After work of my colleagues on yesterday, the system works again. she said it is the microwave delay time and crystal angle.
We also do statistic on data. each setting, we repeat at least 4 times. and we found out that the 1.0W power give maximum FID area, about 435 unit in average. and the FID area from all set of microwave power have ±30 uncertainty.
i was accidentally, forget to pull out the NMR coil and let it hit by laser directly. after that, the reading start falling. the coil was token out and checked. it has a tiny spot melted by the laser, the rest is in good shape.
I am not sure it is the reason. is it is the case, then the data should not falling, should be a sudden change. this phenomena is observed before on 2 days ago. the original signal is 400, then after the accidence, it went to 270, and another measurement is 200 and stable.
i had tuned the tuner, but it just getting worst. and i found that, in order to get the minimum reflected wave, there are many solution for the capacitors.
at last, the laser polarization has no statistical trend to tell it depends on angle.