Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.

# help help! Water clean for CSP/BGA package

karlin

help help! Water clean for CSP/BGA package | 5 October, 1999

Hi,

Help! Could anyone help to enlighten me on this?

Question:

If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI)

Brian

Re: help help! Water clean for CSP/BGA package | 5 October, 1999

| Hi, | | Help! Could anyone help to enlighten me on this? | | Question: | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | Thank you in advance! | I'm sorry, you ask an impossible question. There are a hundred unspecified parameters in your hypothesis, not the least of which is defining a "normal water clean washing machine". You do not even state whether it is batch or conveyorised.

In my 30-odd years' experience of aqueous cleaning, the biggest myth I have come across is that of water pressure being an important factor in cleaning quality. Sure, it plays a role, but a very minor one. To illustrate this, imagine a machine running at 100 bars (a VERY high pressure) fed to atomising nozzles producing a droplet size of 10 �m. This would produce a quasi-mist which would float in the air and never clean a PCB in a hundred years. The important parameter is the kinetic energy imparted to the water at the moment when it hits the board. After hundreds of experiments, I have determined that a medium pressure (3 - 6 bars) and a high volume of water impinging on the boards in coherent jets at about 25� to the board surface gives the best results. The kinetic energy required is provided by about 1,5 kW of pump consumption for every linear metre of spray bar. This will give maximum cleaning efficiency under tight components. However, there are many other parameters such as shadowing. The board must be designed to be cleaned, with adequate spacing between components of different heights. The wash cycle must last at least 2 - 3 minutes -- no problem with batch machines but a big one with conveyorised ones. The temperature must be just right, depending on the flux/paste chemistry, but typically about 55 - 60�C. The surface tension of the water should be less than about 30 dynes/cm. With W/S chemistry, the surfactants in the residues should maintain this, but you may need to add a little anionic surfactant to the wash water in a freshly filled machine.

Equally important is the rinsing. Here, you can halve the pump power, but the nozzles must provide a spray such that the droplet size at the moment of impingement is less than the smallest stand-off, again at an acute angle. To understand this, the under-component spaces are filled with contaminated water, held in place by capillary forces produced by the surface tension. Flooding around the component is not going to move this: it requires the rinse water to be able to "shoot" itself under the component to replace the dirty water droplet by droplet. To get really clean results under the components requires typically 10 rinses (6 minimum) therefore you need the appropriate number of spray bars in conveyorised machines, the last 2 or 3 with uncontaminated DI water, which has the highest surface tension. With batch machines, the rinse spray bars (separate from the wash jet bars) should turn, or preferably oscillate, a minimum of ten times in the rinse period.

Not the least part of the process is drying. You will appreciate that rinsing is a series of successive dilutions to what is hoped to be a safe level of residual contamination. If you simply allow the water to dry off the assemblies, what will happen? As it evaporates, what remains will start to congregate, by surface tension, to where the capillaries are smallest, i.e. round the solder joints, which is exactly where you don't want contaminants. It is essential to remove as much excess water as possible with high-velocity air knives or by centrifuging the assemblies, before allowing evaporation to start (this is also energy-saving).

Please be warned that, in all the subsectors of electronics assemblies, there is more nonsense talked about in cleaning than any other subject (soldering included!). Very few vendors really know much about it. After all, in pre-Montreal Protocol days, it was said that "slosh it around in a bucket of Freon and it will be clean", yet a real scientific study, because of environmental pressures, shows that CFC-113 azeotrope cleaning was one of the worst from the point of view of quality. Many of us were conned by this for years upon years. Water cleaning will give excellent results (up to ten times better than CFC-113 azeotropes), but only if you apply it correctly.

For further reading, details of my book "Cleaning and Contamination of Electronics Components and Assemblies" may be found at http://www.elchempub.com/epfiles/ep06.htm . (This is not really commercial, because it is the publisher's site, not mine.)

Although it is commercial (sorry!), the reference below also gives a host of technical information about aqueous cleaning.

Hope this helps,

Brian

Graham Naisbitt

Re: help help! Water clean for CSP/BGA package | 5 October, 1999

| Hi, | | Help! Could anyone help to enlighten me on this? | | Question: | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | Thank you in advance! | Karlin,

For modern Chip-Scale pakages, you MUST look to switching inot a saponifier in your wash zone. Pure OA using only water is NOT proving adequate. Refer to Precision Cleaning magazine Feb or Mar 99 issue entitled: "Cleaning under LCC devices, by someone from Lockheed Martin" I have a copy I can fax you if you get me your details.

If you are using pure OA, you MAY find it necessary to change your cleaner if it is equipped with only 1 sump.

As to Brians comments about batch v in-line - The school of hard knocks is the most effective I know. In-line is the better PROCESS. However, each has its own merits and problems which are far too complicated to discuss here.

What system are you using? Don't be shy - it will help get you a simple answer.

Regards, Graham

Re: help help! Water clean for CSP/BGA package | 5 October, 1999

| Hi, | | Help! Could anyone help to enlighten me on this? | | Question: | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | Thank you in advance! |

Re: help help! Water clean for CSP/BGA package | 5 October, 1999

| | Hi, | | | | Help! Could anyone help to enlighten me on this? | | | | Question: | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | Thank you in advance! | | | | Hello Karlin,

Every day I speak with many individuals who are experiencing difficulty in cleaning underneath fine pitch components, BGA's and CSP packages. It usually comes down to providing the customer with a result that meets the cleanliness requirment.

It is the right combination of flow and pressure that will provide satisfactory cleaning results. Most cleaning systems today recommend pressures of 80-100 psi, with flows rates between 60-80 gpm. The number of spray wands, nozzles, wash sections, etc. will vary depending on the cleaning company. the angle of pentration, flow, and pressures onto the PCB are critical when BGA, CSP and fine pitch components are present. With the use of a chemistry to lower the surface tension, and longer cycle times you willbe able to compensate for the decreases in pressure and loss of proper spray wand penatration that are found with older inline technologies and batch type cleaning systems.

In agreement with the other responses, cleaning underneath components involves many factors. I have several referances of manufacturing companies that are sucessfully cleaning underneath fine pitch,BGA's and CSP's. I would encourage you to contact these individuals to discuss what testing they have performed to qualify their cleaning process. Ultimately, performing various tests on different cleaning systems will provide you with concrete answers. Short of actual data, it is a speculation game.

I will be happy to further discuss cleaning processes and provide the contact information of other companies running similar processes. You may contact me at 602-315-9486.

Best Wishes,

Debbie

Karlin

More informations on water clean for CSP/BGA package | 6 October, 1999

| Hi, | | Help! Could anyone help to enlighten me on this? | | Question: | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | Thank you in advance!

Hi,

Thank you for the informative feedback specially to Debbic, Brain and Graham.

Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues).

Perhaps with the following informations, your could throw some light into the discussion:

Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test.

Any comment on whether could we establish some form of guideline? Thank you,

rgs karlin

Brian

Re: More informations on water clean for CSP/BGA package | 6 October, 1999

| | Hi, | | | | Help! Could anyone help to enlighten me on this? | | | | Question: | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | Thank you in advance! | | Hi, | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | Perhaps with the following informations, your could throw some light into the discussion: | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | Any comment on whether could we establish some form of guideline? | Thank you, | | rgs | karlin | Karlin

Thanks for this additional information. Unfortunately, I don't know the design of your machine, but I'll talk about conveyorised machines in general.

You say you have three wash and three rinse compartments. I presume that each one has one or two spray bars. I suggest that this may be insufficient, if this is the case. Say you have coherent jets (as you ought) in the wash zones. They will resolve into an area of high-energy impingement equivalent to not more than 5 cm long. At a conveyor speed of 1,2 m/min, this will give a cleaning time of 2,5 seconds per bar. Even with 6 spray bars, this is still very insufficient for difficult circuitry, even assuming the pump energy is sufficient (which it may not be). Worse, most conveyor machines spray at an angle close to 90� to the plane of the PCB. This means that about 85% of the kinetic energy imparted to the water is dissipated uselessly by forcing its way through the puddles of water, splashing back and excessive turbulence on impact. This energy is not being used. The more acute the angle of impingement, so the use of the kinetic energy to force the water under the components becomes greater, with losses reduced from about 85% to less than 30%.

My practical advice would therefore be a) slow down the conveyor to the lowest speed you can get away with b) make sure you have sufficient pump power (1,5 kW/metre of spray bar length) c) turn your spray bars so that the jets do not hit the assemblies at 90�, to as acute an angle as you can get away with d) make sure your wash jets are coherent up to the point of impingement e) make sure your rinse spray jets have a fine enough spray reaching the corners of where the components meet the substrate f) if necessary, add more jet and spray bars and pumps to extend the time that washing and rinsing is actually occurring.

The whole secret of this operation is that of energy at the right place for the right time.

Graham recommends a saponifier. With the right chemicals and equipment, this is not necessary and could be detrimental. I have many customers cleaning off water-soluble products without a saponifier, but the choice of the chemistry is critical. Many so-called water-soluble products (especially pastes) do not leave truly water-soluble residues but water-washable i.e. they form a suspension in the wash water rather than a solution: these require more energy to remove successfully than the truly water-soluble ones. Some even contain modified rosin with a longish hydrophilic chain attached to ensure it can be removed in water, but no imagination can call it soluble. Others are strongly micellar in nature. It is true that saponifiers may help to remove the residues, but the saponifiers themselves are even more aggressive than the fluxes, with pHs way up in the 11-12 region. They therefore require removal just as efficient as required to remove the flux itself, or more so, implying that they should be used only if you can add an extra ad hoc module to the front end of the cleaning machine. In addition, saponifiers cannot be used if you use polyimide substrates (Kapton), either rigid or flexible, as they attack this insulant fiercely. Caution also has to be exercised if you use saponifiers in the presence of amphoteric metals or alloys containing them (Al, Mg, Zn and possibly Fe). Any saponifier residues left under components could be detrimental to reliability.

I would therefore strongly recommend you do qualification trials with straight W/S products from as many sources as possible. The criterion of choice should be best cleaning with adequate soldering, rather than best soldering with adequate cleaning. This is a compromise that must be faced.

If you do not feel that you are sufficiently knowledgeable to implement this, there are consultants who can help you, but it is largely a question of systematic trials, especially if you are stuck with equipment which may not be optimal for the job in hand.

Best regards

Brian

Re: More informations on water clean for CSP/BGA package | 6 October, 1999

| | Hi, | | | | Help! Could anyone help to enlighten me on this? | | | | Question: | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | Thank you in advance! | | Hi, | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | Perhaps with the following informations, your could throw some light into the discussion: | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | Any comment on whether could we establish some form of guideline? | Thank you, | | rgs | karlin | Karlin: So, what is this white residue? Is it soluable in water? Ta. Dave F

Re: More informations on water clean for CSP/BGA package | 6 October, 1999

| | Hi, | | | | Help! Could anyone help to enlighten me on this? | | | | Question: | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | Thank you in advance! | | Hi, | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | Perhaps with the following informations, your could throw some light into the discussion: | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | Any comment on whether could we establish some form of guideline? | Thank you, | | rgs | karlin | Karlin,

Just a couple of quick suggestions to try. First raise your wash and rinse temperature to 135 -140 degrees F. I have seen white residues present when running lower temperatures with certain fluxes. I would also recommend using a higher quality DI water through out your systems. We see several contract manufactures using 10-18 Meghomm water.

As Brian referanced, problems can be encountered when using a saponifier. It is important to remove all residues with a good final rinse. The mention of time exposed to the spray also warrents merit. By reducing the conveyor speed, the product is exposed to more mechanical action, thus producing better results.

Please feel free to contact me directly for further information.

Debbie

Graham Naisbitt

Re: More informations on water clean for CSP/BGA package | 6 October, 1999

| | Hi, | | | | Help! Could anyone help to enlighten me on this? | | | | Question: | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | Thank you in advance! | | Hi, | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | Perhaps with the following informations, your could throw some light into the discussion: | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | Any comment on whether could we establish some form of guideline? | Thank you, | | rgs | karlin | Karlin,

I venture to suggest that you are using a flux which contains trace amounts of rosin. Which flux? I'll guess Alpha.

You might want to try a fully OA water SOLUBLE product. This may yield superior cleaning results.

Fot the white powder residues, contact Trace Labs, CSL or Robisan and get them to run an Ion Chromatography test. This shoudl ID the powder and help sort out the issue.

As I stated previously, I suspect that you will need to resort to a saponifier in the wash. This should not cause you too much inconvenience with the unit you have, although both the saponifier and flux MUST be water miscible to avoid entrapment problems.

Hope this helps, Graham

karlin

Re: More informations on water clean for CSP/BGA package | 6 October, 1999

| | | Hi, | | | | | | Help! Could anyone help to enlighten me on this? | | | | | | Question: | | | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | | | | Thank you in advance! | | | | Hi, | | | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | | | Perhaps with the following informations, your could throw some light into the discussion: | | | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | | | Any comment on whether could we establish some form of guideline? | | Thank you, | | | | rgs | | karlin | | | Karlin: So, what is this white residue? Is it soluable in water? Ta. Dave F

Hi Dave,

The white residue is soluable in water and we has confirmed that it is our flux residues from the EDX lab analysis.

karlin

Brian

Re: More informations on water clean for CSP/BGA package | 8 October, 1999

| | | | Hi, | | | | | | | | Help! Could anyone help to enlighten me on this? | | | | | | | | Question: | | | | | | | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | | | | | | | | | | Thank you in advance! | | | | | | Hi, | | | | | | Thank you for the informative feedback specially to Debbic, Brain and Graham. | | | | | | Guess there is no simple way out. My initial intention was to give a "rule of Thumbs" guideline to our designer/manufacturing what would be our washing capabilities when a package reach a certain size.( At this moment, we still require to perform water clean process as our boards are sensitive to flux residues). | | | | | | Perhaps with the following informations, your could throw some light into the discussion: | | | | | | Our washing machine is a inline cleaner Aquestrom 200 with 3 wash and 3 raising snd two dryer zone. The conveyor speed we fixed is at 4ft/min with wash pressure at 90 PSi and temperature set at 115�F. No saponifier is being used. We used ionic contamination tester ( Alpha metal)as our gauge for residue check ( we keep our standard as 5 micro gram / square inches instead of 14). However when we perform destructive by peeling off the CSP/BGA ( 48/386 leads) direct from boards, we see some white residue on these component joints even though it pass our ionic test. | | | | | | Any comment on whether could we establish some form of guideline? | | | Thank you, | | | | | | rgs | | | karlin | | | | | Karlin: So, what is this white residue? Is it soluable in water? Ta. Dave F | | Hi Dave, | | The white residue is soluable in water and we has confirmed that it is our flux residues from the EDX lab analysis. | | karlin | | If the residues are water soluble, this is all the proof you need to say that the machine design is such that a) the wash is inadequate and b) the rinse is probably doing little or nothing, under your conditions of use. You will seriously need to alter these (or change the machine for one designed for what you are doing). If the machine was sold to you recently with the promise that it would do the job, go to your vendor.

Brian

Re: help help! Water clean for CSP/BGA package | 8 October, 1999

| Hi, | | Help! Could anyone help to enlighten me on this? | | Question: | | If I have a CSP/BGA package of size X by Y and the standoff gap between the component and PCB is Z, What is the maximum allowable Y/Z or X/Z that using a normal water clean washing machine that able to effective performing the cleaning on component's bumps after attachment on PCB? ( Assume that water pressure is 60 PSI) | | | Thank you in advance! | Karlin: Most BGA and uBGA fabs use standard ball sizes and materials, optimized for the particular package. They will punish you if you try to spec non-standard materials. Fix your cleaning process.

My2�

Dave F