Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


BGA ball and PCB pad

Bryan

#27226

BGA ball and PCB pad | 12 February, 2004

Hi all, These days our customer asks us to figure out ,during reflow,the BGA ball and solder paste,which will first melt? I think it's very hard to get the rusult.coz I think it'll vary at diffrent part of the BGA,and how can this affect the performence of solder joint? Please advise!

Thanks&Regards!

Bryan

reply »

Kris

#27229

BGA ball and PCB pad | 12 February, 2004

what is the composition of the paste and ball ?

Interesting to know why the customer is asking you to look for it

Thx

reply »

Evtimov

#27251

BGA ball and PCB pad | 13 February, 2004

Hi Bryan

The answer is in the paste compozition. Balls of the BGA are usually made by eutetic solder paste.

reply »

#27252

BGA ball and PCB pad | 13 February, 2004

Is your customer trying to reduce voiding? If the device is using Sn90 / Pb10 and the paste is Sn63/Pb37 the board will melt first. However, if both solders are 183C eutectic, it's a tossup? PCB design, copper weight, bga type (pbga, micro bga, Tbga...etc). The key here is the entropy of the PCB vs the part.

What is the customer trying to achieve?

reply »

John W

#27289

BGA ball and PCB pad | 17 February, 2004

There's normally a number of reasons people want to know what melts 1st, 1 - to see if you know, 2 voiding issues - whole can of worms 3 device alignment. One of our collegues has said that most balls are Eutectic, that's in fact not totally correct. Not all chip manufactureres actually package the devcie them selves, they do the silicon and sub the packaging. Depending on the package type the ball contact will change, there is actually a fairly dramitc split between std Eutectic 63/37 tin / lead and also a 2% silver ball (62/36/2), this is usually found on TBGA devices which are highly prone to voiding issues due to the kapton base as much as anything else. if your usign a cermaic device then 9 times out of 10 it'll be 90/10 and wont 'melt', there are soem exotic ones out there like tin/lead/bismuth which starts to melt at 170 and fully at 180 again voiding a plenty. Solder paste will have a large impact as will the volume and shape of the print. if you print square appertures so over print onto the resist for example that will melt 1st irrelevent of the ball material. Strangely a nice complex subject but hey that's why people hire people like us right?

reply »

Steven

#27316

BGA ball and PCB pad | 18 February, 2004

OK, if the sphere of the ball and the paste on the pad are both eutectic, 63/37 then the paste will reflow first causing a drop of the component, after which the spheres will reflow causing a double drop and a good intermettalic. You can see this by getting your hands on a cd put out by Ersa or go on line to Ersa.de. The Ersascope has shown us this. If you have a high temp solder ball on the bga the solder on the pad again will reflow first. In fact I am told that the high temp solder ball really doesnt reflow. It is encompased ? by the eutectic solder.

reply »

Jay

#27336

BGA ball and PCB pad | 20 February, 2004

First, it depends on the alloy composition of the solder. As you might know, eutectic Sn/Pb solder has the lowest melting point among available conventional solder systems. (except those based on Sn-In and Sn-Bi systems.) For example, 63Sn/37Pb - 183C, 96.5Sn/3.5Ag - 221, 95.5Sn/3.8Ag/0.7Cu - 217 (this particular one is not exactly eutectic, though.), etc.

But, if you are dealing with the same solders for paste and BGA balls, it's always the paste which melts first. The reason is very simple - the dimension. Actually, paste has a lot of very small size of solder spheres surrounded by flux. Thus, flux removes the oxide layer at the solder surface easily and quickly. Besides, small dimension of spheres in the paste requires less energy for full melting. Just imagine soldering with solder wire and melting a big ingot of solder.

Regarding the effect, I can think about self alignment and voids. Both of them were well explained by John already. But voids are more related with flux type and amount - really complicated problems.

reply »


agp

#27340

BGA ball and PCB pad | 20 February, 2004

I agree with Jay. The paste will reflow first before the balls. The main reason is the flux content and the dimension of the spheres.

reply »

Bryan

#27404

BGA ball and PCB pad | 27 February, 2004

Very strange question,but interesting.Is there anyone who have conducted any experiments to figure out the results?I made a profile board,one thermal couple on pad and other one drilled into the ball of BGA.the 2 points are perpendicular to conveyor direction.From the thermal profile I found that the point on pad reaches 183C less than 1C earlier.even if the two points reach 183C simultaniously,I think solder paste will melt first for less sphere dimension(paste and ball are of the same alloy).but,if we choose 4 points?(2 sets).we found that BGA ball on the side earlier entering reflow oven melt earlier than the paste does on rear side .I mean on the full BGA board,along the conveyor direction,the ball first enters oven will melt earlier than paste later enters the oven.thus, for how this affects the solder joints performance ,I think it has little things to do with alignment.perhaps,to reduce the voids,it's better to let the paste melt first,but I don't know whether it works?

Thanks

reply »

PCB Soldering Tools

Panasonic OEM Solutions