Thermosonic Wedge Bonding


In the past, if we’ve been doing smd soldering, we’ve used pretty basic hot plates. This project takes that idea a bit further.  Since [kc6qhp] will be using parts that aren’t conducive to soldering, he has to use wire bonding. After locating a fairly cheap wire bonding machine and microscope, he built the heated stage to fit perfectly with his other tools.  You’ll notice that he has machined a lip around the heat plate for small custom C-clamps as well as made it adjustable height. Very nice work [kc6qhp].

12 thoughts on “Thermosonic Wedge Bonding

  1. i’ve played this game at work :D
    glueing dies down with a pair of tweezers and a toothpick is the hardest bit in my opinion, tho i was using a semi-automated wire bonder.

  2. Thanks for the link!

    So far I’ve only tried glueing down one die and yeah its really hard! Although I am uing arctic silver epoxy which is horrible for this use. Luckily its a low power chip so I hope not to have thermal issues. In any case, some real silver epoxy is in my future which will require me to build a dedicated freezer. There’s still a couple more things left to get the bonder working perfectly. I’ve made a ton of mechanical modifications, as the bonder was originally set up for insulated wire bonding, not 1 mil gold wire which I what I’m using.

    Luckily where I work there are a number of skilled wedge bonding techs whose brains I have been picking for learning what is important and what isn’t.

    Once I have the controller finished I’ll put up an instructable on the entire process.

  3. Just to say I hate instructables.

    You have to go through pages of parts lists and instructions (as if it was a cooking recipe) when all you wanted was the good idea that makes the project work.

    It’s really obscure to see for instance a step named “machining the block” when you have no idea what that block is for… and you will not have that idea until you’ve read through and analyzed all the pages… what a waste of time.

    Make a proper website guys… especially for interesting projects like this.

  4. I actually have a need to re-wirebond a chip. (I took something apart and didnt realize the bonds were just there in the open and I think I broke a few. Can anyone point me in the right direction to get this done?

  5. I agree that Instructables is a pain and poorly put together for this kind of stuff. I think its put together in a way to increase advertising views more than anything. If I’m interested in a project I either download the pdf of the project or “view all steps on one page” which makes it a lot easier to understand. One or both of those options require a login. I really ought to start a blog, but I’m too lazy to figure it out, so Instructables is the next best thing.

  6. mike m:

    Kinda depends on the type of chip, bond wire, etc. If its a chip that uses gold metallization (typically GaAs MMICs, also InP (but thats pretty rare)) chances are they use gold wire or ribbon for bonds and I could try fixing it for you. If its a silicon chip they frequently use aluminum and potentially use ball bonding rather than wedge bonding. I don’t have any aluminum wire and its hard to say whether I could fix that. There are companies that do wire bonding, but you are likely going to ahe to pay a lot (think at minimum hundreds of dollars) unless you can get to know someone who does it, etc.

  7. Hey guys, very rare that wedge bonding/ball bonding comes up in a hacking group. Feel like I should at least say hi.

    I am an engineer for a major automated wire and die bonding manufacture (not that hard to figure out). If you have any questions on process setting or design I would be more then happy to help out.

  8. matt: I’m definitely interested in picking your brain. email me at tony wavelen dotcom.

    I have a few questions… So what I’ve got is an older Westbond manual bonder that was originally used for disk drive manufacture. It came with no wire spool, no clamps, etc. The bond parameters (time and power) are set up for only a single setting, whereas I may need two settings for first and second bond (MMIC and then substrate). So I’m going to make a controller to set those. As far as bond pressure goes, I’m calibrating with a small pocket scale, and using just maybe 10-20 grams of force, 200 to 500 ms long. Now the big uncertainty I have is with respect to power. Apparently this unit has a ~20 watt transducer, but I have no idea what units the setting is in. Do you know of a way to measure that? Or even to extrapolate from electrical measurements?

    Also, I’m not sure if my unit is phase locked, is that a critical feature?

    Finally… I’m planning (obviously) on heating up the part to 150 C and then bonding with cold wire. Westbond typically heats their wire before bonding, how critical is this?

    I’ve made some cold bonds on a piece of alumina substarte and had mixed results depending on the settings. Some come out perfectly, others result in a sheared tail, and others just don’t get good adhesion. I think its going to be a matter of refinement to get these to work well, but hopefully adding heat will help.

    I’m using Gaiser Tool wedges, made from tungsten carbide (I believe) and the slightly concave tip as reccomended for gold.

  9. Have you measured the amplitude of the sine going to the bonder transducer? I wanted to compare with mine. It’s hard to find data on these. I was surprised that it wasn’t as high as expected: about 10vpp at 64khz

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.