Abstract
We took back the BIO cables (CB{I,O}_L32_C1) for the BFV coil driver from the coil driver switcher,and stole again the BIO cables (CB{I,O}_L32_C0) for the coil driver switcher from the PI coil driver.
After then, only one BO cable (CBO_H32_C1) is unused, so it connected to the CTRL port of the PI coil driver.
As a result, there is no readback of de-whitening filters for PI though it can be switched.
After our work, we engaged two stages of de-whitening filters for PI with confirming it by a sounds of relay switch.
Details
In the previous work klog#33152, we stole the BIO cables (CB{I,O}_L32_C1) for the coil driver switcher from the BFV coil driver. After then, Kenta and Ushiba-kun found the DAC noise contamination of GAS and BF stages (klog#33197). At that time they couldn't enable de-whitening filters for BFV because of the stolen BIO cables. As the temporal solution for this issue, Ushiba-kun inserted a config board for the BFV coil driver (klog#33223). In order to remove this temporal solution, we rewired the BIO cables.Trying installation of a new BIO cards
At first, we tried to install a new BIO card as the 6th one on the IO chassis in the K1EX1 rack. But real-time front-end couldn't find any PCIe cards on the IO chassis when the 6th BIO cards were installed. According to my experience in the test stand, we was able to use 6 or 7 BIO cards on one IO chassis. Ambiguity of a number of cards seemed to come from an used cable between real-time front-end and IO chassis, the power supply situation, etc.
Though I'm not sure an exact reason why we reached a limit of a number of BIO cards as only 5, I guessed that it's also related to the total number of cards including also ADC, DAC, and BO. I was able to use 6 or 7 BIO cards with 1 ADC and 1DAC cards on the test stand. On the other hand, 3 ADC, 3 DAC, and 1 BO cards are installed on K1EX1. Anyway, we couldn't find a way to use the 6th BIO cards on K1EX1 though we tried to reboot some times and to use another PCIe slots.
Rewiring BIO cables
Because we found that it was impossible to use the 6th BIO card, we changed our strategy to the next best way as that BIO cables for the coil driver switcher is stolen from the PI coil driver instead of the BFV coil driver (DAC noise contamination from PI should be a smaller impact than one from BFV). So we removed the config board on BFV coil driver added by Ushiba-kun and took back the BIO cables from the coil driver switcher to BFV coil driver. Then we stole BIO cables again from PI coil driver in order to use on the coil driver switcher.
After rewiring, we found only one BO cable which was not used. A paired BI cable with this unused BO cable is apparently used for release sensors and the release sensors has only monitor channels. So we used this BO cable for PI coil driver for a control of de-whitening filters. Because there is no enough BI cables for PI coil driver, we cannot check the readback signal to ensure the state of filter engagement. So we confirmed to engage two stages of de-whiteninig filters by hearing a sound by relay switch. In common case, a state of the de-whitening filters for PI is never touched. So it might be enough to use a config board. If we obtained a benefit of using BO even an absence of BI, it's a time when we need to do a health check for ETMX. It's difficult to do the health check with enabling de-whitening filters because of DAC saturation.
Operation check
After the installation, we confirmed that de-whitening filter for all stages can be switched one-by-one and the test mass can be driven with LPCD. Fig.1 and Fig.2 shows the TM oplev signals when excitation was enabled from HPCD path and LPCD path, respectively. (I manually stopped two ndscope window for oplev signal and DAC output, so time epoch of two window has slightly a bias. Sorry for the lack of clarity.) We can see some excess on the oplev signals, so the coil driver output seemed to work fine. After then, we requested ALS_DARM and confirmed that changes in this work don't affect the lock acquisition.
