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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


CSP/BGA Assembly Problems

Kelvin Chow

#13167

CSP/BGA Assembly Problems | 29 December, 1998

As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages.

Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste.

Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. Any Comment??

reply »

Michael Allen

#13168

Re: CSP/BGA Assembly Problems | 29 December, 1998

I don't have answers...just comments.

We're just starting to assemble CSPs on protos, and I've got the same concerns you have. The MicroBGA package I'm working with has a ball diameter of 0.325mm (0.0128") and a standoff of just 0.010" BEFORE soldering.

We'd like to continue to use aqueous solders (at least for the near term). Most people I talk to say that no-clean must be used for CSPs; however, their companies had previously switched to no-clean for other reasons (and they haven't ever tried washing CSPs), so I'm not convinced.

I recently did my own SIR test (Class II) and the initial results were positive, so we're proceeding cautiously with aqueous. We're still developing our stencil design and land pad design for this package. One goal we have is to maintain the 0.010" standoff AFTER soldering -- or possibly increase it -- by maximizing the solder volume. This is a challenge, obviously.

I'd like to hear from anyone else who's cleaning boards with CSPs (?).

| As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | Any Comment?? |

reply »

Kelvin Chow

#13169

Re: CSP Assembly- No-clean vs Aqueous | 29 December, 1998

It is still a big concern on using No-clean or aqueous cleaning for CSP assembly. I prefer to use no-clean, however, not all components are suitable to use it. Especially those dirty capacitors or connectors which may cause solderability problem. It may the reason someone prefer to use aqueous clean paste.

Besides, I want say that SIR test is used to determine the characterize fluxes under high humidity and heat conditions. If you want to know the cleaningness of the CSP assembly. You may need to know the "salt" content under CSP after assembly. If the salt content is very low, it mean the assembly is safe.

There is several standards cover this test. e.g.MIL-P-28809 OR there is a machine call Omegameter / Ionograph

Take care.

Kelvin Chow

| I don't have answers...just comments. | | We're just starting to assemble CSPs on protos, and I've got the same concerns you have. The MicroBGA package I'm working with has a ball diameter of 0.325mm (0.0128") and a standoff of just 0.010" BEFORE soldering. | | We'd like to continue to use aqueous solders (at least for the near term). Most people I talk to say that no-clean must be used for CSPs; however, their companies had previously switched to no-clean for other reasons (and they haven't ever tried washing CSPs), so I'm not convinced. | | I recently did my own SIR test (Class II) and the initial results were positive, so we're proceeding cautiously with aqueous. | We're still developing our stencil design and land pad design for this package. One goal we have is to maintain the 0.010" standoff AFTER soldering -- or possibly increase it -- by maximizing the solder volume. This is a challenge, obviously. | | I'd like to hear from anyone else who's cleaning boards with CSPs (?). | | | As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | | Any Comment?? | | | |

reply »

Michael Allen

#13170

Re: CSP Assembly- No-clean vs Aqueous | 30 December, 1998

I'm afraid we don't agree regarding the best test method to use here. My understanding is as follows.

The omegameter test measures board-level, ionic contamination (i.e., average for the entire board area and all components); it does not tell you whether the area under a specific component is contaminated.

The Surface Insulation Resistance (SIR) test does focus on the cleanliness beneath a specific component. It's conducted at elevated temperature and humidity (with a voltage bias) in order to accelerate electromigration and dendritic growths -- otherwise the test would take years. I believe that the results of SIR testing are applicable to end-uses in "normal" environments as well as severe environments.

| It is still a big concern on using No-clean or aqueous cleaning for CSP assembly. I prefer to use no-clean, however, not all components are suitable to use it. Especially those dirty capacitors or connectors which may cause solderability problem. It may the reason someone prefer to use aqueous clean paste. | | Besides, I want say that SIR test is used to determine the characterize fluxes under high humidity and heat conditions. If you want to know the cleaningness of the CSP assembly. You may need to know the "salt" content under CSP after assembly. If the salt content is very low, it mean the assembly is safe. | | There is several standards cover this test. e.g.MIL-P-28809 | OR there is a machine call Omegameter / Ionograph | | Take care. | | Kelvin Chow | | | I don't have answers...just comments. | | | | We're just starting to assemble CSPs on protos, and I've got the same concerns you have. The MicroBGA package I'm working with has a ball diameter of 0.325mm (0.0128") and a standoff of just 0.010" BEFORE soldering. | | | | We'd like to continue to use aqueous solders (at least for the near term). Most people I talk to say that no-clean must be used for CSPs; however, their companies had previously switched to no-clean for other reasons (and they haven't ever tried washing CSPs), so I'm not convinced. | | | | I recently did my own SIR test (Class II) and the initial results were positive, so we're proceeding cautiously with aqueous. | | We're still developing our stencil design and land pad design for this package. One goal we have is to maintain the 0.010" standoff AFTER soldering -- or possibly increase it -- by maximizing the solder volume. This is a challenge, obviously. | | | | I'd like to hear from anyone else who's cleaning boards with CSPs (?). | | | | | As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | | | | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | | | | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | | | Any Comment?? | | | | | | | | | |

reply »

Kelvin Chow

#13171

Re: CSP Assembly- SIR Test | 31 December, 1998

Dear Micheal,

SIR test is used to characterize fluxes by determining the degration of electrical insulation resistance of rigid printed wiring board specimens after exposure to the specified flux. (IPC TM-650 #2.6.3.3 http://www.ipc.org/html/tm2.6.htm). Do you follow the same spec? I suppose that this test method can only be applied on a specific assembly; not general PWA assembly. Please correct me if it is not true.

Besides, I agree that SIR test is excellent to determine the long term reliability. Yet, it take a long time to do a test (168 hrs??). Do you consider to use Omegameter as a daily QC test?

Take Care

Kelvin

| I'm afraid we don't agree regarding the best test method to use here. My understanding is as follows. | | The omegameter test measures board-level, ionic contamination (i.e., average for the entire board area and all components); it does not tell you whether the area under a specific component is contaminated. | | The Surface Insulation Resistance (SIR) test does focus on the cleanliness beneath a specific component. It's conducted at elevated temperature and humidity (with a voltage bias) in order to accelerate electromigration and dendritic growths -- otherwise the test would take years. I believe that the results of SIR testing are applicable to end-uses in "normal" environments as well as severe environments. | | | | It is still a big concern on using No-clean or aqueous cleaning for CSP assembly. I prefer to use no-clean, however, not all components are suitable to use it. Especially those dirty capacitors or connectors which may cause solderability problem. It may the reason someone prefer to use aqueous clean paste. | | | | Besides, I want say that SIR test is used to determine the characterize fluxes under high humidity and heat conditions. If you want to know the cleaningness of the CSP assembly. You may need to know the "salt" content under CSP after assembly. If the salt content is very low, it mean the assembly is safe. | | | | There is several standards cover this test. e.g.MIL-P-28809 | | OR there is a machine call Omegameter / Ionograph | | | | Take care. | | | | Kelvin Chow | | | | | I don't have answers...just comments. | | | | | | We're just starting to assemble CSPs on protos, and I've got the same concerns you have. The MicroBGA package I'm working with has a ball diameter of 0.325mm (0.0128") and a standoff of just 0.010" BEFORE soldering. | | | | | | We'd like to continue to use aqueous solders (at least for the near term). Most people I talk to say that no-clean must be used for CSPs; however, their companies had previously switched to no-clean for other reasons (and they haven't ever tried washing CSPs), so I'm not convinced. | | | | | | I recently did my own SIR test (Class II) and the initial results were positive, so we're proceeding cautiously with aqueous. | | | We're still developing our stencil design and land pad design for this package. One goal we have is to maintain the 0.010" standoff AFTER soldering -- or possibly increase it -- by maximizing the solder volume. This is a challenge, obviously. | | | | | | I'd like to hear from anyone else who's cleaning boards with CSPs (?). | | | | | | | As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | | | | | | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | | | | | | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | | | | Any Comment?? | | | | | | | | | | | | | | | | | |

reply »

Michael Allen

#13172

Re: CSP Assembly- SIR Test | 31 December, 1998

Kelvin,

Yes, we're referencing the same spec. I performed the test under the Class 2 conditions (except that I used a 100V bias rather than 50V...but my water-washed microBGAs still passed).

Regarding the omegameter test: we have this instrument, and we use it to test every prototype as part of our production process validation. As far as I know, the contamination readings are always well below the specified limits.

Good luck, and please pass on any further info on this subject!

| Dear Micheal, | | SIR test is used to characterize fluxes by determining the degration of electrical insulation resistance of rigid printed wiring board specimens after exposure to the specified flux. (IPC TM-650 #2.6.3.3 http://www.ipc.org/html/tm2.6.htm). Do you follow the same spec? I suppose that this test method can only be applied on a specific assembly; not general PWA assembly. Please correct me if it is not true. | | Besides, I agree that SIR test is excellent to determine the long term reliability. Yet, it take a long time to do a test (168 hrs??). Do you consider to use Omegameter as a daily QC test? | | Take Care | | Kelvin | | | I'm afraid we don't agree regarding the best test method to use here. My understanding is as follows. | | | | The omegameter test measures board-level, ionic contamination (i.e., average for the entire board area and all components); it does not tell you whether the area under a specific component is contaminated. | | | | The Surface Insulation Resistance (SIR) test does focus on the cleanliness beneath a specific component. It's conducted at elevated temperature and humidity (with a voltage bias) in order to accelerate electromigration and dendritic growths -- otherwise the test would take years. I believe that the results of SIR testing are applicable to end-uses in "normal" environments as well as severe environments. | | | | | | | It is still a big concern on using No-clean or aqueous cleaning for CSP assembly. I prefer to use no-clean, however, not all components are suitable to use it. Especially those dirty capacitors or connectors which may cause solderability problem. It may the reason someone prefer to use aqueous clean paste. | | | | | | Besides, I want say that SIR test is used to determine the characterize fluxes under high humidity and heat conditions. If you want to know the cleaningness of the CSP assembly. You may need to know the "salt" content under CSP after assembly. If the salt content is very low, it mean the assembly is safe. | | | | | | There is several standards cover this test. e.g.MIL-P-28809 | | | OR there is a machine call Omegameter / Ionograph | | | | | | Take care. | | | | | | Kelvin Chow | | | | | | | I don't have answers...just comments. | | | | | | | | We're just starting to assemble CSPs on protos, and I've got the same concerns you have. The MicroBGA package I'm working with has a ball diameter of 0.325mm (0.0128") and a standoff of just 0.010" BEFORE soldering. | | | | | | | | We'd like to continue to use aqueous solders (at least for the near term). Most people I talk to say that no-clean must be used for CSPs; however, their companies had previously switched to no-clean for other reasons (and they haven't ever tried washing CSPs), so I'm not convinced. | | | | | | | | I recently did my own SIR test (Class II) and the initial results were positive, so we're proceeding cautiously with aqueous. | | | | We're still developing our stencil design and land pad design for this package. One goal we have is to maintain the 0.010" standoff AFTER soldering -- or possibly increase it -- by maximizing the solder volume. This is a challenge, obviously. | | | | | | | | I'd like to hear from anyone else who's cleaning boards with CSPs (?). | | | | | | | | | As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | | | | | | | | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | | | | | | | | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | | | | | Any Comment?? | | | | | | | | | | | | | | | | | | | | | | | | | | | |

reply »

Kris Ewen

#13173

CSP/BGA Assembly Solutions | 4 January, 1999

Where I worked, we used 0.014" square aperatures (several mfgs. have proven the benefits of square & I've seen a better transfer efficiency myself). Some mfgs use smaller edge lengths & have had success. 0.005" thick stencils give better transfer efficiency for smaller aperatures.

Use a dedicated fixture for screen print!!!!!!! Without a fixture, paste volumes were all over (from 50-100 mils^3) [measured using 3D inspection equipment. WITH a fixture the paste volumes followed, in most cases, a tight normal distribution. The November 1998 issue of SMT (www.smtmag.com) has a good article on CSPs in general (High Speed Assembly with CSPs by Julian P. Partridge & Vern Solberg). I've seen some different volumes than what they show, but we did use different aperatures) & have also found that there is a big difference in the volumes you will get on a small board and a big board, even with a dedicated fixture.

K.

Hi PTT!

| As CSP and BGA packages become more and more popular, I want to know if anyone get problem to assembly CSP packages. | | Currently, most CSP use 20mil solder ball diameter and it is a trend to go down to 16mil or even 12mil. I wonder if there is a problem to apply solder paste. | | Besides, as the ball size become smaller and smaller, the gap between CSP to PCB become smallar. I bet all of us will use No-Clean solder paste for this CSP assembly. | Any Comment?? |

reply »

ICT Total SMT line Provider