Issue link: https://iconnect007.uberflip.com/i/1389320
JULY 2021 I SMT007 MAGAZINE 19 in the pads themselves that can help give you a little more standoff. You can design a win- dowpane for the ground pad instead of one solid pad and increase standoff height. ere are a couple of QFN manufacturers I've seen that have gone to an actual windowpane on the component side as well that helps give you a little bit more standoff. I think that when we see better QFN designs that will help increase the standoff height to four mil, somewhere in that neighborhood. You'd like to see 3-5 mils anyway, to help you better out- gas those no-clean activators as well as being able to clean them. QFNs and all bottom ter- minated components are still a huge percent- age of failures. Holden: Because of density and the fact that chipscale packages and wafer-scale packages keep getting denser, with smaller solder balls, which means less and less offset, evolution is going to make that an increasing issue. Espe- cially with IBM announcing two nanometers geometries on their latest set of wafers. All of that will be more IOs, and more ground IOs; things like that add smaller and smaller pitch. Camden: You are absolutely right, and this is the price they pay. e extra scrutiny they must put on cleanliness, washing it, or what- ever it might be—that is the trade-off. ey must put a finer eye on reliability. It's certainly not a one-to-one relationship, but the smaller your spacing, the lesser amount of contami- nation is allowable. And as voltage goes up, that also has a part to play in it. Cleanliness is always important, but as you said, especially Z-height, when the miniaturization goes that way has even more importance to it. Johnson: Eric, I know I'm oversimplifying with this question, but is no-clean obsolete? Camden: No, not even close. It's the number one choice for CMs out there. As I mentioned earlier, we see a lot of failures related to par- tially cleaned no-clean flux that removes that outer shell. I am positive with no data to back it up that this trend of cleaning no-clean flux was made by someone who's never built a board in their life. ey said, "Well, it's no-clean flux. So, if we don't get it all, it won't matter. Anything that we can't see won't matter." And it's abso- lutely the opposite of that. A partially no-clean flux can be just as detrimental to an assembly as water-soluble flux being le aer the wash process. You're still looking at open active flux residues that are very hygroscopic, and then it just doesn't take that long with enough mois- ture and enough voltage to set up that electri- cal leakage path. With no-clean, it will be 95 to 5 if it's not already. No-clean is absolutely the wave of the future or the wave of the last 10 years really, but it will continue to dominate outside of certain sectors of the industry. A lot of the DoD still uses water-soluble and then cleans that. Aero- space uses a water-soluble and cleans that for their components, but there's just this mind- set that you can partially clean it and it will be okay. People are learning the hard way that this is not the truth, and that goes back to what we were just saying about the washing pro- cess with QFNs. ose are the components we see most of our partial cleaning failures with because it's so hard to penetrate that. I always say that if you're going to wash, use water-soluble. Don't make it any harder on yourself, just make sure you're doing a good job. It's so much easier to clean off the no-clean. You're not trying to break down this rosin or resin material through a two-minute wash pro- cess. Being able to properly process no-clean fluxes, whether you're leaving them in place or washing them, is just so critical. Matties: e demand is still there on a grand scale, that's for sure. ank you so much for all your time and insight. It's really helpful. Camden: It was absolutely my pleasure. SMT007