Contact Us
  • Home
  • BLOG
  • Can You Mix Leaded Paste With Lead-Free BGA Balls?

Can You Mix Leaded Paste With Lead-Free BGA Balls?

Yes. A lead-free SAC305 ball soldered with Sn63Pb37 paste is a normal joint in defense and legacy work, and in the published accelerated thermal cycling studies it matches the full lead-free joint. It only works when the profile peak follows the ball rather than the paste: 235 C and above. At a leaded peak of 205 to 215 C the ball never melts, the joint looks closed under X-ray, and it is a mechanical contact sitting on a fillet.

The reason is one number. Sn63Pb37 melts at 183 C. SAC305 starts to melt at 217 C and is fully liquid at 220 C. A profile built for leaded paste peaks around 205 to 215 C, which sits under the solidus of the ball, so the ball never becomes liquid.

What melts, in what order

During ramp the paste is the first thing to go liquid, at 183 C. It wets the underside of the ball, the flux activates and the oxide on the ball breaks down, and then nothing else happens: the ball is still solid for another 34 C. If the profile turns over at 210 C, the paste solidifies again around an unmelted sphere and what is left is a collar of solder joined to the pad and touching the ball.

Take the peak to 235 C and the order changes. The ball melts, the paste and the ball become one liquid volume, and lead from the paste diffuses up through the ball. Published mixed alloy work shows the difference plainly: identical boards reflowed at 220 C show partial mixing with the lead concentrated near the pad, and the same boards at 235 C show lead distributed through the joint. The 235 C leg is also the one that survives accelerated thermal cycling, which is the opposite of what the lower temperature is supposed to buy.

Time above liquidus matters as much as the peak. Sixty seconds above 217 C gives the ball time to collapse and the lead time to move; a fast spike that touches 235 C for ten seconds collapses the ball but leaves the mixing half done. Our own reflow is a twelve zone SAC305 profile under nitrogen below 500 ppm oxygen, and a mixed build runs on the lead-free profile with the peak set by the ball alloy.

The pitch decides how much mixing is even possible. At 1.0 mm pitch there is enough paste per ball to make a real alloy change. At 0.5 mm pitch the stencil aperture holds a small fraction of the ball volume, so the joint is mostly the ball with a thin leaded layer at the pad. That is not a defect, but it does mean the joint is not uniform, and a thermal model that assumes one alloy for the whole joint is wrong.

Where it goes wrong quietly

Running the leaded profile because the paste is leaded. This is the classic head-in-pillow. The ball sits on a paste deposit, the paste melts and wets the bottom of the ball, the ball does not collapse, and the two surfaces touch without metallurgical bonding. X-ray shows a proper ball with a collar and no open, because the ball and the solder are the same density. The joint passes continuity at goods in and opens in the first few thermal cycles. The only reliable detection is dye and pry, or a cross section, on a sample that has been through the temperature excursion.

Treating an X-ray pass as a joint. Void limits do not catch this. Our X-ray reads voids to IPC-7095 with a 5 percent limit on thermal pads, and a head-in-pillow joint can show almost no void at all. Void data and collapse data are different questions, and the second one needs a section or a pull.

Promising a lead-free declaration. A mixed joint contains lead by construction. RoHS limits lead to 0.1 percent by weight in a homogeneous material, and a SnPb paste joint is roughly 37 percent lead at the pad end before any diffusion. Defense and aerospace work is usually exempt, and legacy industrial work often is not. If the file says lead-free, a mixed build cannot sign it, and swapping the paste to SAC305 is the only route back.

There is a fourth trap worth naming: baking. A moisture sensitive BGA that has been out of its bag goes through 125 C for the hours J-STD-033 gives for its level, and parts that have already been through one reflow take a shorter bake. Skipping the bake to protect the ball from heat gain is how popcorn voids appear in the same build that the profile work was meant to fix. The cycle arithmetic sits in how many reflow cycles a PCB survives.

What SUPERB Assembly does with a mixed build

The alloy pair is fixed at quote stage and written on the traveller: ball alloy, paste alloy, peak, time above liquidus, ramp. A mixed build is not run on the leaded profile, and a leaded build is not run on the lead-free one, because the peak belongs to the ball. Nitrogen stays below 500 ppm, the twelve zone profile is recorded per run, and the profile board goes through with the panel.

Inspection is set to the risk: 3D SPI before reflow for paste volume per aperture, 3D AOI and X-ray after, CT on the fine pitch parts where a 2D image cannot separate a collapsed ball from a resting one, and dye and pry on the first article of any mixed build or any BGA at 0.5 mm pitch and below. X-ray void reporting follows IPC-7095 with a 5 percent limit on thermal pads, and the report format is the one described in BGA void acceptance. The assembly side of the house runs under IPC-A-610 Class 2 and Class 3, and the same traveller follows the board into turnkey PCBA assembly and box build.

The honest boundary: a mixed joint is a process decision, not a default. If a customer sends a lead-free BOM with a leaded paste instruction and no profile, we quote the profile work as part of the build rather than running the paste on the leaded recipe. Where the peak cannot be raised because a part on the same side is rated to 220 C, the answer is to move that part to a second pass, not to lower the peak.

What to send with the BOM

Ball alloy for every BGA, paste alloy, the peak and time above liquidus on the report you qualified, moisture level and floor life for the packages, the finest pitch on the board, whether a lead-free declaration is required, the thermal cycling class, and whether underfill is called out. Eight items, and they decide the profile before the first panel rather than after the first field return.

Need a quote or a review of your files against these criteria? pcba@superb-tech.com · WhatsApp +86 133 9608 1443