A new mixer was tested (elog2626) to be more stable than the old one (elog2616). Actually, from elog 2616, the error is only about 1 count.

Anyway, the detuning was monitored with the new mixer. The result is attached here. The detuning change follows the correction change as well.

Especially, the correlation of detuning change and temperature change is shown in the attached figure. From this figure, it looks that they are quite correlated.