SHG cavity locked

Today we found out the reason why there is some flat part in the spectrum, it is because the PZT driver reached its limitation, so we reduced the offset of the PZT driver and got a normal spectrum. Then except the TEM00(the highest peak) we found a vertical first-order mode and a laguerre mode which can not be killed together. So we chose to reduce the TEM01 as lower as possible and keep the laguerre mode and try to lock the cavity.
Before start to lock the cavity, I set the temperature to the phase matching temperature which we found quiet a while ago. The readjust the mirror a little to bit to get the best mode matching I can. Then I set the the sinusoidal wave with frequency of 15.2MHz and amplitude of 400mV, sent it to the EOM to do the modulation,and this is also used to be the 'LO' signal in the demodulation process. Then used the PD to get the transmitted beam of the cavity, which is the 'RF' signal of the demodulation. Then with the demodulation board, the 'LO' and 'RF' signal mixed together, the output( error signal ） sent to the oscilloscope.Also the error signal will go through the low pass filter and send to the input of the PZT driver. The other part of the PZT input is a triangle signal(parameter: 28Hz, amplitude: 2.5V, offset:2.5V) As soon as I got the error signal I stopped scan the cavity with triangle wave and adjust the PZT by changing the offset of PZT driver until the cavity is locked. After the cavity locked, the green light become much more powerful, and the reflection of the green light goes back into the Faraday Isolator, and the reflection from the Faraday point to the height around people's eyes which is really dangerous.