reballing problems

[scooby01]

Ok I have been trying to re all gpus and Southbridge chips. I have jigs and stencils. I prefer the jig but as soon as i try to melt the balls i always get a few that join together. I have a jovy and heat guns, but both do the same. The flux i have tried is amtek and kimbo, both with the same results. Also tried different thicknesses and temp. Stencil have more success. Any tips would be appreciated.

[JayArr_BC]

Re: reballing problems

Hi Scooby

You haven't really given us much to go on. Can you give me a little more detail about the stencil and the jigs?

I use a tool that looks like this

http://www.ebay.com/itm/Blue-BGA-Reb...item27c7d3cd96

and I have pretty good results. Would you call this a stencil or a jig?

Can you also tell us if you are using leaded or leadfree balls or are you using paste. What temperatures you are using and the exact procedure you are doing.
Which Jovy - the model and what profile are you using.

There are a few different ways to reball an IC and some are easier than others.

Jeff

[scooby01]

Re: reballing problems

Hi Jeff that is what I call a jig. I use leaded balls for you. We have a achi station using default profile 4, which reaches 226 c with bottom heat at 180. The profile starts @90 dwell 2mins then 180c dwell 2mins then at about 200 the balls will start to move. Not all but maybe 3 or 4 areas. I use kimbo flux and have tried armtek with the same results. Even tried liquid flux very thin. Same result. There is nothing worse then getting a set of balls stuck. To recap I have expensive flux at different thicknesses.

[JayArr_BC]

Re: reballing problems

Hi Scooby

So you are putting flux on the IC,

then putting it in the jig with the proper template

then putting the balls onto the IC through the template

then removing the template

then removing the IC from the jig

then puting the IC into the Achi to melt the balls onto the IC,

is this your procedure?

JayArr

[scooby01]

Re: reballing problems

Yes exactly.

[momaka]

Re: reballing problems

If the chip isn't too big and it doesn't have a heat spreader, you can use hot air to melt the balls.

The way I normally do it is like this:

So you are putting flux on the IC,
then putting it in the jig with the proper template
then putting the balls onto the IC through the template
then removing the template

After removing the template, I use the hot air to *slowly* heat the chip until the solder balls become shiny and almost start to melt. Then I spray them with liquid flux (from about 10 inches or 20 cm above so as not to move them around from the spraying). After this, I continue heating the chip. When the balls become shiny again and are almost melted, I pick a corner of the chip and start heating there only. After usually 10-20 seconds (depending on chip size) the balls will start to melt. When I see them start dropping, I pick (randomly) a direction to go around the chip and continue to slowly move the hot air nozzle along the outer edge and basically making a "lap" around the chip. When I do this, I move the nozzle slowly so as not to leave any unmelted solder balls.
Once I complete the lap around the outer edge of the chip, most of the solder balls would be melted at that point. Finally I move to the middle and heat, just to make sure all of the solder balls in there have melted as well.

What I have found through experience is that the amount of flux you put on the chip before putting the balls matters quite a bit. Put too much and the balls will slide and clump together like you mentioned. Put too little and they won't stick to where they are supposed to go, so eventually they will move and stick to each other. The type of flux is also really important. For this particular step, I use paste flux only.

Also, if the balls are not aligning for some reason or another, keep spraying them with liquid flux and keep heating with the hot air. If the balls are partially melted and already stuck to the chip and they are not aligning, let the chip cool down a bit, then brush some paste flux (either Kingbo or Amtech) on all of the balls. Then reheat the chip again until they melt and align. This final touch should give you perfect results and very shiny joints.

[scooby01]

Re: reballing problems

Thanks for a very good method. We have heat hot air guns. What temperature and air pressure would you suggest and nozzell?

[JayArr_BC]

Re: reballing problems

Hi Scooby

I agree with Momaka, use the hot air, we use Aoyue 968 stations and I set the hot air to about 350 or 400 depending on the size of the balls. I turn the volume of air down really low and I don't put any restrictor nozzles on the gun.

I'll add that I also apply a little hot air before I remove the stencil/grid. With the balls all held in place I heat everything up just enough for the flux on the IC to melt, the balls then sink into the flux and then I let it cool. This seems to 'set' the balls in place and they don't roll off their pad as readily.

You should also note, if you haven't already that the top plate has four set screws that allow the height of the stencil/grid to be set and this changes from IC to IC so make sure it's set so that the balls are held accross their mid diameter by the stencil and not by their top or bottom. If it's too low it may hit the IC itself and if it's too high the ball may not be centered on the pad.

One other thing to watch for as you make your 'lap' as Momaka describes is that the area around the jaws that hold the IC are harder to get the balls to melt to the IC pad since the jig acts as a heat sink. If I can't get the outer rows to melt near the jaws I'll pass over them and after the rest of the IC is done I'll carefully lift it out of the jaws and set it on the bench and go back and reflow the four spaces that were too hard the first time. Until I realized the heat sink problem I would spend too much time around these areas and would overheat the balls without heating the IC and that's where I got the majority of my 'twinning' and mistakes.

I also have an Achi IR station but have found that the hot air, when you get the finesse of it, works better and faster than running a profile on the station.

Jeff

[scooby01]

Re: reballing problems

I am correct in understand that heat the jig with the stencil in place and heat the jig to allow flux to melt and set? Also leave gpu in jig for heating fully after removing stencil.
What profile settings do use for refit or reflow. We have been using default profile 4 with some success on reflow. Max out at 227c on leaded.

[JayArr_BC]

Re: reballing problems

Hi Scooby

I am correct in understand that heat the jig with the stencil in place and heat the jig to allow flux to melt and set? Also leave gpu in jig for heating fully after removing stencil.

yes, that's what I do.

My machine is not a Jovy so I don't know what a default profile 4 is. We took an Achi IR900-3 and upgraded the bottom elemets with four 650W Salamanders and then we replaceded the controllers with Omega RS422 CN7533s and hooked them up to a laptop. We also changed the thermostats so that they both touch the pcb (originally the bottom sensor was atttached to the IR plate.) We find it's important not to allow the temp difference between the top and bottom of the pcb to get too large, 10-20 degrees max. otherwise we see delamination and bubbles.

We can control the heating in 1 minute steps starting from ambient and going quickly up to about 175 in about 3 or four minutes, we pause there for 90-120 seconds to allow the flux to melt, activate and clean everything then we raise the temp to the solder melting point. If we are using leaded solder balls we only go up around 195. For lead free we'll go to 235 but we don't like lead free and will sometimes buy a new IC and clean off the lead free balls and reball it with lead balls before installing on the board.

Jeff

[scooby01]

Re: reballing problems

Thank you Jeff for the info on different heat from top to bottom. Getting confused with leaded and non leaded. We have a lot of success with reflow. But re balling has always been a pain in the neck until now. Tried heatgun method today with some success. AHCI are similar to the jovy. We haven't got it connected to a laptop at the mo for profiles but its all ok to use using control panel on front. We cannot rely on the thermo coupling so I have 2 other manual ones each side of the chip. What temp and time would U use for the gpu yo be resolderd back? Svooby

[ecking767]

Re: reballing problems

Ok I have been trying to re all gpus and Southbridge chips. I have jigs and stencils. I prefer the jig but as soon as i try to melt the balls i always get a few that join together. I have a jovy and heat guns, but both do the same. The flux i have tried is amtek and kimbo, both with the same results. Also tried different thicknesses and temp. Stencil have more success. Any tips would be appreciated.

Hey scooby01,

I feel you, I was having the same problem when it came time to reballing GPUs. I first brought the blue jig and some universal stencils, and all of the soldering balls.

The problem I was having was the balls will move or get stuck together while placing the ball into the universal stencil. So I purchase another jig and the universal stencil that work with this jig. See picture.

Its simple but I have more control of how the ball fall into place and how the gpu is place underneath the universal stencil. And I can apply the heat directly to the stencil with the gpu in place.

I think this jig work best and you have more control when positioning the gpu and there is a good flat contract between the stencil and gpu

hope this help
ecking767

[scooby01]

Re: reballing problems

Thanks to all! I am using heat gun. Hot air with much more success, only failed 1of 5. Just takes a lot of patients. Anyone have any manaul profiles for AChI not computerised for refit you and reflow inn leaded. I have fried many a mobo and gpu but have 2 other thermo couplings helping and saving burn outs.

[sasser]

Re: reballing problems

like ecking767 use direct heated stencil which i think the good method when we go heat directly to BGA. i hav direct heat stencil and also hav non direct stencil. with non direct heat stencil much more control monitoring. sometimes your ball stick together sometimes your ball stick inner hole of stencil. so my conclusion use heated direct stencil like ecking767. more exercise then go for another tools if not suit with your work

[savo.lekic]

Re: reballing problems

Hello evreryone!

Is it possible to directly heat stencyls made by Jovy Systems? I have black jig with set of universal stencyls.

Also have problems with balls joining together, becouse of universal stencyl.

[xboxhaxorz]

Re: reballing problems

if you have direct heat stencils then yes, if they are 90mm NO

[savo.lekic]

Re: reballing problems

if you have direct heat stencils then yes, if they are 90mm NO

These are 90 mm. Thank you!

[Ltank]

Re: reballing problems

I am very interested in this and am makining a DIY preheater and will be ordering a Hakko FR-803B from China for a very good price. Darn stands are high priced!
I worked industrial electronic 30+ years. From my experiance reheating solder on wires/post/terminals etc did not work well. I generally removed all the old solder, redress wire ends and apply new solder. What concerns me with the process is reheating the solder. Here you remove the old solder, apply flux, place new solder balls, heat them untill the TAL (temp above liquidus). Then it cools, then you put the balled chip in PCB ball socket and Tal the balls a second time for reflow. I guess there is no reasonable way to place the ball in the PCB and place IC inplace without dislocating some balls.
No coice I guess.
The problem I see with solder balls flowing and attracting each other is the solder ball size is too large for the IC that your reballing.
Please correct me or enlighten meto this topic that I am learning.

[rtstorm]

Re: reballing problems

Hi, please i dont understand what did you mean by direct heat stencil and non direct heat stencil??

[rtstorm]

Re: reballing problems

did you use some epoxy after reballing?