Looking at the OFI TEC servo we noticed that the response was much slower than what one could expect from the given ugf frequency. The gain was increased fomr -1 to -15, whereas the ugf frequency was decreased from 50mHz to 10mHz. This resulted in a response time around 10s, which is about 3x faster than when we started. Increasing the ugf beyond this can make the servo unstable.

As a side note: The applied voltage at the driver was about 3x higher than the voltage readback from the TEC, indicating a rather large voltage drop in the cable.