Ansel and I put the end station CPS timing fan outs on separate +/-18V power this evening in hopes of eliminating the comb of odd harmonics of 1 Hz in DARM. A similar comb had been evident in the magnetometer that I set up at the EY ESD to look for blip glitches, and it went away when we powered the fan out with a separate supply. We think that this is because the magnetometer was detecting ripple in the field radiated from the +/-18V connector on the low V ESD drive and that we eliminated the ripple on ESD power by powering the CPS timing fan out on a separate supply. For this reason we varied from the work permit and didnt put them in a different rack. Ansel will fill in details of the studies he did that made us think that this is the source of the comb (especially if it really does go away in DARM), but we think that the comb isn’t GPS synched because it is driven by a clock on the PSoC board of the CPS timing fan out. Isolation of BSCs 9 and 10 has been restored to nominal.
Ansel Neunzert, Robert Schofield
This shouldn't be right. The PSoC board is supposedly clocked by half of the provided 71 MHz signal. However, looking at the photos on E1400155 I see that the 2 crystal oscillators of the development are still soldered in. Please tell me this is not true for the units installed in the field.
Our understanding of the electronics is based on the same DCC document/photos that Daniel linked to.
Some background info on what led us here: the comb was first reported in May by Brynley Pearlstone (alog), and again by Keith Riles in his ER9 narrow lines report (alog). Per Keith's report, the spacing is 1.999951-Hz, visible on approximate odd-integer-Hz frequencies. Robert and I looked back in older data for its appearance, and isolated the date range between February 26 and April 2, but there was not enough good DARM data in that window to track it further. Turning to magnetometer data, we found the comb in a channel at EX (H1:PEM-EX_MAG_VEA_FLOOR_X_DQ), and tracked its appearance down to March 14th using data from that channel (with help from one of Keith's students, Eilam Morag, who has written code for tracking the strength of known combs). We initially suspected the computer system restart on that date, but have not found evidence to support this hypothesis by examining data around subsequent restarts. As a second guess, we looked into the changes to the CPS timing system made on that date (alog) (both end stations). Coincidentally, Robert had placed a magnetometer in the CPS rack at EY to search for blip glitches, and this magnetometer showed the comb strongly, which motivated us to look into the CPS timing fan out.