The datataking setup
  xmitting a birdie from building 2 to find the az peak.
The  shutdown sequence
Monitoring with 327 during the power cycling
    dynamic spectra of 327 power cycling
    327 rfi frequencies that turned on,off during the power cycling

Page in Progress (28jan13); The multi frequency data taken after the on or off has not yet been reduced

---

Intro:

   The power at building 2 was shutdown on 24jan13starting around 17:00 hours. Data was taken with various receivers, before, during, and after the shutdown to investigate rfi coming from this area. Prior to the shutdown birdies were transmitted from building 2 to find the azimuth of the strongest signal.
    The shutdown included: building 2, ac on top of building 2, and the two trailers between building 1 and building 2.

---

The datataking setup:

• telescope: azc 329.13720, dome=19.04, ch at stow.
• this position was determined by transmitting a birdie from building 2.
  
• Mock setup:
• 8192 channels, polA,B dumped at 1 second integrations.  Bandwidths varied by receiver.
  
• Receivers used. The sequence listed is the order the data was taken (by position on the turret floor).
  
• 327 receiver:  entire band. Resolution bandwidth (rbw):5.2 khz.
• lbw receiver: entire band. rbw: 7.0Khz
  
• 430 receiver. entire band. rbw: 1.9 KHz  .. taken for the first off,on, then discontinued since the receiver was turned off.
  
• sbw receiver. bands within our filters:
• 1700-1880. rbw: 10.5Khz
• 2040-2400. rbw: 7khz
• 2600-3100. rbw:7khz
• prior to turn off, after turn off, after turnback on:
•  data was taken for 200 seconds for each receiver (sbw needed two sets of 200 secs to cover the entire band).
• The command file is in /share/obs4/usr/x111/shutdown.cmd
  
• During the turn off, turn Back on.
• the 327 receiver was left running during each transition to catch changes as they occurred.

---

The  shutdown sequence: (top)

•   prior to the first shutdown,  200 seconds of data was taken the different frequency bands:
  

• 16:53 327
• 16:59 lbw
• 17:02 430
• 17:07 sbw1
• 17:11 sbw2
  

• Turn off 1 (dynamic spectra)
  

1. time	minutes from start	action
   17:17:53	0	start 327 continous monitoring
   17:20:10	2.3	start turning off 1st floor
   17:26:15	8	1st floor is off
   17:26:55	9	start turning off 2nd floor
   17:37:25	19	2nd floor is off (as well as trailers)
   17:38:10	20	start floor 4 turnoff
   17:38:50	21	start floor 3 turnoff
   17:47:35	29	floor 3 is off
   17:53:40	35	floor 4 is off (ups in phil's office got left on)
   17:55:00	37.12	main breaker (and ac breaker) off

   
   

•  start data taking cycle during first power off

• Turn on 1 (dynamic spectra)
  

• floors 1,2 and the trailers were left off for the duration.
  

1. time	minutes from start	action
   18:24:33	0	start 327 continuous monitoring
   18:27	2.45	turn on main breaker
   18:29	4.5	start turning on floors 3 and 4
   18:31	6.5	turn on main AC breaker
   18:32:20	7.5	AC compressor turns on
   18:40:35	16	floor 3 is all on
   18:43:25	19	floor 4 is all on

   
   

•  datataking cycle all off. 1st time

• 
  

• Turn off 2 (dynamic spectra)

• 
  

• time	minutes from start	action
  19:06:18	0	start 327 continuous monitoring
  19:07:20	1	start shutdown floor 3
  19:07:45	1.5	start shutdown floor 4
  19:16:40	10.25	floors 3,4 are off (but some ups still on)
  19:18:11	12.	ac main breaker off
  19:18:34	12.3	floor 3 breaker off
  19:18:55	12.62	floor 4 breaker off
  19:22	16	last hidden ups off

• datataking cycle all off:  2nd time

• Turn on 2 (dynamic spectra)

• time	minutes from start	action
  19:43:50	0	start 327 continuous monitoring
  19:45:00	1.17	main breaker on
  19:45:29	1.5	ac breaker on
  19:48:00	4.2	start turning on floors 1 and 2
  19:48:47	5	ac compressor starts
  19:55:49	12	floor 1 is on
  19:59:18	15.5	floor 2 is on
  20:02:27	16.6	trailer 2 is on
  20:05:51	22	trailer 1 is on
  20:06:17	22.5	start floor 4 turn on
  20:06:27	22.6	start floor 3 turn on
  20:13:08	29.3	all on

  
  

---

monitoring with 327 during the power cycling (top)

    The dynamic spectra  images show the 327 receiver during the turnoff and then turn back on.

• There are two bands for each measurement: cfr-315 Mhz and cfr 339Mhz.
• processing to create images:
• create median bandpass
• fit a 20th order harmonic function and then divide each 1 sec spectr by this
• remove the median value from each spectra
• scale range to +/-5 sigma (using robust sigma estimator).
• The image had to be decimated for display (there were 8192 data channels but only about 1300 for display).
• The second column  in the above tables should get you close (along the y axis)  to what was happening when...
  

EVENT	34 Mhz band images (25 Khz resolution)	Image blowups: 6 x 6.2 Mhz images (4.1 Khz resolution)
1st Turn off sequence	315 Mhz Band	315:1 Off	315:2 Off	315:3 Off	315:4 Off	315:5 Off	315:6 Off
	339 Mhz Band	339:1 Off	339:2 Off	339:3 Off	339:4 Off	339:5 Off	339:6 Off
1st Turn on sequence	315 Mhz Band	315:1 On	315:2 On	315:3 On	315:4 On	315:5 On	315:6 On
	339 Mhz Band	339:1 On	339:2 On	339:3 On	339:4 On	339:5 On	339:6 On
2nd Turn off sequence	315 Mhz Band	315:1 Off	315:2 Off	315:3 Off	315:4 Off	315:5 Off	315:6 Off
	339 Mhz Band	339:1 Off	339:2 Off	339:3 Off	339:4 Off	339:5 Off	339:6 Off
2nd Turn on sequence	315 Mhz Band	315:1 On	315:2 On	315:3 On	315:4 On	315:5 On	315:6 On
	339 Mhz Band	339:1 On	339:2 On	339:3 On	339:4 On	339:5 On	339:6 On

processing: x101/130124/shutdown_327onoff.pro, img327.pro

Plotting the rfi that came and went with the building turn on,offs:

    All of the high resolution 327 images were manually searched for rfi that turned off when the building power was turned off or turned on when the building power was turned back on.

    The plot shows the 327  frequencies that turned on, off with the building power (.ps) (.pdf):

• The vertical scale(time)  is broken into 4 sections:
• 1st power off, 1st power on, 2nd power off, 2nd power on.
• Each section has the time scaled from 0 to 1 to cover the data taken for each section.
• an offset was added to each section for plotting.
• Each color is a separate system.
• Within each section, an * is plotted at the frequency and relative time when rfi appeared or disappeared.
• Horizontal dotted lines were drawn at each  freq/time to see if the same frequency repeated in various turn off,on cycles.
• How to interpret the time axis:
• Each section starts with the start of data recording. There may have been delays before any action was taken (see the shutdown sequence for that section).
  
• Main Breaker  dashed horizontal line when main breaker changed for this section
• AC Breaker   dashed horizontal line when AC  main breaker changed for this section
• Looks like a lot of the rfi is occurring when the AC breaker is being turned on or  off.
• on 28jan13 a spectrum analyzer at the building II A/C unit verified that this 327 rfi was coming from the a/c uint.
  

processing: x101/130124/findrfi327.pro

---

Processing, plotting the data taken with power on or off.

Processing the data:

   The data from the datataking sequence: powerOn,powerOff,powerOn,poweroff,powerOn was processed as:

• compute the average spectra, and rms spectra for each of the 200 second integrations in the on,off, and on power setups.
• This was about 150 files.
• for each 200 second average spectra fit a robust 50th order harmonic function (8192 channels in each spectra). Use this fit as the bandpass correction.
• Divide each average spectra by the bandpass correction then subtract 1.
  

Plotting the dat, finding the rfi:

• The power off,  power on  were over plotted using color to separate the plots (offsets were also added to the power on data).
• Each page contains 1 band.
• Top: polA, bottom polB
• For each frame:
•  the 2 power off spectra are over plotted in black.
• The 3 power on spectra are over plotted in red,green,blue. These are offset from one another.
  
• rfi from buldinng 2 was identified by:
• The rfi was not present in the offs. It's ok to over plot them with no offsets since a signal in any off  means it's not from building 2.
• The rfi should be present in all 3 of the ons. This is why they are offset so you can see if the signal is present in each.
• It was ok (although suspicious) if the rfi was only in polA or polB
• The 327 band is so dirty that it's hard to see when rfi comes and goes.
• I should probably blow this up into 6 Mhz sections like the dynamic spectra.
• Since we have the dynamic spectra, it's probably not needed.

The table below shows:

• the plots for each receiver with
• number frequency bands measured
• list of rfi frequency that wer in the ons and not the offs.
• Some notes onn each freq.
  

Table of rfi freq from visitor center shutdown

receiver	#numbands	rfi Freq Mhz	Page	Notes
327 (.pdf) (1 Mbyte) 302-352Mhz	2			only seen in dynamic spectra 327 is very noisy. alfa motor controller was on
lbw (.pdf)  (7 Mbytes) 1120-1750Mhz	14			polA,B.. maybe something weak in Boff?
				polA
				polA
sbw (.pdf)  (13 MBytes) 1700-1790 2040-2400 2600-3100	24	-	-

---

Notes on the test: (top)

• The 327 band has lots of narrow birdies.
• The alfa motor and the dome cameras were turned off for this  test.
• Repeating the power on,off cycle multiple times really helps to identify the  rfi in the data taken during the turn on,off.
• next time it might be helpful to spread out the individual turnoffs.
• floor 3,4 could be spread out more in time.
• A/C and main breaker could be separated by more time.
  

processing: x101/130124/shutdown_327onoff.pro, siteshutdown.pro, onoff1rcvr.pro,baslin.pro,fitbas.pro