Pierre Auger Project
Progress Report
Summary (photo album for this month html
More progress is being made with the fluorescence prototype at Los Leones. The relative calibration has been set up and measurements have been made. Fluorescence telescope images were made of the beam from a zenon flasher and from a pulsed laser set up 2.5 Km from the Los Leones building. The transit of stars was also recorded to measure the optical properties of the optics. The systems continue to perform well.
The software for the surface detector processor has advanced daily but the system is not yet ready for use in the field. The electronics packages for all but a few of the Engineering Array detectors are at the site ready for installation once the software is loaded. Freezing tests are under way in two detector stations deployed near the Assembly Building. The weather has been cold but clear and dry with pampa access possible away from the wettest areas of the Engineering Array.
The roof beams are in place on the Auger Center Building. Electrical work, HVAC duct placement and finishing is in progress on the lower office floor. Completion is now expected in August.
A press release relating to the success of the prototype fluorescence detector was prepared and given to the local news media. We have had good coverage particularly in Los Andes.
Fluorescence Detector
Report data-taking 20. & 21.6.
1. Is the hall really dark?
We try to find light leaks and sources within the bay. I put a curtain (black) on the detector and measure with the CM (current monitor) and variance. We read out only the middle part of the detector column 5 to 8 and 11 to 16, because the available curtain was to small. The noise in modules 9,10 are at present 25% higher we will look next week for the reason of it.
With curtain: CM = 3.7 +-0.3
variance = 4.5 +-0.1
Without curtain: CM = 3.5 +-0.3
Variance = 4.9 +-0.1 (all values averaged over all enabled pixels)
We will do extra measurements Monday on this topic to analyze CM and
Variance method more carefully. We found one light source, which may be reasonable and to be seen by the camera, some orange indicators in the HV system. We use now the curtain to cover this light into the direction of the camera. The change in the variance was not reasonable (4.85).
2. Counting rate with 100Hz hit rate regulation on, shutter closed: 0.85 Hz data rate no problem for the DAQ, data rate didnt change very much if we open the shutter, see point X?
3. Power down roughly at midnight for about 15 minutes.
Start of a small learning disaster in AUGER experiment. Problems of FD:
a. Mirror PCs not on UPS, neither Monitor of eye PC -> you cant switch it off controlled because of the missing Monitor on power, nor the UPS do it.
b. UPS should switch (if power dont come back) PCs off, also slow control one of the next jobs in Karlsruhe?
c. CDAS until 21st (late afternoon down), very complicated UPS system, is that necessary?
d. The general power distribution box makes a lot of noise until 1h after the event then it guts more quiet.
The 21st was very much more successful.
Plan was to measure counting rate acceptance of the DAQ as it is at present without interrupt and DMA, only by polling. Karlsruhe has the hardware modifications ready, but we fear the software-changes necessary to get again a running stable system are to dangerous for the measurement last time the hardware changes (in May here) use more than 2 very busy days last week.
What we have done: HV off, only crate and LV on, external trigger from a
pulser with precise adjustable rate (Agilent 33120A). We dont use the normal DAQ, only a simplified version
Run# , results:
1. Simple loop installed for readout reading page status register (2*32
bit) and first and last timestamp register (4*32 bit). If an event has appeared evaluate the parameters and clear the event.
2. Same as before, but read also the SLT-values (1s = 32 bit * 20 words) for the time span of 10s corresponding to 10 SLT matrices.
3. Same as before, but time span of 100 s (full SLT = 100 matrices).
4. Same as before but also readout of 10 s ADC data (100 slices of all pixels)
5. Same as before but readout of 1000 slices of ADC-data. This last limit should be in the range of the data taking of the actual DAQ the 4th may be done with some more efforts of Hermann. But 1 Hz is the value we promised. To do special external trigger runs for calibration with more than 1.3 Hz we have to look for a special calibration
Furthermore we looked in run #6 for the hit rate in the dark as function of an additional threshold on the regulated 100Hz 1st level trigger threshold. We regulate the threshold to a stable value and then stop regulation and add different thresholds. You have to have in mind that the hitrate threshold is a factor 16 (at present) higher then pixel value of the pedestal, because of the sliding box-car-sum window of 16 time slices. The curve shows two components the noise and the direct hits (of the PMTs). The noise levels off very quickly from 0.85 to 0.2 Hz. The second component around 0.2 Hz we attribute to the direct hits of the
PMTs. That could be verified by an extra (run #7) DAQ-run taking this data. The replay of this data shows clearly direct hits of the PMTs and/or the surrounding aluminium. There is only one difficulty: run #7 uses the readout of the ADC-data and should contain additional readout noise. From a separate measurement with open shutter under similar conditions run #8,9 we know, that the readout noise contribute to 16% and a very much higher rate of the 1st level trigger. That will then be regulated down (to get again 100 Hz) affecting the normal noise triggers but mostly not the readout noise! We have to do a more careful analysis and look at the thresholds.
Also for the open shutter run the rate was 0.85 Hz. Only if the do no
ADC-data readout we got 16% lower rate.
-------------------------------------------------------------
2nd report
-------------------------------------------------------------------
Report data-taking 22. & 24.6.
1. Dome tests (Rishi Meyhandan, Jeff Brack)
The dome/cylinder flat fielding device was mounted on the exterior ring of bay 4 and run for approximately 5 hours. The entire aperture was illuminated with relatively short duration pulses (0.5us to 7us) for AC mode, and two intensities in DC mode. We collected around 1000 events in each mode. This is a proof of concept test, which was successful, and will be improved upon in later trials. Uniformity measurements of the dome were made using the area between bay 1 and 3. The complete absolute calibration will require further measurements and calibration of the dome intensity.
2. Xenon Flasher (Paul Sommers, Brian Fick and Micheal Roberts)
The flasher was operated at 2.61 km out north of Los Leones on the road at the exact position 35 28.341 S and 69 26.860 W (GPS reading).
Data taking conditions: To avoid noise triggered events the threshold was increased by 40/16 ADC counts. We could see at once with 100 % efficiency the flasher tracks. First we got some manual shots (run2). Then (run3) 0,5 Hz on GPS second toward the eye, sweep in zenith angle, trying to hit all pixels. Finally a rapid sweep down and up around in azimuth at fixed zenith and back. We saw really very nice long (also in time) tracks over the detector. The flasher was iced and did therefore relative broad tracks, see album picture. We used for this measurements for the first time the online histogramming module of Pedro Facal, embedded at the same day by Andreas. The whole run was a complete success for FD calibration.
2. Current-Monitor measurements (will be provided separately by Daniel
Camin)
WBS 1.2 Fluorescence
Detector Electronics Tests (Matthias Kleifges FZK)
The following report describes the status of the installation at Los Leones end of May 2001:
Slow Control System:
(WBS 1.2.4.1)
The proposed slow control system based on field bus technology was installed. Its main components are the field PC (slow control PC), the general slave station (addr. 99) and the slave stations for bay 4 and bay 5 (addr. 41 and 51). These stations are all powered by an UPS system. Most of the work (done by A. Grindler (IK), G. Sequeiros (UTN-San Rafael) and P. Vitale (UTN-Mendoza)) was the wiring of the stations with the actors and sensors, e.g. the left and right shutter of each bay, the fail-safe-curtain (FSC), relays to switch the power of the FE sub-rack, the signals to and from the GPS clock, the door sensors and the indoor (per bay) and outdoor light sensors. The installation of all slow control hardware for both prototype telescope was finished and even those for more general actors/sensors (except for the interface to the HV system and the out side weather station). A lot of time was lost for tests with the shutter mechanic and the FSC. Despite the electrical wiring is correct, the shutter sometimes get stock (one or both sides) and has to be moved manually. Unfortunately, there was not enough time to install all control features of the program, but an environment to make basic operations. The program allowed a manual operation of the shutters and the FSC via push buttons, an updated revision of the software will be installed soon.
Lessons learned:
- improve the shutter mechanic (end switches, gear transmission ratio) and the FSC (clutch)
- provide cable trenches in the floor of the next buildings to ease cable installation
- safety features of the slow control program urgently required
FD computing facility
in Los Leones and LAN: (WBS 1.2.1.1)
N. Kunka installed and tested the LAN system in Los Leones: the internal DAQ subnet and the Eye Network. He configured the Eye-PC as boot and file server for the MirrorPC in all telescope bay. The LAN wiring for bay 4 and 5 is complete, it was tested to boot the MirrorPC 4 and 5. The Eye Network was setup with a laser printer, a connection to CDAS via the radio link, the link to the field PC and to several LAN plugs in all rooms of the building (calibration room, electronic room, DAQ room). A problem with the Eye-PC it didnt start up properly - could be solved by reducing the CPU speed. Norbert also suffered from the hard work of installing all the cables.
Installation of the
Front End Sub-rack: (WBS 1.2.2.1.3/ 1.2.2.1.4/1.2.2.1.5/1.2.2.1.6 etc.)
A one week delay in the delivery of the camera was used to make intensive tests with the FLT and SLT boards. We found some bugs in the time and page memory controller of the SLT. E.g., the SLT now stops the data taking automatically in case all pages are full. It is restarted when the PC sets any memory pages free. The installation of the sub-rack and wiring to the camera was done within 1 day. As usual we first decreased the electronic noise be an improved grounding. With a nominal gain of ca. 50 000 we achieved a variance of 4.6 LSB^2 with all detector parts (HV, LV ) on, but the shutter closed. Following features were tested:
- creation of test pattern at the analog board input to test the full analog and digital chain, the uniformity in gain and the pattern recognition of the SLT. By this test we found 1 Analog Board (1.2.2.1.2) with one defective channel.
- by turning on the HV we found one defective HE unit (1.2.2.1.1) as it had a short cut on the HV line, which broke down the HV of its group of 44 pixels. This channel had to be disconnected.
- Illumination of the full camera with the calibration flasher (Xe-lamp) (WBS ? J.Matthes). We found one pixel with a defective HE as it produced negative pulses.
- The RAM test on the FLT & SLT was successful, no short cuts on these boards.
- The regulation of the hit rate was tested both with pure electronic noise and with night sky background (see below).
Software for Tests
/DAQ: (WBS 1.2.3.1, especial 1.2.3.1.8)
All measurements during the commissioning phase were done with a test program of Sascha Menchikov running under C++ and WinNT. It was easy for him to adapt the program to changing data taking conditions and demands from us, and he could also implement and test the latest features of the FLT and SLT firmware as he did the VHDL programming of the hardware as well. First measurements (with closed shutter) showed that we are still affected by the cross talk problems at the Analog Board. As we know already from the test in Rome, this board a) induces noise in the analog part by the current monitor and b) picks up noise from the digital board during data readout times in the even channel numbers. Therefore all measurements were done with an increased threshold and a nominal single pixel rate f = 40 to 60 Hz (compared to the design values in the range of 100 to 200 Hz), but at the expense of decreased sensitivity (**).
Jonny, Paolo, Sascha and me started physical measurements at 18/19 May, which the determination of the optical telescope properties. Without the optical filter in (otherwise its to dark) we took a picture (with a CCD) of the image of a star projected on a screen fixed in the focal plane. The spot size was about 1.5 cm (or 0.5 °) as expected.
During the data taking the next nights we distinguished 3 types of measurements:
a) background measurements: periodic readout of (auxiliary) information on pixel threshold, variance (noise), pedestal, hit rate and current monitor (CM) value. With this measurements we intent to learn more on the DC night sky light level, its dependence on the threshold, variance and CM value. Paolo wanted to analyze this data. We found that the hit rate regulation is working at night sky conditions.
b) track measurements: readout data of triggered pixels in the SLT; this are tracks (4 out of 5) recognize by the SLT, e.g. FD showers, muons through the camera body or direct Cherenkov light.
c) calibration measurements: readout data of the full camera after illumination with a Xe-flasher. We started every night with this measurement to use it as in comparison with data from previous nights.
As we have already reported we saw early in the morning of May 23rd the first event, which has the full properties expected for an FD shower. Meanwhile more measurements have been done, now with the corrector ring in place (uncovered).
Summary & Outlook:
- The commissioning of the 1. telescope is finished, the system is able to take physical data.
- Slow Control system works in a very basic manner, software upgrade necessary; reliability of the shutter can only be achieved by some design changes.
- Next step is transition to LINUX DAQ with higher data rate and root data output.
- Solve noise problems with Analog Board to increase sensitivity.
- Implementation of software trigger, online data analysis etc.
- Dedicated measurements for determination of telescope parameters, comparison of design options; installation of 2. telescope
(**) Modifications to cure effect (a) have been done by S. Argiro on the boards of the 2. telescope in Karlsruhe, but effect (b) is now even better visible.
WBS 2.1.10.1 Surface Detector Deployment (Humberto Salazar Puebla, Laudo Barbosa CBPF)
EA Surface detector status:
There are a total of 42 tanks with water in the Southern Observatory,
41 in the pampas EA, and 1 at the assembly building.
There are 6 tanks more from Rotoplas provider at the assembly building.
Two of them are being used for thermal studies of freezing effects on the film domes) with liners provided by the Mendoza firm.
There are 21 tanks of the EA with PMT's installed, plus one more at the assembly building. During early July will be installed the 60 PMT's for the last 20 tanks of the EA.
One of the detectors, CARMEN is in data taking mode since 9th June and transmitting monitoring data.
WBS 2.1.1.1 Tanks
(Humberto Salazar Puebla, Laudo
Barbosa CBPF)
Shulman resin will be shipped in early July to Brazil to produce 50 tanks.
Rotoplas at Pilar, has made 4 tanks more and they are ready to be inspected and shipped to Malargue. Next set of 10 tanks will be make with the new Mold to be defined at the CDR.
Plastrong is still working with Shulman's Resin in order to make a good
tank with this resin.
The CDR for tanks will be held at Fermilab next 18-19 July.
WBS 2.1.5.2 Liners
for preproduction (Dave Warner Colorado State, Humberto Salazar Puebla,
Laudo Barbosa CBPF)
Dave Warner visited IHEP pointing to improve and assure 50% of the preproduction lines from China. He tested the sample welds from IHEP and found they were greatly improved from earlier samples.
CSU have received the controller for the new welding machine, and they hope to have the mechanical setup allowing them to make welds with it ready in a week or so.
CSU is still on track to produce window kits for liners by the end of
July, and it is needed to start to plan in detail where the next 100 liners will be produced, and when.
Two liners has being completed by the Argentine liner firm in Mendoza. They will be installed in test tanks in Malargue this week
The US liner manufacturing firm will be taking delivery of a CSU-style welding machine next week. CSU hopes to be getting weld samples from them in 2-3 weeks, and liners in ~1.5 months.
Dave Warner visited the liner company in Mexico to help them with their development work (Humberto Salazar and Oscar Martinez were present). We were able to make some improvements in their techniques, but it seems to be there is still a great deal of work to be done there before they are ready to make liners. They are using a heat seal machine to make their seals, but it is not going to be able to do many of the seals needed to make liners. They did produce ~1000 meters of laminate which should be to the Auger specification. We need to investigate this laminate further.
WBS 2.1.2.1 Antennas and brackets (Humberto Salazar Puebla, Laudo Barbosa CBPF)
The antenna mast will be increased in the height 50 cm, to give headroom for collaborators working on the top of the tank. The square lower section should be increased by 25 cm. and the round upper section should be increased by 25cm too.
Studies and dynamic calculation are in progress to improve the oscillation situation with antenna brackets and the modification to the mast installation scheme.
WBS 2.2 Surface
Detector Electronics Installation (Jim Beatty Penn State)
In the time since the May meeting, the SDE task has conducted extensive work on electronics deployment. All large hatch covers have been modified to include the feedthroughs required to connect the PMTs to the electronics, and the electronics enclosures have been prepared and installed on nearly all tanks. A tank near the assembly building (Laura) has been fully equipped and is being used to support development of station, LSX, and CDAS software. Twenty-one tanks in the field have phototubes with bases installed, and additional PMTs are en route to Malargue. Approximately ten tanks in the field have electronics installation partially completed, awaiting a few missing components. One station, Carmen, is running a simple monitoring client and awaits only final software.
We have recently completed several key items which had been impeding progress. The version 2.2 front end has been tested and all units are in Malargue. The problem leading to frequent catastrophic failures of the tank power control boards has been traced to an incorrect resistor value, and these boards are in Malargue (except for a few which have been sent for repair). The last ten slow control boards are entering the shipping process. The station software is nearing completion. During July, we plan to continue installation and can reasonably expect to turn on most of the engineering array tanks.
Work is also proceeding on planning for preproduction. In particular, we are conducting a review of costs and work assignments. Testing of the prototype extended dynamic range photomultipliers is nearing completion, and we expect to obtain cost information from the vendors and recommend a PMT choice to the collaboration in mid-July.
2.2.1.1 Surface Array
Phototubes (Arun Tripathi UCLA)
We have received 5 modified PMTs each from ETL, Hamamatsu and Photonis. We have measured Gain, Dark Current, Linearity and Singel PE sectra on all these PMTs. We are currently in the process of compiling all these results into one document, which will be made available as a GAP note soon.
Briefly, Photonis is still unable to produce good single PE spectrum, but the reason is known: foil type first dynode. Since we can not see good single PE from these PMTs, the question of absolute gain calibration for these PMTs is still open. They show good linearity though.
One of the new Photonis PMTs had a crooked dynode structure. We need to understand the cause, whether it was shipping or manufaturing.
New Hamamatsu PMTs show good linearity, as well as good single PE peak. The operating voltage for 1E6 gain for these PMTs can some time exceed 2000 V. This voltage will have to be brought to a lower value, since the production base can only supply 2000V max.
The new ETL PMTs show good single PE peaks, but linearity shows behavior that we don't quite understand yet. The nonlinearity actually becomes positive (even up to +4% for one PMT) as the anode current increases up to a certain point, and then it drops very rapidly.
The production PMT test facility has seen much improvement in the past month. The linearity system has been automated and the LeCroy 1440 HV supply is now controllable by computer. The linearity of some 8-inch PMT has been measured and is consistent with what is seen in the original test station at the 5% level. It is necessary to remove these smaller artifacts to be able to measure linearity to within 2% reliably.
A DC measurement system is being designed by T. Ohnuki (a grad student) to allow gain vs voltage curves to be measured for the production system. As well an amplifier to measure spe pulses at low gains being built.
WBS 6.0 Site
Development (Ingo Allekotte Instituto Balseiro)
- A preliminary survey of the proposed preproduction area has been undertaken in June. The alluvial cone of the Salado River presents a compact gravel soil, accessible all year long. This area can encompass around 55 preproduction positions. Contact has been made with the owners of this ground, who have expressed their intention of signing the contract in the near future.
- The central part of the array, south of the Atuel River, also presents good access conditions. Although this very large area is not accessible during all the winter, it is in much better condition than the Engineering Array area.
- On the other hand, there is a wet part north of the EA which has also been selected for preproduction which is as difficult to access as the EA area. Some positions will not be accessible due to swamps.
- The delta area between Salado and Atuel Rivers is difficult to access, which will probably imply a gap (roughly, 40 positions) in the full array.
- A specification for the preproduction survey and position marking has been written. Presently, we are in the process of obtaining prices for these works.
- The site web-page (cabtep5.cnea.gov.ar/particulas/ingo/surveyindex.htm) has been improved, with new images. It is expected to be transferred to the Malargue server soon.