West Bond Wire Bonder

West Bond Wire Bonder


Description

The West Bond Wirebonder is for electrically connecting our MEMS chips to standard packages. We mainly use 40-pin DIP packages, but we also have workholders for TO-5 and TO-8 packages.

The bonder is currently set up to use 1 mil (25 um) gold wire through the 45-degree wedge tool. The machine performs bonding by using the wedge to mash the wire against the device surface, while applying sideways ultrasonic vibration for a short amount of time. Heat is also applied by a radiating coil around the wedge tool. Novice wirebonder users can use contact pads of 200 by 200 microns in die area, whereas more experienced users may push the threshold down to 150 by 150 microns or smaller. The size of the wire, when mashed down, determines the minimum bond size, and this is about 75-100 um.

 ... to top


Usage Policy

Several training sessions are required in order to use the wirebonder, which will last roughly an hour each.

The West Bond bonder has a memory for digitally storing recipes for various wire materials and package types. Most users will find "NORM" sufficient, however there are 40 spaces available, so you may store your own recipes if you like. You are encouraged to claim and name your own memory area if you need to make any changes to the basic recipes. The "NORM" recipe is suspected of applying too much power to the pads on some chips, shorting them to the substrate. To help with this, there is a recipe called "LITE" which is recommended if you are concerned with this.

If you are listed as an inactive user, please talk to the existing users or superuser before using the machine to learn about any updates that may have taken place recently

 ... to top


Contact

Name Office Extension Email
Nathan Lazarus HH 1209 Link Link

 ... to top


Qualified Users List

Name Office Extension Email
Active Users
Gregory Barchard   Link Link
Kristen Dorsey HH 1209 Link Link
Peter Gilgunn HH 1209 Link Link
Larry Schultz HH C114 Link Link
Mustafa Emre Karagozler SH 422 Link Link
Jingwei Liu ReH 337 Link Link
Katey Lo ReH 336 Link Link
John Reinke HH 1209 Link Link
Cheng Yuan Wen   Link Link
Amy Wung HH 1209 Link Link
Amelia Wright   Link Link
Inactive users
Umut Arslan HH A212 Link Link
Sarah Bedair HH 1209 Link Link
Peter Buetler     pbuetler@andrew.cmu.edu
Bahareh Behkam     bbehkam@andrew.cmu.edu
Sergio Camacho      
Chando     chando@andrew.cmu.edu
Malcolm Gregory     mgregory@andrew.cmu.edu
Jichul Kim     jichulk@andrew.cmu.edu
Anna Liao   Link Link
Bo Li     boli@ece.cmu.edu
Chun-Teh Li     chuntehl@ece.cmu.edu
Shahab     shahab@cmu.edu
Brian Smith     brs@andrew.cmu.edu
Keivan Yazdani     kyazdani@andrew.cmu.edu
Shu Zhang     szhang@andrew.cmu.edu
Industry Users
Brett Diamond Akustica    
Jeff Hibner Akustica    
Todd Keebler Akustica    
Handong Li Bridge Semiconductor    
Matte Zeleznik Akustica    

 ... to top


Standard Operating Procedures

The following gives you some idea what is involved with the operation of the bonder. This is far from a complete description, and you really need hands on training to understand it. You must get training from before using the machine.

1) Register that you will be using the bonder on the sign-in sheet.
Follow the sequence in the text below as a guide:

These instructions are not meant to be taken as permission to use the bonder without training. All users must meet with Chiung-Cheng Lo (chiungcl@ece) for training.

As of October 2002, we have a new wire bonder for the lab, a West Bond 4546-E79. It is similar to the old K&S wire bonder we had in that it does ultrasonic wedge bonding. However, the machine-human interface is very different, in that much of the movement of the wedge tool is automated. You can also do manual bonding, but my experience so far with the machine is that the automatic bonding is extremely accurate and easy to use, so I recommend this mode. The machine is also less likely to become unthreaded in automatic mode.

Users of the old machine will be delighted to learn that it doesn't need to be threaded very often. The main way the machine could come unthreaded is if the user moves the bonding tool backward and the wire comes out of the clamp. This is extremely unlikely in the fully automatic mode, and the wire feeding configuration of this machine also helps keep a stable path for the wire. Another way unthreading happens is if the wedge tool runs into an obstacle of some sort and the wire is knocked out of place. try not to do this because that can also damage the tool. If the tool does come unthreaded, please ask me for assistance, because threading this machine is quite tricky. The wire must come down a narrow tube on its way to the clamp. This is REALLY hard to thread and I don't want to repeat it. If the wire comes out of the wedge only, it is not as big a deal, being only slightly more difficult than the old machine to thread.

Here is the general sequence for operating the machine in fully automatic mode. It serves as a reminder of the steps to follow, assuming you've alrady been trained in person.

Turn on machine, crosshair generator, TV monitor

If you are using gold wire (that is how the machine is set up right now) turn the tool heat setting to 4.0 and wait a few minutes before bonding. Don't touch the metal coil around the wedge tool, because it will be very hot.

Position sample in the workholder under camera. Note that when you can see through the camera and monitor, the wedge is actually pulled back (away from operator). It will snap back under the camera (and block its view) under many circumstances, like when it's actually bonding.

SET WORK HEIGHT

edit menu (4)

machine settings (6)

restart ht (9)

work height (7)

It is best to do this so the work height is optimized for the chip surface, because then that will be in focus. Press "G" then wedge will come down (and won't stop in this mode- be ready to position landing spot). Follow instructions- turning knob clockwise raises work table. The machine will beep and the display will briefly read "Good" when you get there.

BOND-OFF/ALIGN CROSSHAIRS AT HEIGHT OF CRITICAL BOND

THREAD/BOND OFF menu (6)

You want to do this on your chip (the critical bond surface) because the position of the crosshairs in the camera is dependent on height because of parallax. press "0" to do it slowly, "G" to do it faster. After the bond is made, manipulator arm will lock in place, and then you take your time to set crosshairs on bond, and hit OK (G).

SET HEIGHTS, WIRE LENGTH

Next you will set the bond heights and the wire length. Before you do this, you need to make sure that the two bond targets are lined up in a straight line away from you. Do this by using the "7" key (LOCK/MOVE X) along with the manipulator handle to go back and forth while you adjust the angle of the workholder on the table. When you get the angle adjusted don't touch the workholder anymore.

Make sure that the package finger is the location of your first bond, and your chip is the second bond. Before going into the edit loop motion menu, use the manipulator arm to aim the crosshairs at the first bond, and then don't touch it again until you aim the crosshairs just before making your bond. Now you go to EDIT LOOP MOTION to set up the machine to make the bond:

EDIT menu (4)

EDIT LOOP MOTION (0)

This will take you through a sequence of steps that set the search and loop heights, and the wire length. Generally, if you are using a recipe that is already tested for your chip and package, you will only need to make adjustments to the wire length ("Y offset"). But it can't hurt to go through the adjustments listed below, in case there are some variations in the package thickness or the chip thickness.

Before you go through the steps by pressing "G", though, you should set the Y offset (the wire length) to something small, e.g. 100. This is important because you can damage the machine if the wedge is run into the package wall, which can happen because the machine doesn't ask before moving the wedge while stepping through the items using the G key. Do this by pressing 4 to go to the "edit loop" menu, where you will see displayed all the parameters of the loop motion. Here you can select them individually, and edit without actually moving the wedge. Select "Y offset" and key in "100" for the Y offset of bond 2. Now you are safe to go through the steps using the G key.

The first step that you will need to adjust is search height for bond #1. This is done by positioning over the 1st bond area (the instructions will lead you to this), and hitting SUGGEST (8). The wedge tool will come down, contact the surface, and pop back up to the appropriate height. (If this doesn't happen, make sure that the wedge is not in contact with the surface already. If you think this is the case, press "inch up" (5) until the wedge rises above the surface. This applies to bond 2 as well.) The next few settings you can leave alone (Z-before-Y, backbend) or use suggest again. Use SUGGEST to choose a loop height. Unless you have an unusual height difference between bonds, whatever the machine suggests will be fine. Next step is to set the Y-offset (which is the length of your bond wire). Make sure the height of the wedge is above any surfaces so you don't hit the side of anything, and use 5 and/or 0 to move the wedge approximately above the bond pad. You will have a later opportunity to fine tune this position. Next few options will be search height #2 and something about the clamp. Use suggested value for the clamps. The search height adjustment is similar to bond #1. Position the wedge above the surface, and hit "8" for "SUGGEST". Then you will have a chance to set the final adjustment of Y-offset. You can also go back to "previous option" as many times as you want to get it right. When you are satisfied, continue through the options by pressing "G" and you eventually pop back to the home menu.

MAKE BOND

Most of the time you want to bond from the package to the chip, which means you want to aim the crosshairs at bond 2, because you want to aim where you have the least room for error. This bond where you aim just before bonding is called the "critical bond". Make sure if you are expecting the critical bond to be X that the display says "ALIGN BOND X OF 2". Otherwise, when you do the bond, the machine will start where you are expecting it to end or vice versa, and end in some possibly bad place, like on your structures, smacking into the package wall and dethreading or damaging the tool.

If the critical bond setting is not what you want, then edit the wire type (a dumb name, but that's what it's called):

EDIT (4)

WIRE TYPE (9)

EDIT TYPE (4)

CRITICAL BOND (7)

To make the bond, aim the crosshairs at the critical bond pad, and press "G" and wait until the bonder makes the bonds. You can see the result through the TV or microscope. Most people prefer to use semi-automatic rather than full automatic mode, because you can stop (by hitting the "3" key) before bonding to the device, in case something goes visibly wrong with the first bond. To toggle between full- and semi-automatic, press the 8 key twice.

THINGS TO WATCH FOR:

wire comes unthreaded from wedge: notify Chiung-Cheng Lo

Two most important buttons: "3", which wil take you back to the home menu no matter where you are, and abort whatever you're doing (even in the middle of a bond, which is useful if you see the first bond didn't stick). "4" takes you from the home menu to all the edit menus. You need to remember this one because it usually isn't shown on the LCD screen.

make sure machine is in full/half automatic mode: More options (8)/ Toggle mode (8)

make sure there is enough tail (use FEED, then bond-off)

crosshairs drift from true target. Use bond-off occasionally to reposition. Camera mount may need to be tightened.

To name a recipe, go into the "Learn mode" menu, and the naming option is a couple options into the sequence. DEFINITIONS:

Many of the menu items may be hard to find. There is a complete map of the menu system on pages 30 and 31 of the West Bond manual.

critical bond: the bond on which it is more important to have precise placement of the bond. Usually the bond pads on your chip are much smaller than on the package, so for most cases this is the bond on the chip.

device: It is possible to store 40 different "recipes" in the bonder, and give them short names. I've created a few already, but some are for the aluminum wire settings. The settings for making the bonds (power, force, time, etc) are associated with each device. "NORM" (device 1) should work for chips on the DIP packages we have around the lab, using gold wire.

You can copy devices into other locations and edit the copies if you're doing something new or unusual or if you just want your own device recipe. Go to EDIT (4)/DEVICE EDIT to copy from one location to another. You can create or edit the name by going into learn mode and going to one of the first few options.

bond edit: You probably shouldn't mess with this. I've programmed in settings that work well with our wire, packages, and chips.

wire type: This you can edit for the number of bonds per wire, and set which is the critical bond.

loop edit: Here you can edit the search and loop heights, as well as the Y-offset (length between 1st and 2nd bond). It is the same as "learn mode", I think, but you can access things randomly rather in sequence.

deep access wedge: Currently there is a 45 degree wedge installed, meaning the wire is fed from a clamp that sits at a slight angle behind the wedge. This means the clamp can run into the package walls if you are bonding close to the walls.

Updated April 13, 2004.

 ... to top


Application Notes

Chiung-Cheng Lo sent out these suggestions on May 17, 2005:
  1. CHECK the wire by OM making sure the wire is right under the central wedge before bonding
  2. strongly recommend use "0" instead of "G", don't do SEMI-AUTO
  3. if wire is right at the center of wedge, it's IMPOSSIBLE not getting bonding.
  4. talk to my or Peter personally, if you get into any difficulties.

 ... to top


Useful Links

 ... to top


Miscellaneous

John Neuman set up much of this web page in 2004.
You can also ship your chips off-site for bonding and packaging.

 ... to top



overview | projects | people | publications | intranet | resources         © 1998-2009  Carnegie Mellon