Re: [SI-LIST] : Bonding options to short Busses on a die

[email protected]
Thu, 17 Jun 1999 10:14:52 -0700

The other option you may consider in terms of wire bonding is to wire bond from one
lead (or bond finger) to another lead rather than from bond pads to lead. Usually
bonding area on leads is much bigger than that on die. So, it may be able to take
multiple bonds.

Won

From: "Pat Zabinski" <[email protected]> on 06/17/99 07:28 AM

Please respond to [email protected]



To: [email protected]

cc: (bcc: Won Chang/NVidia/US)



Subject: Re: [SI-LIST] : Bonding options to short Busses on
a die

Krishnan,

There are several issues to be addressed with such a question.

Let's start with the sizes of the bond pads on the die. For most
die, the pads are between 75 and 125 microns square. These small
sizes do not allow for second bond wires to be placed on the same
pads. So, you'll need to find pads that do not already have
bonds on them or find a vendor that can "stack" bonds on top of one
another.

The "stacking" approach works for research (we've done it several
times), but it stinks commercially (i.e., low yield = high cost).

If your pads are larger, then it's sometimes possible to place two
bonds on the same pad. However, if your die are already bonded up,
the bonds are likely in the center of the pads, and you're not
going to be able to place a second bond.

Let's now say you were able to find two open bond pads that you wanted
shorted together. From a mechanical point of view, most wire bonders
perform best if there is a minimum distance between the bonds.
On co-planar surfaces, a "short bond" is often considered less
than 20 mils; anything less reduces the repeatability of the
bond (i.e., the low yield thing again).

In the worst case scenario, if you do not have any open pads to bond
to, some research shops can open the dielectrics with lasers and
bond to the busses, but this should only be considered for research;
the cost is high and the yield is low. However, it can be a
very helpful approach to fix a problem on prototype/experimental
die.

One other approach outside of wire bonding is to use conductive
epoxy. To do this, you need to find two open "areas" on the respective
busses, and an electrically-conductive epoxy is "smeared" on the surface
of the die. It's not pretty or commercially viable, but it often
does the trick on prototypes.

In all, if the die has open pads with a large physical separation,
it is commercially viable to "patch" the busses together. Essentially,
you'll simply be adding one more bond wire to the existing set, which
is no big deal. However, it is rare to find a die designed to allow
for this, and you'll likely need to re-run the silicon to make a commercial
product.

One other concern is that any surface bond wire is likely to protrude
above the other wires. Make sure this is addressed, or it might
stick out of the SSOP plastic. i.e., not good!

In terms of who can help you out, I suggest starting with the vendor
you're presently using for wire bonding. If they are not of much
help, the International Microelectronics and Packaging Society
(IMAPS, formerly ISHM and IEPS) has several companies in their
membership to help you out. I suggest contacting them (imaps.org)
and requesting a copy of their 1999 Industry Guide (like a yellow-
pages of packaging folks).

Pat

> Hi,
>
> I am investigating options to short power / ground busses using
> bonding techniques rather than having to tape out another mask.
>
> I would also appreciate information on vendors, cost, and commercial
> viability of the method.
>
> The die is bondable to 56/48pin SSOP plastic package.
>
> Thanks in advance for any help
> Krishnan

--
  Pat Zabinski                                     ph: 507-284-5936
  Mayo Foundation                                  fx: 507-284-9171
  200 First Street SW                     [email protected]
  Rochester, MN 55905       www.mayo.edu/sppdg/sppdg_home_page.html

**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****

**** To unsubscribe from si-list: send e-mail to [email protected]. In the BODY of message put: UNSUBSCRIBE si-list, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu/si-list ****