09.15 Following overnight pumping of CARME chamber by Leybold turbo
      Pirani 1.00e-3 mbar IE514 2.83e-7 mbar

     The Leybold IE514 gauge currently in use is the CARME unit ( serial # 1054 ). The calibration data can be found at https://elog.ph.ed.ac.uk/CARME/56


09.44 Pirani 1.00e-3 mbar IE514 2.71e-7 mbar

10.10 Leak tester attached to backing line

      Valves to leak tester and turbo (connection to Edwards scroll pump in between) closed.

      Stop Edwards scroll pump with magnetic valve 'red' button - magnetic valves closes and Edwards scroll pump is switched off.

      Open valves to leak tester and then (slowly) turbo.

      Wait for leak tester to settle in e-12 mbar.l/s range. Leak tester pressure 3.6e-3 mbar. IE514 2.69e-7 mbar.

11.00 Pirani 1.00e-3 mbar IE514 2.42e-7 mbar
      Leak tester 1.8e-10 mbar.l/s

13.04 Pirani 1.00e-3 mbar IE514 2.06e-7 mbar

      Leak testing.
      Downstream DN100 flanges from CARME to dipole OK
      Lower NEG power connector feedhtrough (downstream chamber) OK
      Left hand side (looking downstream) bellows OK

      c. few e-9 mbar.l/s leak bottom left (lookimg downstream) of COF600 wire seal

      Additional torque to M16 bolts bottom left - c. 10 bolts -> 220Nm+

      c. 0.1-0.3e-11 mbar.l/s

      Further checks around circumference of COF600 seal reveal some small leaks c. e-11-12 mbar.l/s.? However background rate
      in leak tester too high for sensitive test - will await recovery to resume checking.

14.00 Pirani 9.8e-4 mbar IE514 1.96e-7 mbar
      Leak tester 4.9e-10 mbar.l/s


14.35 DSSSD test
      IE514 OFF

Ambient temperature 21.5 deg C

Top right
-10V -4uA
-20V 10uA
c. -30V > -20uA trip

Top left
-150V -5.15uA

Bottom left
-150V -4.97uA

Bottom right
-150V -6.33uA

Top right (repeat)
-20V -10uA
-25V -14uA

15.28 IE514 ON
      Pirani 9.8e-4 mbar IE514 1.87e-7 mbar

15.40 c. few e-9 mbar.l/s leak around magnetic valve - probably not an issue in terms of backing performance but may interfere with/confuse leak testing

16.05 Pirani 9.8e-4 mbar IE514 1.74e-7 mbar
      Leak tester 1.2e-9 mbar.l/s

Will leave CARME pumping with leak tester on roughing line and see whether leak rate recovers. If so, will continue leak tests tomorrow. 

18.50 Pirani 1.00e-3 mbar IE514 1.48e-7 mbar
      Leak tester < 1e-12 mbar.l/s (leak tester valved off)

Grosmantage are scheduled to arrive 10.00 to inspect COF600 seals and their fit to the COF600 flange. They may decide to adjust the seals so that they fit the flange 
under 
tension.

Once Grosmontage are satisfied with the seal and its fit we can close the CARME chamber and start pumping down. 

Attachment 1 - photo of bakeout tent parts including those adapted/modified by Davide.


18.40 Grosmontage were re-tasked this morning.

      Davide, Omer, Cosmas checked alignment of COF600 male and female flanges.

      4x COF600 seals were checked - for some seals we observed wave over a ~ 10cm section of the circumference and/or surface scratches/roughness.
      Also observed that COF600 seals appeared to be too large - consequently the wire seal would be seated well ( i.e. in contact with the flange surface )
      for the majority of the circumference but would lift from flange surface for part(s) of the circumference.
      The fourth COF600 seal seated well over all of the circumference and we decided to proceed with closing the chamber.

      COF600 flange bolts tightened sequentially 5, 10, 20, 30, 40, 60, 80, 100, 120, 140, 160Nm.

      ML roughed CARME chamber slowly ( < 1mbar/s ) to protect Be window. Turbo started at 1.3mbar. Pirani gauge indicated e-3 mbar range when turbo at speed.


18.53 IE514 gauge ON pressure 4.05e-5mbar.

19.10 Pirani 1.42e-3 mbar IE514 3.02e-5 mbar

09.20 CARME chamber at atmospheric pressure as expected.

      Will open CARME chamber and inspect COF600 wire seal for similar (or other) issues to the seal removed yesterday. One possibility is that the COF600 flange 
surfaces/structures are misaligned and part of seal is 'escaping' the seal zone. Given the flanges close easily the misalignment would have to be small.

13.30 Observations

      The bolts on the upstream feet of the far detector chamber are all loose and have probably been so for some time

      Based on measurements with calipers and the insertion of shims into the *radial* gap between COF600 flange surfaces the COF600 flanges may be misaligned and rotated 
transversely by a fraction of a millimeter. The challenge is that the chamber is not designed for fine transverse adjustments and it is not easy to measure the changes.

      The Pfeiffer recommended procedure for closing COF seals is to start tightening bolts *diametrically* and then to tighten sequentially

https://www.pfeiffer-vacuum.com/en/know-how/mechanical-components-in-vacuum/connections/detachable-flange-connections/cof-flanges/

     The plan

     Adjust position of far detector chamber and secure upstream feet bolts

     Close COF600 flange using Pfeiffer recommended procedure

     Today is Friday so it will probably be Monday before we can pump the CARME chamber

13.50 Another view

     Grosmontage also inspected the COF600 seal this afternoon. They said that it was important that the seal should be installed under tension to ensure that it remained 
in position as the flanges were closed and the bolts begin to be tightened. They think the seals are a little too large in circumference but that they can shorten one on 
Monday morning. 

09.30 CARME chamber status

      Edwards scroll pump ON
      Edwards Speedivac valves #1 & #2 OPEN
      VAT gate valves OPEN

      Re-connect Leybold Pirani gauge to local display/control
      CARME chamber pressure 2.4mbar

      Scroll pump should be able to achieve < 0.1mbar - indicates significant leak

      Davide will first check bolt torques for all flanges opened during most recent vacuum cycle

12.05 Additional torque did not resolve problems
      ML test COF600 seal with isopropanol and observed increased chamber pressure

      Will open COF600 seal after lunch

13.55 COF600 seal removed - one section marked by Kapton tape shows some distortion and a burr on the inner diameter - see attachments 1 & 2. 


      ML speculated that perhaps the seal was lightly too long and a section of circumference was forced outwards?
      Vacuum surfaces cleaned with isopropanol and tissues - red/brown discoloration probably indicating copper observed

14.30 New COF600 seal installed, flanges closed, bolts finger tight, bolts then systematically torqued in small increments 

      COF600 torques finger tight, 10, 20, 30, 40, 60, 80, 100, 120, 140, 160Nm x2
      DN100 flanges doenstream finger tight, 8, 16, 24Nm

16.40 Start scroll pump and slowly rough chamber

18.00 Roughing CARME chamber pressure flatlines c. 30mbar indicating a large leak.
      ML ran some isopropanol around COF600 flange and observed response in roughing pressure of c. 0.2-0.3mbar at c. 30mbar which appears to confirm COF600 is leaking.


      

The CARME chamber was closed and pushed back into position on the ring. 

Prior to moving back onto the ring, the detectors were bias tested and the distance between the detectors was also checked using the motors. Bias test was performed before the detector movement tests and before moving the chamber back onto the ring.  

Attached images shows the separation between the detectors when moved in togther. Separation is < 10 mm. Motors logs will be checked to place safety limits on the movement.

The chamber was pumped down. At 2 mbar the turbo was turned on, but the pressure would not decrease below 1e-2 mbar. The turbo was switched off and the pressure increased to 2 mbar. Isopropanol was sprayed on the wire seal causing an increase in the pressure. Leak on the wire seal. 

The left (inside ring) bellow for the motors rod was replaced.

Replacing the bellow involved;

• One person holding the metal plate the detectors are mounted on (cabling remains plugged in) while the operation is performed. 
• The motors box is supported by the cryring crane by its mounting points
• Bolts on the flange connecting to the chamber can then be undone. The whole motors section (rod included) is then moved out of the chamber using the crane and someone to stabilise the rod as it slides off the detector plate. 
• Once off the chamber, the steel cover can be removed by disconnecting from the flange and from the motors box. Best to perform on a table. 
• The bellow is then removed by removing the nuts on the small flange (kf 25) connecting the bellow to the box. The bolts for these bolts are very long and go through the motors box from the moving metal plate seen in the attached image. 
• Follow these steps in reverse to install the new bellow. 

While replacing the bellow, it was noticed the long bolts for the small flange were too short. It was therefore very difficult to re-attach the nuts with any torque. The thickness of the head of the bolts was shortend to effectively lenghten the bolt which was sufficient to tighten the nuts. New bolts should be bought in case of replacing the bellows again. 

The leak on the actuator arm bellows has been attempted to be sealed using UHV vacuum seal spray. The spray had the opposite to the desired effect and the leak rate increased. When spraying with helium the leak rate had increased to 1e-9 mbar l/s. After lunch a second application of the vacuum seal spray was attempted, however the leak was not sealed. The base leak rate had also increased to the 1e-10 level up from in the morning and yesterday. With the leak increasing in size and risk of a full failure, we are no longer able to move forward with the bakeout. 

The chamber has been vented and we will move forward with plans for replacing the bellow tomorrow. 

Attached plot shows the most current pumpdown.  

CARME and the interaction chamber have been pumping for ~26 hours, P=4.15e-7 mbar, attached figures show the current pump down plot in comparison to the two previous pumpdowns in the last 2 weeks and the pump down with the bare chamber in 2020. The gradient of the pressure reduction appears to be improving for each pump down, either due to conditioning of the chamber or the reduction in leaks.  

CARME and the interaction chamber were leaked tested again today. The base leak rate has reduced significantly from 1.5e-10 mbar l/s to <1e-12 mbar l/s. All flanges on the interaction chamber were tested. A leak is still visible on the gate valve, however the magnitude of the leak has decreased significantly from 2.2e-10 to 2e-11. No other leaks observed on the interaction chamber. 

All flanges on CARME checked again, no leaks observed on flanges or wire seal. Leak was observed on the bellows of the left hand side (inner ring) actuator arm. Leak is not on either the small flange box side or the large flange on the chamber, but on the bellows itself. Image attached. Leak rate is 2e-10 mbar l/s. The actuator was moved with the pneumatic motor once to observed any effects, no effect on leak rate or the chamber pressure.

 Replacing the bellows is very challenging and will likely require removing the detectors installed on the rod. Current plan is to continue with the baking and re-evaluate post bakeout. The bellows will be baked but should not experience the same temperatures as the flanges on the chamber. Future experiments do not require extremely forward angles so that the actuator arm can remain fixed in place to not cause a n increase in the leak, the leak rate can be compensated by the large neg pumping power. 

This morning the pressure achieved was 1.44E-7 mbar. Leak testing showed an overall leak rate of 1.5E-10 mbar l/s, this is below the previous leak rate but still significantly above the leak rate from elog 141 of the CARME chamber on its own. Spraying helium, a leak is visible on the VAT gate valve on the interaction chamber (inside of the ring). The leak rate increases to ~2.3E-10 mbar l/s. No increase in the leak rate was observed elsewhere on CARME or the interaction chamber. It was decided to vent again and replace the gasket and inspect the knife edges for the valve. 

The valve was removed and the gasket was replaced. Several scratches are visible on the gasket which may be the cause of the leak (or damage may have been done during dismounting). While vented, the backing line was re-organised. The scroll pump is now on the outside of the ring for easier access (see attached image), the turbo pump was rotated to move all power cables and the backing line to the outside of the ring. 

CARME and the interaction chamber were pumped down again and the ion gauge was activated. Vacuum can be followed via the vac plot or by grafana on the carme-gsi computer (anydesk required). 

26/06/23

This morning the pressure was 4e-7 mbar. The bolts on the windows and the large flange on the bottom of the interaction chamber were tightend. No leak was present on the windows following tightening. However, the leak on the rotatable flange, although decreased in magnitude, was still in the 1e-9 mbar l/s range when helium was sprayed. The torque on these bolts was 36 Nm, which is already past the 33 Nm recommendation for M8 bolts. It was decided to vent the section and examine the flange. 

Tomorrow morning the gasket will be replaced and we will begin pumping again in the afternoon. 

27/06/23

Flange was removed and examined. Several scratches on the surface of the flange including on the knife edge were visible. These scratches were sanded away by the grossmontage and the flange was remounted. CARME was pumped down in the afternoon. Pressure was in the 1.5E-3 mbar range when the scroll pump stalled causing the pressure to increase and the turbo to increase in current to keep up. Error symbol was observed on the scroll, unknown to why it stalled. VAT gate valve to the chamber was closed and all pumps turned off to be safe, P~ 1mbar. 

28/06/23

Jan started the scroll pump on  just the backing line using the gas balast to see if too much water was pumped. Scroll running fine so the gate valve to the chamber was opened and the turbo started. 

The motors were moved using the pneumatic only, in and out for several cycles to gauge the effect on the vacuum. For the left arm no change in pressure was observed. For the right arm a small increase of 0.04e-7 mbar was observed which was quickly reduced by the turbo. 

The detectors were then biased to see if movement had any effect on the leakage current (large effects were observed in 2021 beamtime). Movement of the pneumatics saw no observable change in the leakage current above normal fluctuations. 

Both motors arms had been sucked into the chamber by 5-10mm due to the vacuum before starting the movement cycle. After one movement cycle, the motors returned to their normal fully out positions and remained fully out. 

23/06/23

Pumpdown started with the scroll pump 17:15 . The turbo-pump was started 17:55. Gate valve was closed and the turbo turned of at 19:20 due to no working fan on the turbo. 

24/06/23

Ion gauge was activated at 11:56, P= 3.7e-5 mbar. Attached pump-down curves begin from when the ion gauge was activated.

The first plot is a comparison was a calculation for water outgassing in the chamber (which has many assumptions) but is a good guide on the expected 1/t decrease in pressure over time. The second plot is a comparison to previous turbo pump-downs in 2020 and 2021. In 2020 the chamber was bare with no instrumentation. In 2021, kapton and detectors were installed, but the pump down was conducted in several stages. Pressure readings are taken from the gauge reader in the cave, the vac plot viewer online appears to show a slightly different reading. 

 25/06/23

10:30 P= 7.11e-7 mbar. Opened valve on leak cart and saw a large increase in the pressure (above ion gauge level). Closed valve to leak cart, pressure recovered quickly (1.28e-6 mbar after ~ 30 mins). Pressure in the leak cart ~3e-3 mbar. Possibly small trapped volume between valves on the leak cart. Opened valve to leak cart again, no pressure increase observed in the chamber.  

11:40 Leak testing. Base leak rate ~5-6e-10 mbar.l/s.  Spraying helium on the seal between the bottom of the interaction chamber and the newly installed valve saw a large increase to  >1e-8 mbar.l/s. Closer attention suggests the leak may be from the rotatable flange on the bottom of the interaction chamber. An increased leak rate to 1e-9 mbar.l/s was also observed on the diamond window on the inside of the ring. Spraying around wire seal and feed-through flanges on chamber on the inside of the ring gave a leak rate of 6e-10 mbar.l/s. This may be a small leak, but is most likely noise on the leak detector or drifting helium to the large leak on the interaction chamber. Cannot determine prescence of small leaks while the large leaks are active. Closing valve to leak cart saw the leak level drop to <1e-12 mbar.l/s. 

Attached images show sites of the major leaks on the interaction chamber. 

Attached are images showing the messages for normal turbo-pump operation

23/06/23

The CARME chamber and interaction chamber were connected and section 9 has been fully closed. New valves have been installed on the interaction chamber, in addition to new diamond windows. A leak cart has been connected to the backing line of CARME. Before begining the pumpdown, the manual valve between the interaction chamber and the electron target was closed. 

The pumpdown was started at 17:15,  the valve on the scroll pump was very slowly opened to not damage the berylium window. By 18:00 a pressure of 1 mbar was achieved and the turbopump was activated. By 18:20 a pressure of 2E-3 mbar was achieved, the ion gauge was activated and briefly flashed 5E-4 mbar but this is over its limit so the gauge shutdown. At 19:20, 1.3E-3 mbar was achieved, the ion gauge was activated but the reading was 1.7E-4 mbar which is still over its limit. The turbo fan was not working and temperatures on the turbo were ~50C and increasing, so it was decided to close the main gate valve and turno off the turbo overnight in order to fix the fan.

A bias test of the detectors was also performed while in vacuum, T=20.5 C. All detectors show expected V-I curve and the leakage current is ~1-2 uA lower than the previous test in air, probably due to reduced light leaks. The top right detector did initially produce >10 uA at 20 V but the adaptor card was not properly seated on the p-n feed-through. Seating the adaptor card correctly yielded the expected V-I curve. Further inspection of the pins on this feed-through will be performed tomorrow.  

24/06/23

Overnight the pressure had increased to 0.8 mbar. Opening the gate valve decreased the pressure by an order of magnitude. The turbo fan was fixed and the turbo was re-activated, the pressure dropped to 1E-3 mbar level within 20 minutes.  Ion gauge was turned on after 30 minutes with a pressure of 3.7e-5 mbar. Pump down plot and comparison to previous pump downs will be uploaded in a seperate elog. Turbo pump maximum temperature reads 46C and has been consistent for several hours. 

V-I tests completed for the right hand side detectors. Leakage currents ~0.4 uA higher than yesterday which is consistent with the increase in temperature of ~ 0.5-0.7C. 

The wire seal has been tightend further to 160 Nm. Open neg feed-through flange has also been fully tightend. 

The CARME chamber was then slid back onto the aluminium frame on the ring. Two carriages were damaged during the sliding process, which have now been replaced. Carriages were replaced by lifting one end of the upper frame section using the crane by only ~1mm. This was sufficient to slide out the broken carriage and replace it (see images to be uploaded)..  It appears there is some deformation in the rails on the temporary frame as CARME is slid across, which resulted in the damaged carriages -> further refinements are required for future slidings. The height of the chamber has been checked using the laser allignment.  

The turbo pump has also been mounted to the top of the chamber . Current plan is to close to the interaction chamber tomorrow and begin pumping.  

Internal CARME installations complete. 

• New strain reliefs (flange side) installed on each cable harness.
• 2x sets of SAES UHV1400 NEG pumps installed and connected to 2x power feed-through flanges -> continuity checked OK -> 3x UHV1400 modules gives a resistance of 7.4 Ohms, 2x UHV1400 modules gives resistance of 4.8 Ohms. 
• Pins B & D open circuit wrt ground ( DMM OL > 20M ? )
• Pin C connection to ground OK 
• Internal thermocouples connected. Middle thermocouple connected to strain reliefs, right thermocouple connected to a standoff on the detector, left thermocouple floating (see attached images). Resistance on thermocouples checked for middle and right thermocouples ~ 4 Ohms per thermocouple. 

V-I tests performed prior to closing CARME. All detectors show expected curve, but light leaks are present. Pressing my hand against the aluminium shield yielded 0.5-1.5 uA drop in leakage current. Could not turn off lights for the test due to other people working in the cave. Will repeat test when closed. Temperature in the cave ~ 21.5 C. 

Both chamber sections have now been closed. Wire seal appeared correctly positioned immediately prior to final close. Bolts closing the wire seal have been tightend to 100 Nm, plan to increase to 140 Nm tomorrow. DN100 feed-through flanges have been tightend with a torque of 24 Nm. 

Plan is to move chamber back onto the ring tomorrow and get ready for leak testing. 

Hopefully we will not be opening the chamber for quite some time so find attached images of all four installed detectors and a video of them both moving into the beam axis for future reference

Strain reliefs on the flange side of the cable harness have been installed for two detectors (top right and bottom right detectors). 

We require the strain relief parts from edinburgh in order to install strain reliefs for the left two detectors. 

Find attached images for the strain relief installed and a video of the clamped cables during a movement cycle.

A V-I test was performed for the two detectors which have all strain reliefs installed (see attachment 4).  V-I curve appears similar to the previous test (elog 398). Aluminium foil shield is now a bit battered so may have some small light leaks. Detectors were moved using the motors while bias was applied at 150V. Small change in leakage current observed (~0.01-0.02 uA)

The detectors were moved into the beam axis using the motors to check the maximum distance they can be moved in without collision with one another. 

Movement of both pneumatic motors results in a separation of ~ 4 cm between the detectors.

The servo motors for left and right were each moved in by 1 cm -> results in separation of 2 cm between the detectors

The servo motors for left and right were then each moved in by 5 mm -> results in separation of 1 cm between the detectors

The separation between the jacking plate of one of the detectors and the side of the adjacent detectors is closer than the separation of the detector silicons. The detectors could not be moved in any further than this to avoid collisions from the jacking plate

Attached are photos after movement of the detectors. Parallax may affect the distance on the ruler in the photos. 

During beam time -> Maximum distance that can be moved on the servo motors is 30 mm combined for left and right. (15 mm left and right or 10 mm left, 20mm right etc)  

The new MACOR strain reliefs on the (detector side) have been installed. See attached photos

The two pieces  of MACOR are not fully closed but we are confident that the kapton wires are sufficiently clamped.

In order for the strain reliefs to be installed in the correct poisition, two nuts on the jacking plate of the cable harness had to be removed. This is as the groove in the strain relief was not deep enough to accomadate the nut. 

Attached is also a video of the wires (not) moving with the strain reliefs mounted