|Ask the Experts|
October 22, 2020
HASL Surface Finish and Coplanarity
We are an EMS contractor. We have a customer that has requested a PCB with HASL surface finish. This assembly includes a 256 ball BGA component. The component balls are 0.6 mm. Should we be concerned with coplanarity? Are there any other concerns?
|Expert Panel Responses|
Be concerned? Be afraid! Be very, very afraid!
HASL was wonderful back when it was invented in the mid-70s for through-hole PCBs. With early SMT with 0.050" pitch components it was adequate. But once we went over the 0.025" pitch threshold, HASL floundered. It is about the most non-coplanar surface finish you can come up with.
In these days of BTCs and area arrays, where every micron of solder paste thickness is crucial (as in make it or break it) the uneven topography of HASL just doesn't cut it. Consider any of the other surface finishes (ENIG, Immersion Ag, OSP - with their caveats) but unless you want a lot of headaches, skip the hassle with HASL.
This is acceptable, but you will need to be placing the BGA with "Paste Additive"process. You should also ensure you have a PCB house that can ensure good HASL coverage with minimal crowning on the pads. This is commonly attributed to cleanliness prior to HASL application and some PCB houses are better than others at processing HASL.
Another consideration is solder type. We would use Type IV powder with better than 90% solder content. This approach along with a extended soak in the liquidus phase helps to mitigate solder voiding that would be considered defects. Slowly ramping up temperatures in the active phase for the flux allows the volatiles to outgas and then extending the liquid phase encourages the voids to dissipate better in the solder pedestal of the BGA.
PCB layout should be analyzed to provide consistent thermal dynamics as well. We preferred to have 4 quadrant escape patterns and I would not recommend 'via in pad' for this solder system, due to the time needed to soak the PCB.
A good target for soak time was 3 - 5 minutes and Solder liquidus phase for HASL / Paste Additive described above for our production line is 7 - 10 seconds minimum.
Capital Equipment Operations Manager
Specialty Coating Systems
HASL coplanarity will vary by method (horizontal vs. vertical) and by alloy (Pb-free versus SnPb). I'm sure you want repeatability; to that point here's the IPC's comment on the subject*:"Hot Air Leveling (HAL) or Hot Air Solder Leveling (HASL) processes are considered to have a degree of difficulty in their control.
This, coupled with pad sizes and geometries placing additional challenges on such processes, places the creation of a practical minimum thickness outside the scope of this specification." Variable thickness, and lacking control thereof, can be seen as the root-cause of coplanarity issues.
When it comes to playing the odds on terminal flatness consistency, HASL is your worst bet.
* excerpt from IPC-6012, Qualification and Performance Specification for Rigid Printed Boards
Circuit Connect, Inc.
For BGA components down to about 0.8mm pitch, HASL is an acceptable finish. Your component, with a 0.6mm ball diameter, surely has a pitch larger than 0.8mm;it's most likely 1.0mm or 1.27mm. As an aerospace manufacturer, we do use SnPb HASL fairly regularly. We find that the (lack of) planarity is more of an issue on fine pitch (0.5mm, 0.4mm) leaded component lands than on 0.8mm pitch and above BGA lands.
While HASL has its downsides, uniformity being one, the alternatives all have some downside as well. So as long as the PWB supplier is capable of maintaining a good process, you can certainly have success with the finish.
All that said, the general trend seems to be away from HASL toward more planar finishes, though it may depend on who you ask. The "planar" finishes do enable more repeatable printing and placement for components with very fine pitch land geometries.
At some point your customer may need to consider both the technical and supply-chain factors and decide whether HASL makes sense for their future. It's one thing if they are choosing HASL because of specific technical requirements, and quite another if they are choosing it because"that's what we've always done."
Hope you execute DFMA and its your responsibility to notify this risk to customer and latent failure risk if customer want to go with HASL on cost perspective. HASL do not have coplanarity very good and is highly variable, preference to be given to ENIG in case BGA <30mil pitch.
Even nothing works to shift to ENIG, ask specification of coplanarity to be maintained with customer to deploy at PCB fabricator.
Supplier Quality Leader
One of the HASL finish characteristics is variable thickness. This can cause co-planarity issues for BGAs as well as for other fine pitch multi leaded components.
There are some solutions to alleviate this problem and the other comments covered them already. If it is a customer requirement, the tight tolerances at the PCB manufacturer would be your best bet.
Engineering and Operations Management
As mentioned above, if you have a 0.6 mm ball, then you are dealing with a rather large pitch BGA (above 0.8mm pitch). With such a 1.0 or 1.27 mm pitch BGA you could actually be OK with even Vertical SnPb HASL, which I have been using down to 0.8mm pitch BGA with good success. Of course, you need to make sure the "other parameters" are tweaked for good results, also as mentioned above (solder paste type, stencil thickness, aperture size, reflow profile, etc).Steven McLaughlin, ABB Switzerland Ltd., Switzerland
But the more obvious improvement to make is to ensure the PCB is done with Horizontal SnPb HASL. This process method results in much flatter pads. We have been doing down to 0.8 mm BGAs and 0.4 mm QFPs for many years with excellent yield results, in fact even better than with the previously used ENIG on some PCBAs (and at a much lower PCB cost).
There are many reasons to avoid the ENIG surface, but for Industrial applications it is extremely important to utilize surface finishes that have the much more robust SnCu Intermetallics (vs SnNi IMCs) and are the most long term corrosion immune (such as HASL - SnPb or Pb-free, or ImSn). ENIG fails badly in both respects (plus the cost, black pad, rework and many other issues/challenges).
And if you are needing a RoHS compliant finish, then the Pb-free HASL actually provides you with a further benefit for your fine pitch concerns, as it results in even flatter pads. The Vertical Pb-Free HASL process creates pads that are in the same range of Flatness as the Horizontal SnPb HASL Process, and the Horizontal Pb-free HASL process is even flatter.By the way, the Pb-free HASL processes are also resulting in slightly thicker minimum pad coating dimensions, which can accept/accommodate more IMC growth.
The Vertical Pb-free HASL process is more common presently, but there are some of the large high quality PCB suppliers that have already switched over some of their Horizontal machines to Pb-free HASL (Nihon Superior SN100C being the preferred choice). And they are seeing continued increase in demand for Pb-free HASL (as the old SnPB HASL, of course, decreases in demand and is actually being banned in certain areas of China, such as Dongguan, in the near future).
Either Pb-free HASL type is easily usable down to 0.4mm pitch QFP and 0.8mm BGA (and possibly less for the Horizontal type).Ironically, over the past couple years I have had 2 of the largest and highest quality PCB suppliers (globally) perform in-depth analysis of these various HASL surface thickness and flatness dimensional issues, so my comments above are based on actual extensive factual data collection.
The only thing I would avoid is the use of Pb-free HASL on high complexity/layer count PCBs with Blind-vias, especially stacked blind vias. The increased thermal shock can be an issue with such fragile blind via PCBs. The recommended choice here would be Immersion Tin, as it will provide great flatness (basically equivalent to ENIG), very good long term Industrial market Corrosion resistance, a strong SnCu IMC interface, low cost and it will avoid a high thermal shock in the process.
I would however recommend against using Immersion Tin on very Heavy Copper PCBs, as you will have full via fill challenges for any thru-hole joints (this is where the Pb-free HASL really saves the day). Immersion Tin is also a bit of a tricky process for PCB suppliers to keep under control (although so is ENIG), so make sure you use a high quality PCB shop.
You will also have much better and more consistent results if you have your boards leveled on a horizontal leveler. Acceptable results can be achieved in both lead free and tin lead horizontal leveling.Ray Rose, Unicote Halco, USA
Here is what I don't understand, I know that HASL is highly uneven but when the PCB undergoes reflow does not the SnPb flow over the pad surface evenly due to surface tension, etc. and coplanarity is maintained?Robert Hanson, Americom Seminars, USA
Is the issue actually that sections of the pad may not have any SnPb on them and therefore the Cu and SnPb will not intermingle?
The big issue with using HASL for this application is can your PCB house provide consistent coverage on your BGA pads. I have seen in past experience where an acceptable level of co-planarity can be achieved only to have the pads on the board oxidize and not be able to produce reliable solder joints. A quality board house, will be able to answer your question.Paul Dickerson, Matric, USA
I have seen the whole co-planarity argument made by industry experts. My main issue, ask them for a paper where boards have been produced to show the effects on end of the line quality. In other words, we built a given number of boards using a number of finishes, and we experienced the following defect rates. At the end of the day, we are all concerned with building quality boards without defects.
The price difference between HASL and ENIG is well documented. My argument has always been, will the cost of going to an alternative finish may erode margin verses defects caused by HASL.
In answering Robert's question, the HASL boards will fail to gasket the stencil to the board surface. This will allow solder paste to escape around the edges of the pads and can lead to solder balls, and in extreme examples solder bridges.
It is very very difficult to get a highly coplanar area for populating a 256-ball BGA component with 0.6mm pitch in a PCB fabricated through HASL process. This is by virtue of the HASL process itself. A slight offset in the planarity can eventually result in the formation of cracks in the device, because of the difference in the stiffness of the solder joint, the associated lead frame and the over molded epoxy. I would suggest to use an ENIG PCB for the application.Dr. KUTTIYIL THOMAS OOMMEN THARAKAN, VSSC
|Submit A Comment|
Free Newsletter Subscription
Circuitnet is built for professionals who bear the responsibility of looking ahead, imagining the future, and preparing for it.
Insert Your Email Address