I went to maiden a new build this morning (KISS V2 and T-Motor 50A 4 in 1) and every few flips the quad would just tumble to the ground.  One motor is warmer than the others but everything is brand new so I'm not sure if that's the problem. 

i would check for any loosen part, flexibility in the frame.. a screw not toghtenet?
a screw touch motorwindings?

capacitator attached?

35V cap attached, no screws touching windings, frame is solid. Motor 2 seems a bit warmer than the others but no visible damage to the windings or anything.

I took it out a second time and it seems to be happening more often.

im not familiar with kiss.
you can do the following to decrease the chance of a desync;
-lower startup/rampup power
-increase motortiming
if you still get desyncs, you can set demag compensation to high.

a single warmer motor isnt a good sign it could be a hardware issue. does it freakout over the same axies all the time?
maybe you can switch two motors to figure out esc or motor as source of the issue, compare how it freakssout and the motor temp.
you can also mesure continuity between motors and frame...

check the connection between esc and fc twice, bend some pins very gently to get tight fit helped me a few times

additional you can try to not go zero throttle during the flip or increase the motor idle speed

Thank you - I ordered an extra motor anyway and will try it. It freaks out over any axis, even a simple throttle punch.

I was able to hook up a friend's Betaflight FC to get to the 4 in 1 in BL Heli32. I followed UAVtech's advice and lowered the ramp up and raise the demag compensation and also updated the firmware to the most current version. This only helped in that it would lift the arm instead of just spinning out of control.

Motor 2 got VERY hot when I brought it in. I can see some of the windings are darkening. Do you think is this a motor issue...or do ESCs smoke motors?

you can try figure out by switch two motors and see if the heat wanders with the motor or stay on that particular esc position.

increase motortiming could be another try

Well it looks like it was the motor after all. Thanks, Bat!