Screaming Circuits: September 2010


Random Via-in-Pad Myth #3

Myth #3: It’s okay to put vias in BGA pads because the extra solder (the solder ball) on a BGA will prevent any problems

9x13 via in pad BGA land Here's an example of a via in pad situation we don't like; A large number of vias in the BGA land pads. They're pretty big via holes too. The primary reason we don't want to see vias in pads is that when left open, those via holes act like little capillary straws and suck solder off of the pad. A couple of undesirable events can happen depending on the method used during board fab.

If your vias are left open, solder will tend to wick down into the via hole. The larger the diameter, the BGA via in pad worse the wicking problem can be. You might end up without enough solder left to secure the component, or even a solder bump on the bottom side of the board which could interfere with other components or lead to shorts.

The extra solder in the BGA ball can just make a bigger mess on the back side of the board. And the extra space that BGAs have between the component and the PCB make opens even more likely. BGAs are more susceptible to this problem. Not less.

Some people will try to use soldermask to cap the bottom side of the via. However, the solder ball can still be wicked down into the via, leaving an open. Or the soldermask cap map pop off due to thermal expansion or out-gassing, leading to a purely open via. Internal air bubbles can migrate up, leading to voids in your solder joint.

Duane Benson
Burma Shave

Dewy Defeats Truman!

No he didn't!

Dewey Defeats Truman But why do I care? Why do I state this? Well, an email went out from an independent external survey administrator on our behalf. The email was referring to a customer service survey - we do that now and then; ask our customers how we're doing. Our customers are nice and we like hearing from them - but that's not the point.

The point is that the subject line of the email was pretty misleading. It read "Final Reminder - Screaming Circuits Closing September 30th." I don't know about you, but if I saw that subject line on an email from my favorite PCB assembly place, I'd be rather worried.

Well, worry no more because we're going to the White House, I mean we're still healthy and happily assembling PCBs and will continue to forever. Yes. I know that "forever" is a long time, but that's the way we think around here.

By the way, I know that "Dewey" is misspelled in my blog post title. I did that on purpose to throw off the copyright police. And I used a picture from the Truman library, not the cool famous one owned by some big newspaper someplace.

Here's some humor to lighten the mood.

Duane Benson
Mark Twain said it too

Funky QFN Land Patterns

I've described the optimal way to create your land and solder paste layer for QFNs a couple of times before. Complex QFN land pattern But that was for a standard square QFN or rectangular DFN. What happens if you look at the bottom of your QFN and it's all weird like this one?

Does it require a different philosophy for the big pad areas? Should it just be a solid opening because their is more than one thermal pad and they don't cover the whole area?

Well, this pic is an Intersil ISL8200 power module. It's pretty cool and Intersil was kind enough to actually put the paste layer recommendations right in the data sheet. Unfortunately, not all chip manufactures do that.

The bad news is that it's a pretty complex pattern. The good news is that the data sheet gives a diagram with great detail on the required dimensions for the lands and the stencil. And, yes, you treat this just like any other QFN thermal pad. They recommend 50 - 80% paste coverage for the thermal pads just like everyone else. That means that you'll segment the paste cut-outs in the paste layer for each of the four thermal areas just like you would for the whole pad area on a standard QFN. The data sheet for this part has the specifics.

For similar parts from other manufacturers, you should go to their datasheets and app notes first, but if you don't find a recommendation, we would suggest you do the segmenting and shoot for somewhere between 50 and 80% coverage. Putting down too much paste is a bad idea for any QFN or DFN, but it's probably even more critical with a part like this where the solder areas only cover half the part. If there's too much solder on the underside, it will likely tilt and most likely not solder reliably.

Duane Benson
Don't eat paste.

ESC Boston, 2010

I just got back from the Embedded Systems Conference in Boston. We go there every year, along with the ESC show in San Jose. After the gloom at the show last year, I was pretty apprehensive.

With so many mixed messages relating to the state of our economy, I really had no idea what to expect. Last year, our booth traffic was decent, but the theme of the day was "I'm between jobs". Not good.

ESC Bosto 2010 day one Looking back at the two days of this year's show, I'm feeling a lot better about the state of at least the engineering segment of our economy. We started the show with a mob of people at our booth - all five of us blue shirts had people talking with us and a few more in waiting. I don't think I stopped talking for the first two hours. After that, traffic slowed a bit and I was able to snap a few pictures. We had good crowds throughout both days of the show.

We had our best opening day in the four years I've been here. There was no gloom like last year. I only spoke with one between jobs engineer the whole day. The throng of folks at the ESC booth for the backpack giveaway was intense. They were giving away 500 backpacks full of goodies, including a Ti robotics development kit. I'm not sure, but it looked like they got rushed by all 500 people just as the doors opened.

We may not be out of the economic woods yet, but based on my view from the show floor, I'm feeling very upbeat and certainly am confident that our industry is back.

Duane Benson
"I love it when a plan comes together"

Re-runs

I've been pretty occupied with the upcoming Embedded Systems Conference in Boston. The exhibition is next week on Tuesday, the 21st and Wednesday the 22nd. Screaming Circuits will be in booth 809. Stop by if you happen to be at the show.

In any case, I've been pretty much wrapped up in show preparation so I haven't had much time for original writing here. That being the case, I'm going to play an old TV sit-com trick and just select some old, but good, content to re-run.

And, there you go.

Duane Benson
Hide Wally Bee. Andre is back and he's got a fly swatter

Four Years and Blogging

The Screaming Circuits blog celebrates its fourth year blogging today (September 14, 2010). In that time, it has had 447 (448, including this one) posts and 250 reader comments (more if you count the spammy ones I deleted).

Popular subjects have been:

A year ago, I noted that the blog had 311 posts, or about 100 per year. That gives 137 posts this last year, Ant_wideweb__430x317 so apparently, I've picked up the pace a little bit. Some other likely non-interesting statistics: that's about 140,000 words, a million key presses and 459 images. My fingers must be tired. Maybe bruised on the tips. Only four of them should be sore though. I skipped all the typing classes and came up with my own variety of modified two-finger hunt and peck using just the index and middle fingers of both hands.

For the most part, all of those images, key presses and words have been attempts at passing on useful information and some diversionary blathering with hopefully not too much self-serving drivel. I think I've avoided getting into any political discussions during that time. Nothing controversial. That's probably best given that this blog is the voice of Screaming Circuits, not the voice of me. Maybe one or two of my tep-ten lists may have come close though.

Duane Benson
Live long and preserve all raspberries

Questions on SMD and NSMD

My prior post discussed SMD (solder mask defined) vs NSMD (non solder mask defined) pads for BGAs. I received a comment on the post asking for a bit of clarification:

"My understanding was that SMD pads were preferred from a solderable area standpoint. To use an example from your picture, the pad in the upper-left corner has a larger surface area than any of the others by virtue of the larger exposed trace leading to it. With an SMD pad, this trace would be masked and the solderable area would be identical to the others. Have you found that other maneuverability factors outweigh this concern?"

IPC-7905B does recommend NSMD pads for the most reliable solder joint. They do note however, that sometimes SMD pads are used to prevent pad lifting so it's a bit of a trade-off. Their suggested compromise is to use mostly NSMD pads and SMD pads in the corners. I guess it you have solder joint issues, make sure that your PCB uses NSMD pads. If you have issues with pad lifting, try the SMD corners. If you have a lot of pad lifting, you might have a warping problem that needs a little more attention.

As far as the assembly process goes, we can build them either way without problems - unless there's an underlying issue causing problems. The difference in pad area isn't an issue for us - unless you also remove the soldermask from the trace going to the via. Doing that is bad.

Duane Benson
Pop goes the weasel is okay
Pop goes the BGA is not

SMD vs. NSMD

In general, we and just about everyone else on the planet recommend NSMD (non solder mask defined) pads for your BGA land. Using copper to define the land gives more precision than using solder mask and, more importantly, gives a more reliable solder joint.

BGA on HASL close with vias between pads There are a few cases where SMD (solder mask defined) pads might be more appropriate. The Beagleboard folks, with their 0.4mm pitch Ti OMAP processor found that at that small a pitch, NSMD pads increased the chances of bridging.

Interestingly, IPC writes that the main advantage of SMD pads is to prevent pad lifting. They further note that since the corner balls are most likely to have lift issues, due to the greater concentration of stresses. If you have concerns about corner pad lifting, you might want to use SMD pads just in the corner for a little extra holding power. (ref IPC-7095B, 6.2.2)

Duane Benson
Danger! 50,000 Ohms

Easy Reading for a Long Weekend

The holiday is upon us and most folks here in the US will have a three day weekend. Of course, when you're an engineer on deadline, all too often holidays don't really mean that much. Here's a little food for thought for those that will be working over the weekend.

  • If you're trying to finish off that layout and need some advice on a pesky QFN or DFN, read these few bits about laying out for a quality reflow: here, here and here.
  • If you're trying to decide what finish to order on your PCB, read this, this and this.
  • If you just want to confuse yourself a bit, try this, this and this.

Now you can get back to some real problems - like finding that last little bit of clock jitter or figuring out how to keep the back-EMF from mucking with your MOSFETs.

Duane Benson
No shorts allowed under that BGA, 'cause shorts cause tan lines

« August 2010 | Main | October 2010 »