for your locomotive.

PC

In addition to what Eric and Preston mentioned, Don't overlook the brusholder insulating posts, they need to be kept clean and occasionally they will develop an invisible path to ground thru the center core if allowed to become oil soaked. In my experience, the creepage areas of the generator and motors would be my first area to look at as Eric mentioned in his post, if you see marks like little welding arcs, they are what is tripping your ground relay. Meggar testing on a regular time schedule is a good tool to prevent the issue you are experiencing. Take your print, and see where you can isolate each cable, armature, and field at a contactor, make a spreadsheet and test annually or quarterly based on your usage. The results will tell you where to look and when to clean. Don't forget the starting circuit/contactors are part of the high voltage section of the generator. Also make sure your generator sump drain is not backed up allowing water/oil to reach the bus bars under the generator.

.

Thank you to everyone for the suggestions. I will report some findings in the near future.