TM#3 and its new source tray have been both moved into the south hot cell.

I moved5shielding blocks from ITE to block storage area, This is a job for preparing ITE vacuum surge investigation and leak check.

The water block sealing surfaces on the module-side of TM3 were polished this week by Chad Fisher, using his air-ratchet controlled rotating head tools (single and double-surface polishing tools).

Photos of the polishing tools in-use are attached (from August 16 - 21, 2013)

Scotch-brite pads were used, followed by 2000 grit sandpaper, then white felt with isopropanol, then lint-free dry polishing pads, each on detachable heads that are used with the polishing tool.

The module-side water block surfaces were inspected by Grant Minor on 21 Aug 2013 using a level telescope and camera looking through the hot cell window.  Results can be viewed here, with photos indicating which block is being viewed in sequence.

https://documents.triumf.ca/docushare/dsweb/View/Collection-11251

After a great struggle with the water block jigs, the TM3 source tray was finally advanced fully into the service tray and all water blocks were connected and torqued to spec in the hot cell by Chad Fisher.

Notes and description of problems encountered during the installation:

- The water lines for the optics tray interfered with the module-side water block jig, and had to be bent out of the way to clear while advancing the tray (see attached photos)

- The water block jig on the source tray side was too low relative to the module side, and had to be eventually unbolted completely from it's mounting bracket in order to raise it up into alignment

- A piece of aluminum plate was placed under the jig, and the pneumatic table was raised in order to move the jig upwards

- After many repeated attempts to engage the blocks by pressing the jigs together, it was discovered that some of the blocks had rotational misalignment with their respective counterparts on the module side, preventing the pins from engaging properly

- By looking through the bolt holes on several of the blocks, it was possible to determine which direction they had to be rotated

- Chads polishing tool was used to apply some torque to rotate the source-tray-side water blocks slightly so proper alignment of the pins could be achieved

- See the attached photo which indicates which blocks had to be rotated (blocks 5, 6, 9, 10, and 11)

- The blocks had the bolts installed and were also torqued in the numbered order shown in the attached photo

Next steps are to remove the water block jigs, re-connect the loosened brackets for the remaining water lines (entrance window, optics tray), re-install the containment box and associated VCR connectors, video inspect, and leak check.

While continuing on the TM3 Rev 3 source tray installation, Chad noticed today that one of the steerer wires is broken and has come out of the thermocouple-style connecter block (see attached photos).

From drawing ITA2826 (Rev D02 attached) it is likely that this is for either the "upper" or "right side" steerer plate.

Today Chad finished making all connections to the new source tray, except for:

- steerers - broken wire, must be repaired at a later date

- IMG gauge - bracket required, old bracket thrown out with previous tray by accident, see ITA3301, must be installed at a later date

- entrance window - the window lines on module side were left blanked-off for leak checking, containment box must be removed in the next few days to complete steerer repairs so these lines were not connected to save work from having to be repeated

The containment box was replaced, and TM3 was moved to the CS for pump-down and leak checking (see other e-log).  David reported at 6:15 pm that the turbos were turned on at the CS and the vacuum looked normal (so far).

As stated above, Chad realized today that he had thrown out the old IMG gauge bracket with the previous source tray, and a new one must be manufactured.  The assembly is described by ITA3301.  Three components must be manufactured (ITA3302, ITA3303, and ITA2673).  A stock split-ring clamp for ITA2673 (MDC #716001) was graciously donated by Dimo Yosifov.  Grant Minor submitted a work order to the machine shop today for 2 each of ITA3302 and ITA3303.

Target Module 3 with no target, and a partially-installed Rev-3 source tray (no steerers connected, no IMG connected, entrance window lines blanked off on module side and not connected to containment box) was transported successfully to the Conditioning Station by Travis Cave and David Wang at around 4pm today.  David attached the vacuum system for start of pump-down to prepare for leak checking.

A huge leak had been found on HS circuit during 60 psi helium pressurize test. Other circuits are free of obvious major leak at present .We will test them again once the TM3 is back to CS. See attachment for details

The heat shield circuit was pressurized with air (15-20 psi) and snoop applied to the water block joint and brazes. Bubbles formed at the joint between the two water blocks indicating that the leak is a c seal problem. Pictures attached but also on docushare Collection-11285.

chad fisher wrote:

The heat shield circuit was pressurized with air (15-20 psi) and snoop applied to the water block joint and brazes. Bubbles formed at the joint between the two water blocks indicating that the leak is a c seal problem. Pictures attached but also on docushare Collection-11285.

 Last night prior to this leak check the Heat Shield was pumped on and helium sprayed on the circuit. The pressure was 0.0 x 10^-4 Torr. The leak was traced to water block. response time was 2 seconds and went from 5.5 x 10^-8 to 1.6 x 10^-6 atm*cc/sec. There was also a response without spraying helium (drift from the nozzle). The results of this testing prompted the testing completed in Chads elog.

 September 3rd

David Wang and myself lowered the leak testing cart into the Ante room and prepared the ante-room for the leak testing of TM3s line. David and Grant then entered the anteroom and inspected the old c-seal and leak tested the the heat shield line. The line pumped down to the lower limits of the leak detector and there was no response from the cart. A presentation with the results of the C-seal inspection results was completed by Grant Minor and email for a design review held on the 4th.

September 4th

David Wang and myself removed the leak detector and surveyed the anteroom and cart. The following results were found with the high spec gamma detector.

Leak Detector body and cart 0 

Leak Detector wheels 200 cpm

Floor off of plastic 300 cpm (there was contamination before)

Plastic in front of HC opening 2000 cpm

Working table after plastic was removed < 50 cpm

Tool box < 50 cpm

First tacky mat 2500 cpm

Second tacky mat 200 cpm

outside of door tacky mats 0 cpm

David then laid fresh plastic on the floor around the HC opening and covered the exposed floor as well.

Following this Travis and David measured the lengths of the wiring and attempted to separate the wiring harness from the Ultem block. It was found that the block was pinned and could not be separated, a new block will need to be made. The measured the following lengths:

Steerering - 26" to 27"

Collimator - 32" - 33"

PNG - 60"

Today Maico measured the thickness of three indium-plated inconel c-seals, presumably Ultra-Seal P/N 50606.  These seals had NOT been in the Hot Cell.

Uncompressed (new) "thin" c-seal

OD .435"

ID .300"

Thickness .094"

Uncompressed (new) "thick" c-seal

OD .438"

ID .280"

Thickness .096"

Compressed (used) "thin" c-seal (Beginning of Life bench test compression only, not long term compression)

OD .438"

ID .294"

Thickness .079"

Maico commented that he measured the thickness of a few other compressed "thin" c-seals and they all had .079" thicknesses.

The seal space when the blocks are bolted together (shown on ITA2342 Rev F) should be .070" (steel insert CB) + .002" (recess on module side)  + .002" (recess on source-tray side) = .074"

Thus the compressed seals seem to have .079" - .074" = .005" spring-back.

C-seal drawings from Ultra-Seal and Garlock in chronological order from Guy Stanford's design file are attached.  Guy's whole design file is also attached for reference.

Yesterday, Grant entered the Ante Room with David Wang to visually inspect the two sets of failed c-seals from the Heat Shield circuit, to remove the second set from the Heat Shield water blocks, and to leak check the heat shield circuit with o-rings and a water-block-to-leak-detector fitting.

Some photos of the inspected seals are in the attached design review presentation.  Some unusual marks were observed on both sets of failed seals.

The heat shield lines and water block assembly leak checked successfully to the bottom of the leak rate range ("UNDER" ~1x10-9 atm cc / sec on the Varian 979) with a great flood of helium on all joints with no response anywhere.

 Today Maico completed the new leak testing tool and blank off. He tested them with C-seals without springs and all was leak tight. He then inspected the crushed seals and noticed that there was an area that was crushed less on both seals (more prominent in one,  20130906_tm3sourtrayefurb_P1020975) and that this reduced crushed zone only appears on one side of the seal. This is similar to the failure seen on both sets of the heat shield lines. On the seal with the greatest change in crush zone the average crush zone was ~0.025 and the reduced crush zone was ~0.012". He inspected the leak tester and the blank off and determined they were within tolerance and that they had a total seal goove height of 0.074" which is nominal. New seals were selected and one had a visible dimple prior to crushing (20130906_tm3sourtrayefurb_P1020978)  and the other had some defects on the inside (20130906_tm3sourtrayefurb_P1020962). These defects were marked and crushed using the same leak testing tool and blank off. The seal with the dimple showed a reduced crush zone in the same area (20130906_tm3sourtrayefurb_P1030021) and the seal with the defects on the inside showed less or undetectable change in crush zone. Maico then inspected (20130906_tm3sourtrayefurb_P1020991) and crushed the "thicker seals" (more indium coating). When inserting the seals he noticed that they fit tight on the counter bore of the blank off. He then attempted to crush the seal to the point where the copper faces would touch (as design intent) but the seal locked. The gap between the 2 copper faces was measured to be ~0.001". This setup was leak tested and found to be leak tight. When inspecting the crushed thicker seals it was found that the material had actually been pushed sideways causing a lip to form around the crush zone (20130906_tm3sourtrayefurb_P1030003). Maico then fitted the wires for the testing of the module and the heat shield line. 

Bevan Moss wrote:

Today Maico completed the new leak testing tool and blank off. He tested them with C-seals without springs and all was leak tight. He then inspected the crushed seals and noticed that there was an area that was crushed less on both seals (more prominent in one,  20130906_tm3sourtrayefurb_P1020975) and that this reduced crushed zone only appears on one side of the seal. This is similar to the failure seen on both sets of the heat shield lines. On the seal with the greatest change in crush zone the average crush zone was ~0.025 and the reduced crush zone was ~0.012". He inspected the leak tester and the blank off and determined they were within tolerance and that they had a total seal goove height of 0.074" which is nominal. New seals were selected and one had a visible dimple prior to crushing (20130906_tm3sourtrayefurb_P1020978)  and the other had some defects on the inside (20130906_tm3sourtrayefurb_P1020962). These defects were marked and crushed using the same leak testing tool and blank off. The seal with the dimple showed a reduced crush zone in the same area (20130906_tm3sourtrayefurb_P1030021) and the seal with the defects on the inside showed less or undetectable change in crush zone. Maico then inspected (20130906_tm3sourtrayefurb_P1020991) and crushed the "thicker seals" (more indium coating). When inserting the seals he noticed that they fit tight on the counter bore of the blank off. He then attempted to crush the seal to the point where the copper faces would touch (as design intent) but the seal locked. The gap between the 2 copper faces was measured to be ~0.001". This setup was leak tested and found to be leak tight. When inspecting the crushed thicker seals it was found that the material had actually been pushed sideways causing a lip to form around the crush zone (20130906_tm3sourtrayefurb_P1030003). Maico then fitted the wires for the testing of the module and the heat shield line.

 Hello all,

As an addendum to Bevan's e-Log:

Maico also prepared eight (8) new retainer spring windings out of the .025" diameter stainless welding wire for the next seals that we will attempt in the Hot Cell and Ante Room

Maico, Bevan, Keith and I had a discussion about the bolt torque related to the c-seal and water block compression:
- It was found that not much torque is required to compress the "standard" design "thin" c-seals (Ultra-Seal P/N 50606 .001 - .0015 thou indium plating) - basically hand tight only with an allen wrench
- By hand-tightening the bolts with an allen key until the faces of the blocks came together, and then measuring the torque with a torque wrench, Maico discovered that about 8 foot-lbs = 96 inch-lbs (or 5/8ths of a turn past finger tight) was required
- The c-seals are fully compressed when the faces of the blocks are contacting... any additional torque applied is only to pre-load the bolts to ensure that they do not come loose due to temperature cycling and mechanical vibration
- Chad's Hot Cell torque tool is nominally set to about 168 - 180 inch-lbs (or about 14 to 15 foot lbs) based on an e-mail update from him 3-June-2013 - this torque is normal chart torque for a 1/4"-28 UNF SAE Grade 8 bolt pre-load of 3,250 lbs
- see http://www.imperialsupplies.com/pdf/A_FastenerTorqueCharts.pdf
- This amount of bolt pre-load torque may not be required if the bolts are SAE Grade 5.  I recommend at this point that the minimum possible pre-load torque be applied to achieve the chart recommended pre-load for the grade of the bolt, which should be investigated
- This is to ensure that we do not over-stress the thread inserts on the water blocks... if these inserts are damaged, we will basically render the service chase unusable and ruin the module

Cheers,

Grant

 The sealing surfaces of the module side heat shield water block were polished on September 9th in the morning using Chad's polishing tool with the following attachments:

- Scotchbrite pads  (30 seconds)  (after this step indium material from the previous seal was no longer visible on the sealing surface)

- 2000 grit sandpaper (30 seconds)

- Felt material with isopropanol (30 seconds)

- Lint free pad (30 seconds)

After the final step the surface was rinsed with isopropanol and air dried using canned air.  The water block blank-off was then installed with the manipulators, tightened until snug using the air ratchet, then tightened an additional 1/4 turn while the water block was gripped firmly using one manipulator.  The seal was leak checked and passed successfully.

TM3 has been connected in CS.Three turbo- pumps have been started. The pumping down is fine so far.

 This e-log is to cover the work completed from September 9th until the 11th. 

September 9th

Today Maico completed the leak checking blank off and leak checking tool (to replace the one that was contaminated). This tool was then leak checked with C-seals and no springs at the machine shop, both halves were determined to be leak tight. 2 sets of small c-seals without divots were selected and used for leak checking the module side and the source tray side. The heat shield line was tested in the ante-room and was determined to leak. Upon further investigation it was determined that the bore of the seal gland was determined to be over sized (0.450"). Isaac then installed the blank off onto the module and pump down was started, the testing car would not stabilize so Isaac tried turning the screw more allowing for further pump down. The leak check was then completed and there was no response. The flow of helium for the target module was checked but was much lower than typical as the second valve was not fully opened. 

September 10th

With the belief that the larger bore in combination with the small c-seals was the cause of the leak Maico found and polished another block that also had a larger bore. Once polished, small c-seals without dimples were selected and tested with the machine shop leak testing cart. These seals were leak tight but upon investigation the crush zone in one section of the seals was significantly reduced. The smallest width of the crush zone was measured to be ~0.007". This result is concerning as it means that larger bore water blocks can be sealed with small c-seals but once installed there is no way to tell one block from the other. There is no data on the lifetime of this combination of seal and bore. 

Following this test a small c-seal with dimples was selected and checked in the large bore. This combination resulted in a significant leak. Following that test the large c-seals were crushed in the larger bore water block. The torque required to crush the seal and to have the faces touch was more than that of a small c-seal regardless of what size bore it was crushed in. However with the larger bore the large c-seals could be completely crushed allowing for the copper faces to touch (the total force on the faces is unknown). The crushed seal was then inspected, it had a nice uniform crush zone that was relatively large when compared to a properly crushed small c-seal. Another testing with the large seal and a 0.025" wire diameter spring was completed. It was found that the wire interferes with the the seal when it exits from the groove. This interference causes a localized increased crushed zone but does not reduce the crush elsewhere or prevent the copper faces from touching. At this point it was decided to repair the heat shield using large seals. Large seals with springs were tested on the actual HS line in the anteroom and yielded that same results. The line was then installed by Maico inside the HC. It pumped down to the lower limit of the leak testing cart and sprayed with helium, no leak was detected. The containment box was then put back on by Isaac and Grant in preparation for a module move.

September 11th

Today the containment box installation was completed and the module moved. Following that Maico and myself entered into the ante-room and retrieved the blank off block from the HC via the tool port. The large c-seal used in the test and the small ones used in the blank off were then collected and bagged. The large c-seal had typical measurements. The small c-seals used in the blank off did not have typical measurements. They had not been compressed as much and the crush zone was near impossible to see (if visible at all). From the measurements it is believed that the block was not sufficiently tightened. It also calls into question the validity of the leak check due to the combination of less helium and reduced crush. 

After these measurements the tools that were contaminated but were to be recovered were bagged and checked by safety. They have been moved to the jacks area where Maico will decontaminate them. The old wiring harness from TM3 was bagged and given to safety for storage in the cyclotron tunnel. The plastic and all of the garbage was then lifted, bagged, and removed from the ante room. Safety surveyed the anteroom following this no alpha contamination was found but there was 5000 counts on the floor of the ante room. The contamination may have existed prior to this work.

Many pictures of the cseals have been taken and have been put into \\trwindata\remote handling\Photos\2013\2013_tm3_source_tray_refurb. A report detailing all of the testing and results would be invaluable.  

I leak checked TM3 water lines at CS with 60 psi and 80psi helium. The base leak rate is 5.0xe-9 atm.cc/sec. Each circuit 3 minutes helium pressurizing for each time test. No any respose had been found on leak detector during the test. See attachment