Problem solved. The issue was electrical.
Steps I took:
1) thoroughly cleaned track and then ran 4 axle locos, stalling occurred, but if you look at the crossing like an one track "x" and other track "y", then the "x" track had stalling while the "y" track was fine.
2) I pulled out the multimeter, and found out that the "Y" track had 16vAC while the "X" track had 4-5vAC. I started to look at the connections.
3) 4 out of 8 rail joiners were soldered. The "Y" track had all its connectors soldered, while the "X" track had insulated joiners on one side (start of a new power district), and the other side had the track feeders soldered to the joiners, but the joint itself was not soldered.
4) After checking with the multimeter that the feeders were in fact delivering voltage (16vAC), the problem was that the "X" track was not getting enough electrical connection.
5) Since 4 axle locomotives typically have a shorter wheelbase (between trucks) than the 6 axle units, and the 19* Atlas crossing is longer than say a 30* or 45* Atlas crossing, the 6 axle locos could make it on the "X" track since it has more electrical pickups (through 6 pairs of wheels), and when the lead truck would make the crossing, the rear truck would still be on powered track, and keep the motor running. They, in hindsight, probably stalled a little, but not enough for me to immediately notice the issue.
6) My solution was to solder the unsoldered track connectors for the "X" track, and voltage to the diamond now was a constant 16vAC for all track.
7) I tested 5 different 4 axle locos, from switchers (which gave me the most trouble there) to the GPs I have, and everything ran fine.
A happy ending!
