YamaT, Akutsu
Abstract
In the PSL room, measured the open-loop transfer function (OLTF) of the PMC frequency stabilization servo with two different instruments: a spectrum analyzer 35670A by Agilent (now Keysight) and the multi-instrument device, Moku:Lab.
Detail
We started with measuring the 35670A's noise level while terminating both of the input connector (Ch1 and 2) with groud terminators each (note: they are not 50-ohm terminators!!). To convert the binary files (*.sdf) from 35670A, I need to consider the modern way... in the past we could use a series of tools for this, but they won't work for the modern windows; use WINE or so?? Anyway, judging from the display of the 35670A, there seemed no strange noise features at a glance; let's do further investigation later. By the way, we avoid to connect the power supply cable of the 35670A to the outlets that had already too(?) many instruments connected (in a so-called tako-ashi cabling manner). Instead, we used an outlet box on which the label said "For use in the PSL room, new 4 outlets" in Japanese; three pins (both AC and the case ground) were all connected as in the intended manner (When we found the 35670A near the IOO racks, its power supply cable was connected to the outlet via a converter from 2 to 3 pins... maybe for "noise" issue?? Anyway, so far, we were not suffered from the noise even using three pins here). When we found it, no other power supply cables had been connected.
Then we connected the signal cables as follows between the TTFFS D1301751 and the 35670A to measure the OLTF for the PMC lock:
- Common/ OUT1 <--> Ch2
- Common/ OUT2 <--> Ch1
- Common/ EXT <--> Source
To do this measurement, do not forget turning on the switch on the MEDM to bring the EXT signal into the loop.
Actually, before doing this OLTF measurement, we took specra of Ch 1 and 2, and noticed that OUT2 (Ch1) had a gain about 10dB to OUT1 (Ch2)... we do not know the reason of this setup but this is confirmed with the circuit diagram... hmm... not so useful... This was also confirmed by measuring the OLTF with no source... -10dB constant was measured; the data plot will come later.
The OLTF plots will also come later; I need to convert the sdf binary file to ascii. About the source input, we started with 10mVpk, then 5mVplk, and then 1mVpk. With these, all the measurements seemed not so bad. The raw UGF was around ~1kHz, but as explained, need to consider the additional 10dB gain block, maybe. Then the UGF would get higher.
Then, we replaced the 35670A with a Moku:Lab, so now the OLTF could be measured remotely!