From last Friday

I previously attempted to lock the SHG by switching around the DDS1 DAC channels - in the normal configuration DAC0 outputs -9 dBm to +14 dB -> 5 dBm to main laser EOM, while DAC1 outputs 9 dBm to splitter for SHG/IRMC demod. I switched them around but this just makes the SHG oscillate even at 0 gain. So I put them back to normal. The rationale for switching is that the wiki shows an 88 MHz Qubig EOM datasheet which specifies that we need about 12-15 dBm RF power to achieve 0.2 modulation index. However, this is for the OPO EOM "EO-88K3-NIR", while the ML EOM is "PM8-NIR-88". I cannot find the datasheet for this particular EOM but other PM8 EOM datasheets indicate that 5 dBm is sufficient to generate 0.2 modulation index.

I measured the open loop transfer function EPS1/EPS2 of the SHG (fig 1, 2) and GRMC (fig 3, 4) using FFT mode of the spectrum analyzer and injecting about 100 mW noise to PERTURB IN. The coherence is basically 1 around the unity gain frequency. I also checked the SHG transfer function with swept sine mode and it gives basically the same result (fig 5). The SHG unity gain frequency is a bit low. Nominally it should be between 2-3 kHz (Aritomi thesis) but it sits about 1.6 kHz with 50 degrees phase margin when set to 0.09 gain. However, increasing the gain further past 0.1 makes the SHG oscillate quite a lot, even though the sticker on the servo says 0.2 indicating a former optimal setting. I set the GRMC gain to 0.8 which gives 3.4 kHz UGF with 45 degrees phase margin.

The GRMC error signal is clearly visible but also a bit noisy. It's quite curious that the GRMC signal shows up at 88 MHz (fig 6, 7), which is the modulation applied to the IR carrier, despite also propagating through three 1064 reflection dichroics. But evidently it has worked to lock the GRMC since the EOM replacement a while ago. I have a bit of difficulty in locking the GRMC to TEM00 though, a lot of the time it just locks to near zero transmission amplitude and I have to keep hitting the rest button a lot. when I managed to lock it with 4.2 MZ offset (stabilizing GRMC transmission to 25 mW) and noted the GRMC transmission level (GRMC TRANSMISS IN) of 316 mV. I set the GRMC threshold to about 100 mV which helps for avoiding HOMs but it's still not reliable for locking to TEM00.