Differences

This shows you the differences between two versions of the page.

--- decap:ceramic [2016/12/05 17:30] – mcmaster
+++ decap:ceramic [2020/10/21 22:24] (current) – [Glass frit] mcmaster
@@ Line 10: / Line 10: @@
 ====== Metal lid ======
+{{:mcmaster:decap:ceramic:magnet.jpg?400|}}
+Above: 8751 decapped using a torch + magnet
 See also: [[decap:metal|Metal]]
@@ Line 21: / Line 25: @@
 ====== Glass frit ======
+-10-21
+  * General consensus may be that heating to below melting point (softens but doesn't liquefy glass) followed by shearing may be the best approach for most
+  * Mind heat and cool cycles to avoid thermal shock
 JM 2016-11 advice:
@@ Line 33: / Line 41: @@
 Above: typical gentle result
+{{:mcmaster:decap:ceramic:img_20180311_194832.jpg?400|}}
+Above: using a spring clamp to better hold the ceramic. Preliminary results indicate this allows much better control than using pliers to hold the top. Possibly could crack and collapse though
 Gentle procedure:
@@ Line 40: / Line 52: @@
     * DIP28: ~10 seconds
     * DIP40: ~20 seconds
+    * Or: I started pre-heating using heatgun instead. IIRC 500F was a good point. Higher and it didn't give good enough thermal gradient
   - Begin torching top of chip
     * Use the side of the flame, not the tip, for gentler and more even heating
@@ Line 76: / Line 89: @@
   * If the package cracks, finish procedure and then assess damage.  The spreader should keep package integrity until assessment is done.  If the die cracks, the chip is probably a loss
+{{:mcmaster:decap:ceramic_glass:cracked.jpg?300|}}
+Above: typical crack due to not pre-heating bottom
+===== Bandsaw =====
+{{:mcmaster:decap:ceramic:bandsaw_sample.jpg?400|}}
+Above: fixturing for cutting. Sample held in DIP 40 ZIF. Use cross slide to stablize.
+{{:mcmaster:decap:ceramic:decap_saw_cut.jpg?400|}}
+Above: cuts on far sides of cavity. They are above wires but not the die
+{{:mcmaster:decap:ceramic:decap_saw_tape.jpg?400|}}
+Above: tape placed to keep lid in place as its sheered off. Excessive movement may damage bond wires. Tape also helps retain large pieces if it fractures.
+{{:mcmaster:decap:ceramic:decap_saw_peel.jpg?400|}}
+Above: lid peeled off after sheering
+DB-100 diamond bandsaw
+-12-16 procedure for large ceramic packages:
+  * Place chip in DIP-40 socket to help steady during cutting
+  * Saw a slit on both sides of the cavity
+    * As far as possible, usually about 0.1" after the quartz window
+    * Saw just into the glass frit
+      * WARNING: failing to completely separate ceramic can lead to unpredictable breakage
+    * Should be roughly above where the bond wires attach but not above the die
+    * Small debris in the cavity is normal and shouldn't hurt the die
+    * Pins make good guides. I usually saw between two
+  * Fill with nail polish to help protect the die against debris
+    * Use a small pipette to pump it into the cavity from one side
+    * Should push out any water from sawing
+  * Alternatively, for small packages it may make sense to simply grind (see below)
+  * Place in vice with top cavity just above jaws
+  * Place tape over top of package
+  * Gently tap side of lid with chisel to sheer it
+  * Peel away tape and lid together
+  * Wash nail polish away with acetone
+    * See UV masking page for more info
+Grinding:
+  * Grinding was used more initially, but can be slower (especially for large packages) and may have increased change to damage die
+  * Alternate side of blade used to prolong blade life
+  * Slowly side grind and/or cut across the whole cavity
+  * Cut about halfway through the top ceramic
+    * You shouldn't need to go all the way to the frit
+  * Try to make one continuous cut
+    * Multiple cuts increase change to fragment and create larger debris
+  * Wash nail polish away with acetone
+    * See UV masking page for more info
 ===== Thermal fracture (old) =====
@@ Line 81: / Line 149: @@
 {{:mcmaster:decap:ceramic_glass:post_drip.jpg?300|}}
-{{:mcmaster:decap:ceramic_glass:post_both.jpg?300|}}
+{{:mcmaster:decap:ceramic_glass:post_top.jpg?300|}}
-Above: upper chip was opened using this technique
+Above: example result
 Procedure
@@ Line 144: / Line 212: @@
 While this does work, I've had much higher success rate with "top heat" procedure.  The main advantage of this procedure is that, if successful, leads to a very intact package.  However, given that many attacks involve masking the die, the other techniques also provide better die access
+===== Mill package =====
+{{:mcmaster:decap:ceramic:cerdip_milled.jpg?400|}}
+Above: test slot. The 1/8" or so endmill broke after 3 CNC passes despite water cooling, slow speed, and shallow depth
+{{:mcmaster:decap:ceramic:mill_wc_diamond.jpg?400|}}
+Above: only diamond shaped tooth carbide endmills seem to be able to remove ceramic reasonably. From reading online, only carbide or diamond tools should even be considered for ceramics. Diamond studded endmills also exist, but are a bit expensive so I haven't tried them
+===== Laser =====
+W CO2 laser can melt package, but it didn't evaporate. Also rest of package get severe thermal shock and cracked here and there
+{{:mcmaster:decap:ceramic:co2_0.jpg?300|}}
+{{:mcmaster:decap:ceramic:co2_1.jpg?300|}}
+Window was easily melted, but had a tendency to slice through the bond wires below
 ===== Shearing =====
@@ Line 180: / Line 267: @@
 ===== Dissolve =====
+{{:mcmaster:decap:ceramic:hno3_before.jpg?300|}}
+{{:mcmaster:decap:ceramic:hno3_dissolving.jpg?300|}}
+{{:mcmaster:decap:ceramic:hno3_dissolved.jpg?300|}}
+Above: 70% HNO3 dissolving glass frit and (silver epoxy?) die attach
 % HNO3 is known to dissolve the frit.  However, it would probably dissolve other things (ie the pins) as well
@@ Line 217: / Line 312: @@
   * Re-heat package
   * Slide die out of open side of well with die
+-12: tried using talc to weaken glass and make it come off.  Didn't work, but probably worth exploring related techniques.  Maybe just a light dusting?