XTek XT H 160 by Metris, Nikon
Contact Steve Lynch
Original: 06/07/10 by Brian Chow
Update history: 06/08/10 (maintenance) BYC
05/10/11 (changing between targets) KCC
10/14/11 (chiller maintenance) KCC
THE MACHINE MAY NOT BE USED BY INDIVIDUALS WITHOUT ONLINE AND IN PERSON EHS RADIOGRAPHY SAFETY TRAINING, IN ADDITION TO TRAINING BY CBA STAFF. If John D doesn’t know that you have been trained, and by whom, then you may not use the machine.
The x-ray source of this machine has a beryllium window, which could be shattered by crashing of samples/fixtures into it. Beryllium dust that can result is EXTREMELY TOXIC. It is your responsibility to be aware of the consequences of inhaling this dust. If you crash something into this window, do not open the machine door, turn off the machine, leave a notice so that nobody else opens the machine, and contact John D.
If you need to make adjustments for which you have not been trained, or there seems to be a problem with the machine, email John D, kenny, and nadya (john.difrancesco@cba.mit.edu, kenny@cba.mit.edu, nadya.peek@cba.mit.edu) and please wait for a response. Excessive and unnecessary contact with Nikon regarding issues that we can help with may damage our ability to get real more difficult problems fixed in a timely fashion.
PLEASE, DO NOT MAKE PHYSICAL ADJUSTMENTS TO THE MACHINE UNLESS YOU HAVE BEEN LIVE TRAINED TO MAKE THAT SPECIFIC ADJUSTMENT. READING THIS SOP DOES NOT SERVE AS SAID TRAINING. This applies to stuff in red, below. There is cctv coverage of the lab areas, and your access privileges can be permanently revoked.
Machine overview
Cabinet X-ray machine - completely interlocked (door must be closed to power x-ray source)
>0.5 mrad/hr (FDA), company spec is >0.1mrad/hr at full power
Up to 160keV (wall thickness max for Al is a few inches, a few mm for steel, soft materials not limited)
60W tube
Electrical – system draws~8A, chiller ~8A also
Devices internally are 240V, stepped up from normal 120V via transformer
5- axis turntable
14-bit flat panel CsI detector (1k x 1k pixel)
1Gbit Ethernet dump to computer
Red power switch turns off system and PC, so SHUT DOWN THE PC FIRST, unless it is an emergency.
Green light turns on computer, too
Water chiller has light that says coolant failure on machine if off
Water chiller in back has on/off switch
Closed loop water chiller, distilled
White button “x-ray power” is for vacuum (roughing and turbo, synthetic oil)
Can do VBA programming/scripting
Can feed in wiring to control objects (left side of machine). Addition of wiring/cables must be approved by EHS radiography before use.
ask John D for engineering password
Getting interlock and conditioning the tube
Lower left of software “I” = information
Key must be turned to get interlock
Will have readings for vacuum (needs 152 arb units to reach interlock, can take up to 2hrs)
Set Point “BAD” means that the vacuum level is inadequate. If the vacuum level doesn’t change after an hour or so, then there may be a problem with the vacuum pumps (this has happened a few times before). Contact John D.
Won’t get interlock if red light won’t go on (internally in machine, there is a test circuit)
Once interlock, must condition tube to clean it (vac is for tube only), best to do it every time, but necessary when turned on, and if using high kV
Bring to max voltage
Won’t reach max voltage if not conditioned (will auto turn off from vaporizing gas if not good vac) if you turn on
Hit the lightning bolt to condition (will condition to what you set)
When conditioned, it will hold at slightly higher than setting
Leave for 2-3 minutes
Homing the axes (button on computer, top left)
Moves to hit hard limits, may hit the source so DON’T put anything in there before you home
Must be done every time software is opened
Software and control:
Top left: manipulation
Bottom left: x-ray control
Right: Image editing
Motion: Joystick (KEEP AN EYE ON THE OBJECT SO YOU DON’T CRASH)
Speed control of joystick is set by arrows on software; joysticks are analog in addition
Y-axis: up and down vertically
X-axis: to and from you
Z-axis: to and from the tube (Magnification b/c spot size/focus from tube determines mag factor, 3um spot size)
For magnification: + is larger mag, but number on software is smaller b/c distance from home (in mm)
10um resolution of motion
Rotation: 360 deg continuous (4th axis)
Tilt (5th axis): +/- 30 deg
Holders are carbon fiber – affix with tape our built in tweezers, align to notch and lock with set screw
Setting kV and other params
Main control on detector is integration time to limit kV and amps
Lower kV is more contrast (want energy to match the material, generally)
If too bright from panel saturation, just get vertical lines on screen. Good way to find is to get train track artifacts and then lower until goes away to max the dynamic range
When finding, use high kV for short integration for fast feedback, then lower kV and increase integration time for higher quality images
Can use tone curve as well, but losing bits (adjusts panel response counts in vs. counts out – like brightness and contrast and splines on photoshop or Zeiss)
Engineering route (AVOID THIS): Moving detector close increases sensitivity but loses resolution, moving farther vice versa. Linear magnification
Low contrast, low atomic numbers – plastics and bio – low kV (~50kV, keep above 40kV to prevent air attenuation)
Voltage is like spectrum, current is like brightness
Most photons are low power, but can put Cu filters to cut out soft (lower power) x-rays to limit saturating the background if high power (with less x-rays coming out of machines)
Keep power below 10W – max is 60W ~ 160kV X 500uA – to prevent damage and ensure good resolution
Detector will get “train tracks” around value of 52k
Focus slider on on xray control is fine focus of the spot of the beam (EM lens)
Acquisition buttons:
Capture – button on screen (Red button), how many frames to capture
Live image is also button on screen
Recursion – averages images in live mode
Properties – toggle has hand pointing to paper
Exposure – frame rate
Binning max 2 x 2
Accumulation – Average frames (BEST TO AVOID)
Digital gain – keep at a 1
Can save the preset, and it will save in the pull down under acquisition parms
ALWAYS CHECK PARAMS BEFORE SCAN B/C OPENS TO LAST ACQUISITION SETTING
Can save positions: manipulator set
“Inspections” is their tool for programming in lieu of direct VBA
Hard to stitch images b/c point source (parallax issues)
On left screen – there is digital magnification
For measuring in 2D, the measurement and edge detect (button) is going to be inaccurate b/c it assumes object is in center of stage, but if object is closer, then it will think the object is larger.
You can manually override with calibration
IF YOU REALLY WANT ACCURATE MEASUREMENT – Use stage itself (should be same plane)
Reset zeros of stage measurement for a reference point
Look at new measurement value
So, the limit of the resolution is basically how flat your object is and how in-plane it is
Saving and post-hoc processing
Saving: D drive is machine, Z is mounted on computer, both have USB drive (find out where JohnD prefers for saving)
Best to save as grayscale with tone so it applies the correction but you have raw data (i.e. can use Auto contrast but get both versions)
XML will file be saved with image that will tell you the settings used to acquire image (can enter notes into XML file if you hit “NOTES” for things like whether you used a filter)
Can’t use XML files to automatically load settings
Need VG-Studio license to look at the 3-D volume from surface data, as opposed to exported TIFFs (lookup if ImageJ will read it, or write our own)
Auto can be done AFTER capture and then recaptured
Negate is true inversion (i.e. people who look at real xrays)
Best to not apply filters beyond tone, like sharpen, contrast, etc. b/c will lose information and can apply post-hoc in software
Can record movie (under save as “record movie” - saves as uncompressed avi – hit cancel and it will stop saving)
CT SCANNING
Rules of thumb:
Keep object in FOV as rotating
Penetrate part at all times for good definition (so need to check against the thickest portion)
Do not saturate the detector (use filters to harden the beam)
Get position of object
Rotate so it stays in FOV as it rotates
Move as close as possible while still keeping in FOV as rotates to maximize resolution (FOV/1000 pixels)
Get good values for penetration (can check penetration value by moving cursor over object in image; check the light/dark areas for dynamic range)
Right screen: CT window
Setup
Number of projections: Averaging to reduce noise
Optimize checked (to reduce number of projections per projection, to save time)
Click rotate to make sure it stays in FOV
Flux normalization (don’t normally use)
Once sample is setup, it remembers sample position
Can load parms from a previous job
Next: Shading correction to correct for background
Move object out of the way
Create a shading-file (can use 19 points to calibrate, but really only need 3)
Next tab: Centre of rotation (can skip this; can do in software)
Next tab: Reconstruction setup
Hit acquire – Takes two scans at 90degrees of each other
Adjust the crop areas
Next tab: Acquisition
Choose computer to send image
Name image (it will append a date/time)
Enter the x-ray filter
Uncheck “reconstruct”
Minimize ring artefacts (fast if checked, higher quality but longer if unchecked)
Fast mode (artifact unchecked) takes about 30 minutes per projection
Quality mode (artifact checked) adds about an hour per scan
Hit acquire
Bottom right of the computer screen – CT agent – will tell you if receiving data
Can’t use software to reconstruct while dumping
Each image is 2MB, and 1000 images, so 2GB – USB2.0 on computer
Reconstruction:
CT Pro (CUDA) – can’t reconstruct while imaging
File open (.xtekct)
Image: blink (see if image moved - if it moves more than your feature size, the scan is trash)
Next tab: Centre of rotation – pick two areas
Accuracy to standard, hit start
Next tabs: Setup and calibration (can skip, see manual)
Next tab: Volume
Can change resolution
Hit start
Reconstructed image is the .vol file, VGI file is a text file of parms
Open in VGStudio
For manipulation, look up manual
Do not store your files on the machine, they are routinely deleted.
MAINTENANCE
Guts of system:
Left back – Roughing pump (change the filter every year; oil window should be 2/3)
Left front – step-up transformer
Right – motion controllers
Water in back - there is a water level meter in software (1:20 algicide/anti-corrosive to water)
Replacement parts:
Lamp 30W (Compton), 240V, 221mm
Cabinet lamp 240V standard fluorescent lamp
Fuses
Polish – if there is buildup on filament holder – clean with polish and then clean extremely well with IPA to remove the polish
Changing the filament
If you move the detector, need to tell computer that the source/panel distance has changed
If filament is burnt, get kV but no current when turn on x-ray
Cambridge Microscope A054X – SEM filaments
Must use POWDER FREE GLOVES
Turn off the vacuum by hitting white button on machine
Let vent for 3 minutes, and then hit the black button on the tube for 1s to vent the system
Wait a minute, then once more for 1s
Wait a minute, and then push vent until the chamber stops filling
Unlatch the chamber, and it will swing open partially
Unlatch the hinge to open it all of the way
If it won’t open easily, then it needs to vent more
Check the o-ring quality
BARELY ANY VACUUM GREASE
Pull out the holder (must jostle in/out from spring clips)
Holder has no polarity
If needs cleaning, use IPA
Loosen set screw on side (metric allen wrench)
GENTLY pull out filament with pliers by the wire; make sure not to bend the filament; make sure to push it all of the way down
Want tip to be just above the bottom of the cone/aperture (!!!)
Can adjust height of the filament with 3 set screws on the bottom of the cone
Try to center the filament manually (no set screws in the part itself)
Put the part back in
Open the information window in software
Hit the white button on the machine to start the vacuum after closing machine
Starts at 255 and should decrease immediately; there is a leak if not
152 on the gauge to engage the interlocks; should take 5 minutes
10^-7 mbar is ~ 152 on the gauge
Condition the filament
Should get about 600 x-ray hours, but only guaranteed for 200 x-ray hours
To align the tip
Like wobble correction for SEM
Put a circuit very close to tip (i.e. in cone without touching filter) for max magnification (Holder materials: foam, modeling clay, beeswax for biological samples (pliable), double-sided tape, plasticine)
Focus on bond wire at 10 frames/second
Properties, xray controller, focus
“Two triangles” toggle button
Correct at 75kV, 110kV, 160kV
Adjust x-shift and y-shift until it stops moving x-y (should just be coming in/out focus)
Should be in the ballpark of -30, -30 (shouldn’t change much b/w filaments)
After all three adjustments at different powers, then capture images at variety of kV to see if images are good
Properties, xray controller, focus
Filament demand – keep moving up until the image stops getting brighter
Adjust filament demand at different currents and kV to get a good balance (original setting is 80)
Reset filament lifetime
Shading (gradient looking software toggle) - hit acquire – will adjust the background to account for irregularities and dirt on the panel, beryllium window, etc.
Shading only affects the real-time imaging, not the CT acquisition
Do for every frame rate
Can disable the shading correction under properties window
To change the target material
There is a choice of tungsten, molybdenum, copper, and silver target materials, selectable via a knob on the target tip.
“The target material in use is the material that is closest to the marking line. There should be a 7mm wrench [in the left hand side of the cabinet] for this purpose. You just rotate the target – no other adjustments are needed.
I’ll be honest [... I just] use W for dense materials and Mo for low contrast materials . I seldom if ever use the others.” (Steve Lynch)
The target piece is threaded and travels across the x axis when turned, so there is a limit to how much it should be turned in either direction. If you always turn the adjuster CCW to select between targets, the target will unthread (vacuum pressure might keep it seated, but it being unthreaded is likely poor for stability). If you always turn the adjuster CW, a coolant outlet hole on the bottom of the target assembly will start to drip. This would be the inner limit - time to reverse direction for adjustment. Because these limits are easy to blow past, please turn the adjuster slowly - like more than eight seconds per revolution. For everyone’s sanity, please leave the machine with about ten turns of thread showing - fewer than nine turns of thread showing is likely to start making it drip, and we don’t know (nor do we need to find out) how many turns of thread would show if it were to be backed out too far.
Adjusting the panel
Move the z-axis stage all the way to the left so that if you move the panel, it doesn’t hit the stage
Quit the program
Open “Inspect Engineering Tool” program
Enter pwd
Open manipulator in software
Setup tab: “Imaging to source distance”: Closest (380mm), Middle (580mm, default), Farthest (980mm)
Axes tab: “Axes travel positive”: Closest (100mm), Middle (300mm), Farthest (700mm)
DO NOT CHANGE ANYTHING ELSE
Hit save
Exit
old as of march 2011
quit inspect x
move the detector into position
start inspect x
select panel position
start
Chiller maintenance
The chiller water eventually depletes by evaporation, and needs to be refilled. It has a low water level alarm, and a little container light will show. It wants distilled water.