Reply
Jun 14 2010
By: Greywolf333 First Son 4 posts
Offline

ps3 reflow cheap and scientific method

2 replies 20109 views Edited Jun 14, 2010

This is my attempt to compile current guides on the internet into something that will always give the best chance of success, using the same primitive tools as used in the guides (i.e. heatgun).  My aim is to adhere as closely to industry standards as possible for reflowing the solder on your ball grid arrays used in your ps3, while investing as few dollars as possible.  Despite what skeptics say, reflow can be an excellent and potentially long lasting solution to your yellow light of death.  The extent and type of solder failure you have can vary greatly, therefore it is definitely worth a try to reflow the solder.

 

While the people in the guides may be excellent at reflowing bga solder without any temperature gauges or sophisticated tools--it is difficult for you to reproduce their results exactly.  Did you know that an extra CENTIMETER of distance between the bga and your heatgun, can amount to a difference in 50 degrees celcius?  Did you know that changing the fan speed on your heat gun can amount to a 70 degree difference or more?  Therefore, unless you are using the exact same gun as the guy/girl in the guide, and are mimmicking the distances and movement exactly, your results will vary greatly.  You will either be too cold and not achieve significant reflow, or too hot and damage your chip.  I first attempted to reflow my BGA using gilksy's guide and, after failure (I achieved 10 minutes of ps3 operability), realized that I was not achieving significant reflow because the distance from my heat gun to the BGA was perhaps slightly too far, or my fan speed was insufficient.  After doing some research, I discovered that the acceptable temperature window (that is, the temperature needed to reflow the solder, but not damage the BGA package, is potentially quite small.  You will need to achieve a sustained temperature of at least 220 degrees celcius to reflow the solder, and some existing packages using lead-free solder are only qualified to sustain 250 degrees celcius or less.  Therefore it is not in your best interests to approximate the distance and temperature needed for your heat gun.

 

 

Materials needed:

 

Heatgun:  Must have variable temperature control.  You may need to set it as high as 700 degrees F or more (although you will not be heating the board to more than 450 degrees, the set up described in this compilation will put some amount of distance between your gun and the BGA.  I used one with two fan speeds.

 

Aluminum Foil:  You will be fashioning a makeshift nozzle for your heatgun.

 

Digital thermometer: It must go up to 230 degrees C and have a long metal probe.  Many cooking thermometers will do. I got mine at Target, 15 dollars.  Alternately, you can use an infrared thermometer--but if you have the dollars for that, I recommend investing that in getting a new system instead.

 

Liquid no clean (non-corrosive) flux.  Please buy lead-free as you will be generating fumes.  I have not researched whether flux with lead is appropriate, better, or worse, for soldering with non-lead solder balls.

 

A metal surface that you can place the motherboard upon.  I used an old metal computer case, and the prongs on the ps3 motherboard hung off the edge.  You will be heating the board for several minutes, therefore using anything flammable is not a good idea.  In the method I describe, something that conducts heat well (i.e. a thin sheet of metal) will also help in heat distribution.  The goal is to heat your board slowly and spread heat to the rest of the board, to help deal with the thermal stress that arises from applying a hot air gun to a specific portion of the board.  A multi-layer bed of aluminum foil might also be appropriate, or even better.  I cannot say with absolute certainty if it it is appropriate or not--however, the fact that hundreds of people have gotten away with blasting their CPU from room temperature to 700 degrees in 20 seconds leads me to believe that a slow heating like I will describe below will be sufficient in preventing significant thermal stress damage.

 

Total cost of materials:  heat gun, flux, foil, thermometer.   40 + 9 + 2 + 15 = 66 US dollars.

 

Step 1:  Flux underneath the two main BGA chips.  I did not reflow the other 4 packages directly south of the chips as did Gilksy.  Why?  I don't understand what they do, and saw no evidence outside the guides that reflowing the solder on those components would be beneficial.  In the more professional reball guides and videos, those components are never resoldered.  I simply did not/could not find evidence to support reflowing those components or what they do.  I could easily be wrong on this however.

 

Step 2:  We will be fashioning an extension for your heat gun out of aluminum foil.  The logic behind this is that funneling the air over some distance will create a roughly uniform stream of air, in terms of both temperature and velocity.  The dimensions should be a square, just like the BGA, and the opening at the end should be slightly larger than the package itself.  In short the opening should be able to fit directly over the package, with perhaps a half centimeter of room left over on each side.  The entire funnel should extend the heat gun from the tip by perhaps 9 inches.  The longer the funnel, the more uniform (and therefore better) the air will be as it exits.  The concept is that you want the air near the sides of the opening of the tunnel to be closer to the same temperature and velocity as the air exiting near the center.  The downside is, the longer the funnel, the more powerful your heat gun will have to be in order to reach 225+ celcius.

 

I found that a shorter end of a credit card is roughly the length of one side the square funnel will have to be.  The easiest way to make the funnel is to make a template out of cardboard, then wrap with aluminum foil, maybe 5 or so turns, then pull the cardboard out.  Attach to funnel to the heat guard of your heat gun using tape.  I actually taped the end of the aluminum foil to fasten it into shape on the outside edge, thinking it would melt for sure.  It didn't in trial runs.

 

Step 3:  Firmly afix the heat gun to a table or work bench using a clamp or some other method so it is pointing straight down, but you are still able to access the controls without moving it.  Make sure it'll stay there--if it falls off and lands on your chip during high temperature heating--game over--you'll need a reball.  The funnel should be also therefore pointing straight down, and not at any angle.  Place the ps3 board with the first BGA directly underneath the aluminum nozzle.  it should be centered as best as possible, so that the 4 corners of your nozzle are equidistant to the 4 corners of the top of the chip.  The end of the nozzle should be hovering a very small distance above the chip.  I recommend 1 cm or less.  Again, the 4 corners of the end of the funnel should be uniformly above the chip in height.  A difference in height between corners can be detrimental... It could mean you're heating 1 corner at 200 C and another at 250 C.  Greater distances than 1 cm may cause less uniform heat distribution (I cannot verify this).  The main reason for such a short distance is you will find it very difficult, with the funnel, and then some significant distance after the nozzle, to achieve 225+ celcius even on high settings.

 

Note:  Before you continue, do a dry run with just a thermometer to get an idea of how to finely control your heat gun on the fly.  Time yourself going from 100 degrees C to 225 degrees C.   You must pay special attention, especially when using a cooking thermometer, to the fact that the digital readout is not instantaneous--it ramps up slowly until it reaches the correct temperature, rather than provide you with it immediately.  Practice getting from 100 degrees to 217 degrees but not accidentally going significantly over 230.  The trick is to get a rough estimate what the temperature might finalize as, based on the rate its increasing on the thermometer.  If you find yourself at 225 degrees and are still increasing at 1 degree every 2 seconds, then you would want to decrease the setting on your heat gun because the actual temperature may be 240 or higher.

 

Step 4:  Set the tip of the probe of the thermometer on the exact center of the chip you are about to heat and turn it on.  Begin applying heat on a low setting.  Adjust settings until you reach 100 C over a 1 minute or more period.  You can take your time at this phase.  Once you reach 100-120 C, hold steady for 5 to 10 minutes to distribute some of the heat to the rest of the board.

 

Step 5:  Slowly ramp up the heat.  Turning the blower from low to high got me to 150, then I incrementally turned the power up a few notches every 15 seconds and observed the thermometer.  Once I reached 200, if the thermometer looked like it was still going up significantly I stopped or even reduced heat.  You want to be between 225 and 235 degrees for about 90 seconds, and then turn it off.  You don't want to be above 217 degrees for longer than 3 minutes.

 

Here are example times from a source I found:

Total time above 217:  160-170 seconds

Time strictly above 221 and below 245:  60-70 seconds.

 

Step 6:  Wait 20 minutes.  Optionally (its been shown in industry that more rapidly cooling your board is beneficial and closer to the solder process during manufacturing.) point a fan at the board.  Low setting.  Be extremely careful not to touch the board.  The chip, directly after being reflow, is being held in place by a thousand or so tiny balls of solder, each a fraction of a milimeter in diameter.  So any jarring or heavy vibrations at this point could be detrimental. 

 

Step 7:  Reflow the other BGA if necessary or you are unsure which BGA had a solder failure.

 

 

Limitations of study: 

Did not do significant air flow dynamics study. 

Did not research (could not find) what temperature the BGAs on the ps3 are qualified to safely handle.  Welcomehomey says do not exceed 230 but that may be just a preventative measure.  I suspect 250 or so based on other things I've read, but 230-235 is definitely safer

 

 

 

 

 

Sources:

 

http://cache.freescale.com/files/dsp/doc/eng_bulletin/EB635.pdf

http://www.zeph.com/pap1.html

wikipedia solder

 

Welcomehomey's guide on this forum, which I followed and found to be accurate.

 

 

/Quote Welcomehomey

Flux used: Lead-Free noncorrosive, halide free, no clean Flux Dispensing Pen
Tin foil the area on the top of the motherboard so only the chip gets the heat from the heat gun.


Lead free (Sn/Ag/Cu (SAC/NEMI) alloy solder paste), has a melting point (liquidus) of 217 °C.


· Holding the board, put the flux on the board and beside the chip and squeeze the flux liquid under the chip.
So you put the liquid on one side of the chip and it flows to the other side of the chip by traveling under the chip.
You don't flux the capacitors on the opposite side of the motherboard where the bottom of the chip is.

The idea is to get flux around the solder, so add flux from the opposite side of the chip too if you want to be sure the flux is around every solder.


· Put the motherboard on the Zephyrtronics Adjustable Board Cradle. Holds the board flat.

On the top of the motherboard, put tin foil around the chip on all four sides so that only the chip is exposed to the heat from the heat gun and not the other components.


· Put the Zephyrtronics ZT-1-CLS COMPACT, ECONO AIRBATH PREHEATER below the motherboard where the bga is.
Preheat the bottom side of the board to 100°C before heating the BGA on the top of the motherboard.
This is what a controlled process means.
Temperature industry standard, built-in temperature ramping at 2° to 4°C per second to 100ºC


· On the top of the motherboard heat the chip up to around 217°C using the heat gun.
Use a circular motion on the heat gun moving it in a circle around the area of the chip.
· Peak temperature of the chip should be 225°C
· Remain above liquidus (217°C) and preferably below 225°C for 70 to 90 seconds.
· Larger components or heatsinks will necessitate longer heat cycles  


· Keep the work area: chip, motherboard, cooler than 230°C at all times.


· A preferred hot air tool supplies hot air to both the top and underside of the fixture.


Air flow all around the package ensures even heating of the package.
Packages not uniformly heated may develop a temperature gradient within the package.
High temperature gradients lead to high stresses that could damage the package.


This means the heat gun puts heat on the top and the preheater hets the bottom during the reflow process.
Start counting to 70 or 90 seconds when you heat the top of the motherboard on the chip and the chip reaches 217°C.


· after waiting the 70 or 90 seconds, turn off the heatgun and preheater and let the motherboard sit and cool for 20 minutes or so.
So no move the motherboard as things will move because the solder is in a loose form now.


Cost


ZT-4-MIL-120 AIRPLUS FUME EXTRACTOR
$175.00


Zephyrtronics Adjustable Board Cradle
$195.00


ZT-1-CLS COMPACT, ECONO AIRBATH PREHEATER
$325.00


infrared handheld thermometer OSXL450
$70.00


tin foil
$2.00


CircuitWorks Lead-Free Flux Dispensing Pen
$7.15


Total cost $774.15

/Quote Welcomehomey

Message 1 of 3 (20,109 Views)
Reply
0 Likes
First Son
Registered: 06/14/2010
Offline
4 posts
 

Re: ps3 reflow cheap and scientific method

Jun 15, 2010

I forgot to mention the issue of how to apply flux properly beforehand.  It was really impossible for me to tell if the flux I applied was significantly covering the area needed to be reflowed.  My solution was to apply liberally to all 4 sides, tilting the board each time to allow the flux to run underneath.  Too much?  Can you overflux?  No idea.

 

 

Message 2 of 3 (20,109 Views)
Reply
0 Likes
First Son
Registered: 06/14/2010
Offline
4 posts
 

Re: ps3 reflow cheap and scientific method

Aug 15, 2010

Update:  Over 2 months after reflow and still going strong, despite heavy daily usage in a hot apartment.

 

I'm fairly sure, from what I've read and experienced, that reflows have gotten a bad reputation as being inconsistent and short-lived mainly due to the fact that there are so many methods out there (most of which are simply uncontrolled 30 second blastings with a heat-gut being swirled around by hand from an approximate distance).  I've already mentioned the following but I want to reemphasize its importance:  If you're going to try a reflow fix, treat it like you would a lab experiment.  Control all the variables and do practice runs with a thermometer, and try to adhere as closely to industry standards as possible for resetting bga solder.  Pay special attention and test to make sure your method is going to heat the edges and directly around the package to the same temperature as the center.  If you use some sort of self-fashioned aluminum hood like I did, make sure it's sufficient to prevent a large temperature gradient from the center to the edges.  I think you will have to have the hood placed extremely close to the bga to achieve this.  Definitely do some dry runs if you use this setup.  If you have a huge gradient from center to edges, and the solder is cracked at the edges (impossible to tell where it is), then it might be impossible to reflow the solder near the edge of the package, without overheating and risking damage to the center of the chip.

 

Also, from what little I've read, the solder problem usually occurs beneath the GPU and not so much with the CPU.  (And it makes sense that the graphics chip for a console is put under much more stress than its CPU for a console)  There's no point to heating up other components if they are fully functional, so maybe try just the GPU first (I believe it says RSX or some such on the top of it).

 

Cheers,

Greywolf

 

 

Message 3 of 3 (20,109 Views)
Reply
0 Likes