I have been involved in wire bonding since early 1973 and have put COB into production in the early 80’s. It is not easy and as mentioned earlier beginners luck will get you into trouble. There are several pitfalls;
1) Solder paste sizzle or splatter will ruin WB pads regardless on PCB finish and WB type, process and material control is absolute
2) ENIG works really well for Al bonds
4) Au bonds require thicker gold and that means selectively plated boards, like ENIG wit electrolytic or electroless Au after Immersion Gold
5) ENEPIG works well with Au WB but all finishes will change WB characteristics after exposure to reflow
6) Take care with reflow after WB and encapsulation it is better to perform SMT prior to COB operations.
7) Cleaning is critical for good WB pull strength (this is where Murphy steps in)
8) You will need plasma cleaning prior to WB and encapsulation, too clean and epoxy will run where you do not want it to go (another place for Murphy to surprise you)
9) Hire a few experts who know what they are doing or you will not be happy with the results in normal operations
I am still involved in wire bonding to polymer substrates today.
Regards,
John Maxwell
On Jan 30, 2014, at 11:58 AM, Steven Creswick <[log in to unmask]> wrote:
> Jack,
>
> Listen to what Joyce and Wayne are saying. Their inputs are all bits of the
> puzzle.
>
> It all depends on exactly what you hope to do. Many times, rather than
> combining chip & wire and SMT on one board [depending on die size and
> pin-out], it is easier to process arrays of die, glob top them and dice them
> into individual components that can be placed and reflowed with the SMT
> components. This isn't always possible if bypass caps or other components
> need to be as close to the die as possible, but it does open up another path
> for consideration. I call it the jelly bean approach.
>
> If no one has mentioned ENEPIG as a surface finish. It is both wire
> bondable and solderable... It does not bond nearly as easily as the pure
> soft gold that Wayne mentions for Au ball bonding. It is also aluminum
> bondable. One thing that I've always encountered with ENEPIG is that virgin
> incoming boards bond quite a bit differently than those that have been
> reflowed. Moral of the story - don't leave your WIP sitting around. The
> thermal history of the board does have an impact on bondability
>
> Even automatic ball bonders that can run at speeds of 7-10 wires per second
> have targeting accuracies of ±2-3 µm. Isn't that impressive, and that is
> slow... :-)?
>
>
> Steve Creswick
> Sr Associate - Balanced Enterprise Solutions
> http://www.linkedin.com/in/stevencreswick
> 616 834 1883
>
>
>
> -----Original Message-----
> From: TechNet [mailto:[log in to unmask]] On Behalf Of Wayne Thayer
> Sent: Thursday, January 30, 2014 9:39 AM
> To: [log in to unmask]
> Subject: Re: [TN] Wire Bonding Spec?
>
> Jack-
>
> Wirebonding is micro-welding at comparatively very low temperatures. Only
> works reliably if the surfaces are perfect--clean and the right metallurgy.
> Getting the right PCB finish can be a nightmare: For wirebonding, most
> prefer >25 microinches soft electrolytic ultra-pure gold over >150
> microinches nickel (electroless or electrolytic, but virtually all shops use
> electroless). And the surface has to be pretty smooth, but this is a lesser
> effect. With gold that thick, parts like QFNs can become problems (due to
> joint embrittlement), so for ChipOnBoard it isn't uncommon to have two
> separate finish requirements. If you can use aluminum instead of gold for
> the wiring, it is much easier to weld and can be welded to a wider variety
> of surface finishes. ENIG can work--depends on the exact formula and
> application process. But aluminum has poorer resolution than the gold.
>
> The big question/challenge for wirebonded COB is whether to wirebond before
> or after SMT. SMT is filthy, and as noted above, the wirebonds demand
> absolute cleanliness. But if you wirebond first, it messes up the solder
> application process. If you have a quarter inch or more of isolation
> between wirebonded and SMT, then using thick peelable solder mask can
> protect the gold through the SMT process. But we've found that to be
> unreliable in tight quarters.
>
> If you are ordering a COB assembled from another company, then creating
> specifications to assure that the results will be reliable is relatively
> easy, however all reliability is based on visual inspection combined with
> statistical process control. For gold ball bonds, you specify acceptable
> destruct test statistics for pull tests and ball shear tests. For aluminum
> or gold wedge bonds, you just specify pull test statistics. Reading Harman
> will assist you in determining the most applicable statistics for your
> application. IF your design also has challenges such as extreme
> environmental conditions or conducting heat away from the wirebonded part,
> then you'll also have to specify statistics for die bond line thickness and
> shear strength. The equipment to do all of this destructive testing runs in
> the $100K range.
>
> HOWEVER, if you want to slam some wirebonded part down on an ENIG board as a
> test case, and the resolution isn't challenging, that can be done quickly
> and cheaply with a minimal investment in equipment. It just likely won't be
> reliable (although given Murphy's cruelest law, "beginner's luck", you can
> easily be led to believe it is). I've even aluminum bonded parts down to
> bare clean copper--as I said above, aluminum bonds easily.
>
> Good Luck,
>
> Wayne Thayer
>
> -----Original Message-----
> From: TechNet [mailto:[log in to unmask]] On Behalf Of Joyce Koo
> Sent: Thursday, January 30, 2014 9:05 AM
> To: [log in to unmask]
> Subject: Re: [TN] Wire Bonding Spec?
>
> Jack,
> if you are looking for design spec, you are out of luck. Depend upon the
> capillary design, and wire bonding machine, you have set of different can
> do, can't do, regarding clearance to adjacent components, cleanliness if it
> is direct chip on board, height differential (you might need deep access
> capillary), etc. etc. The nowadays WB machine are amazing (if you pay arms
> and legs). Get a good plasma clean machine prior to bonding, like March,
> you got to go. Check UMC website, the operators runs thousands parts a day,
> no sweat. Good luck. (if you are running low volume, get an CM-elite
> prototype house, set up from ground up for WB is a bit of too big a task.
> IMO).
>
> Joyce Koo
> Researcher
> Materials Interconnect Lab
> Office: (519) 888-7465 79945
> BlackBerry: (226) 220-4760
>
> -----Original Message-----
> From: TechNet [mailto:[log in to unmask]] On Behalf Of Steven Creswick
> Sent: Wednesday, January 29, 2014 5:17 PM
> To: [log in to unmask]
> Subject: Re: [TN] Wire Bonding Spec?
>
> Jack,
>
> All good responses thus far - George's books will provide a great deal of
> process and reliability considerations, as well as some plating info. Mil
> Std 883 will define visual criteria, pull strengths, and ball shear
> strengths. ASTM F458, 459, and 1269 essentially repeat the Mil Std. I'm
> sure there is an IPC, JEDEC, or something spec which also repeats the Mil
> Std, but I'm an ol' Mil Std boy and I don't know what the other ones are,
> off the top of my head.
>
> Can you clarify a bit more specifically what you are looking for?
>
> Are you looking for plating detail, wire bonding process detail,
> design/layout details, packaging options etc.?
>
> It is my observation that one doesn't want to just casually get into wire
> bonding. Checking things out is a good first step.
>
>
> Steve Creswick
> Sr Associate - Balanced Enterprise Solutions
> http://www.linkedin.com/in/stevencreswick
> 616 834 1883
>
>
>
> -----Original Message-----
> From: TechNet [mailto:[log in to unmask]] On Behalf Of Jack Olson
> Sent: Wednesday, January 29, 2014 4:17 PM
> To: [log in to unmask]
> Subject: [TN] Wire Bonding Spec?
>
> newbie question:
>
> Do any IPC publications discuss wire bonding?
> What we are trying to do is similar to COB, but this is a new subject for
> us...
>
> thanks,
> Jack
>
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
>
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
>
> ______________________________________________________________________
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please contact helpdesk at x2960 or [log in to unmask]
> ______________________________________________________________________
______________________________________________________________________
This email has been scanned by the Symantec Email Security.cloud service.
For more information please contact helpdesk at x2960 or [log in to unmask]
______________________________________________________________________
|
