Ed didn't say in his original e-mail if the "padless" connections were on IL, OL or both. Happy's comments appear to be limited to OL padless. Has there been testing done to see if the same reliability benefit is realized if padless innerlayer connects are used?
Mike McMaster
RF Product Engineer
Merix Corporation
503-992-4263
----------
From: Happy Holden[SMTP:[log in to unmask]]
Reply To: TechNet E-Mail Forum.;[log in to unmask]
Sent: Wednesday, September 17, 2003 8:33 AM
To: [log in to unmask]
Subject: Re: [TN] Reliability question
Hi Ed,
You will probably get a lot of different opinions on this subject based on your questions. And, YES, microvias boards are being designed today with landless vias. Three pictures of them are attached. The unique thing about landless vias, is that extensive reliability testing have shown them to BE MORE RELIABLE THAN VIAS WITH LANDS! But this is mainly in Europe, because we have not done a lot of testing on this while the Europeans have.
Here is what we found when we tested landless vias nearly a decade ago on small and large THs in thick and thin boards.
In 1988, a large U.S. OEM was working with OKI of Japan. OKI was shipping multilayers with the surface having landless vias. This did not match their standards and IPC Specs, but the Japanese insisted that these were reliable. Not wanting to insult the Japanese, this OEM used their own PC Fab facilities to build and test the landless vias as well as vias with lands of various annular rings, materials thicknesses, finished hole dia and plating thicknesses.
Some results were predictable and other very surprising: Small holes, thin copper plating and high aspect ratios all led to early hole failures. But the holes that refused to fail were the landless ones. Even when the plating in the landed-small holes were thin, the same dia. holes as landless on the test boards had sufficient copper to be reliable (this later test showed was a current density thing).
This data was turned over to the PhD Reliability Scientist to explain. Not only did they explain it, they created a modified Coffin-Manson model that predicted the early failures of the small annular ring holes. The work was so significant, that the decision was made "on-high" not to publish the data and let everybody in on the 'secret'. Landless vias are a significant performance, density and reliability advantage, so why inform your competitors about it. Landless vias have been in production on high-density and form-factor driven products at that OEM ever since.
The issue with your boards is that they were not designed or built to be 'landless'. They may be landless now, but it also sounds like they have SOME annular ring on the holes due to misregistration. True 'landless' TH of 0.062" material and what sounds like 2 mil of plating per wall should be reliable. BUT, holes with 1, 2 or 3 mil of annular rings are the ones the OEM found failed early. The cause was corner cracking. As the material was thermal cycled, the smaller the annular ring, the higher the angle of the land to the barrel, and the earlier it cracked and then propagated around the hole until it severed the trace connection. So your boards may be in this case and not one of true "landless".
Thermal cycling is one way to test it if you can afford the time and materials. Maybe a faster way is to use the new HATS testing machines that some commercial labs have (Highly Accelerated Thermal Shock). New sample coupons will have to be made but the results look like they predict what will happen in service or by IST testing (at least Delphi thinks so). Their Web site is: www.hats-tester.com There is technical papers there, contacts and coupons.
Not a lot has been published on 'landless' and small annular ring holes. Good Luck
Happy Holden
Westwood Associates
I have a question concerning the reliability of a PCB that has little to no
annular ring on the via holes on the external surfaces of a multilayer
board:
-there are about 100 such holes on the board
-all components are through hole
-the board is four layers, 0.062" thick
-solder mask fills most of these via holes
-the pad size is about 0.024", with a hole drill size of 0.020", and
finished hole size of about 0.016"
-the trace coming into the hole is 0.008" wide
-there are a few locations where a trace enters a hole and there is no land
(pad) at the junction of trace/hole, just the width of the trace
-the specification is for 2 ounce copper finished amount
The board house has 100% electrically tested the boards for shorts and opens
and found no failures, and believes that the boards will be reliable. We
have assembled, wave soldered and tested (before discovery of the problem)
about 30 of the units and have found no failures related to this issue.
The boards do not meet IPC-A-600 specifications. The customer designed the
board and did not call out any IPC spec.
My concern is the long term reliability of the trace to hole junction. It
seems that I remember hearing of high density boards actually being designed
without pads: the traces came directly into the hole. Has that been done
successfully?
Is the reliability questionable? Will successful thermal cycling of the
assemblies indicate long term reliability? Any thoughts or suggestions will
be appreciated.
Thank you,
Ed Berti ---------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
-----------------------------------------------------