First we changed the VGA gain from high gain mode to low gain mode for the B1, B1p, B2 and B4 longitudinal photodiodes by keeping the same gain value and without re-starting the internal clocks of the mezzanine to keep the demodulation phase.

As today there are 2 types of DaqBox firmware,  we used 2 different ways to update the VGA gain without fully reconfigure the DaqBox

• for the v4r2 firmware: we used the v14r2 DaqBox server
• for the v3r0 firmware: we used some expert commands to set the new gain value. As consequence for these DaqBox if someone modifies the demodulation mezzanine configuration, the mezzanine will be reconfigured with  the gain restored  to the high gain mode and the demodulation phase  will have to be adjusted

VGA gain:

• B2_PD1 and B2_PD2; previous:0x84 - now: 0x1C
• B4_PD1: previous: 0x89 - now: 0x20 instead of 0x40 to avoid any saturation of the ADC during the lock acquistition
• B4_PD2: previous: 0x89 - now: 0x40
• B1p_PD1, B1p_PD2, B1_PD1 and B1_PD2: previous: 0x86 - now 0x2A

With these new gains, the ITF was back at LOW_NOISE_3_SQZ at 2019-09-09 09:11:32UTC-GPS1252055510 (see the attached trend plot). Here the differents sequences of test perfomed

• From 09h30UTC to 10h05UTC with the ITF in LOW_NOISE_3_SQZ we tried to apply differents 6MHz modulation depth (-28dB to 1dB, 1dB to -28dB) (see this file for the detail)
• Starting from 10h30UTC  we tried to relock the ITF with the 6MHz modulation always present  and by  adjusting the 6MHz depth at ITF step 120 from 1dB to -1dB.  With -1dB  the ITF succeeded to lock at  LOW_NOISE_3_SQZ
• The last trial, from 13h04UTC to 13h20UTC  the ITF was relocked at LOW_NOISE_3_SQZ with a 6MHz modulation depth of -1dB , at 13h20UTC the 6MHz modulation depth was set to  -28dB . The TF unlocks at 13h40UTC