Lulzbot’s NEMA17 (specs:https://www.lulzbot.com/store/parts/nema-17-stepper-motor) is comfortable in a range from -20 °C to +50°C and has an absolute maximum rating of 80°C. So 90°C is far outside it’s specs and the motor does actually stop working at this temperature.
Maybe you mistook Celsius for the strange, evil, old and obsolete Fahrenheit?Previously we thought the problem was the mechanical resistance (https://forum.lulzbot.com/t/brand-new-1week-stock-taz5-x-axis-stepper-getting-very-hot/2903/9), but unfortunately we were mistaken 
The motor is still getting hot.
Doing further research, I discovered that the SPI signal which is sent to the AD5206 digital potentiometeron the Rambo1.3L Board (co-responsible for stepper current limiting) is barely readable by an oszilloscope. In fact, the SPI signal we saw, barely has anything to do with what the firmware is supposed to send to the chip. The logic analyzer could not even differentiate HIGH from LOW. That might be the reason, that setting any kind of current limiting in the firmware, has absolutely no effect. (yes we verified this)