X-Ray for CSP devices

Hello All,

I am setting up a line to mount CSP/uBGA devices and now am looking at the X-Ray process. It is my opinion that this should be done offline - can you tell me if Im right or do I need to put one inline?

Also anyone who is mounting CSP/uBGA devices, can you suggest good X-Ray equipment. What do you use? Does it do a good job?

Any advice appreciated.

Steve

reply »

| | Hello All, | | I am setting up a line to mount CSP/uBGA devices and now am looking at the X-Ray process. It is my opinion that this should be done offline - can you tell me if Im right or do I need to put one inline? | | Also anyone who is mounting CSP/uBGA devices, can you suggest good X-Ray equipment. What do you use? Does it do a good job? | | Any advice appreciated. | | Steve | Steve,

You don't need in line x-ray. Use the process to verify acceptable quality during prototyping. Then, use it on a random sample basis.

I'm using an 8000 pound HP x-ray that does a farily good job. Me and everyone else at HP, would rather use something like a Nicolet. The laminography is great but the machine is NT based but runs the graphics in DOS, if you can believe that.

The stuff being advertised on this forum might be worth a shot. I know nothing about them.

Best resolution and views still are obtained using very high resolution black and white systems. Laminography and speed are not their strong suits.

In all cases, I would x-ray to confirm first there is no shorting. Then, I would look at insufficient solder, voids, etc. Then, you can correlate luck with skill (bounce findings back to profiles, paste, design rules, other).

Earl Moon

reply »

| | Hello All, | | I am setting up a line to mount CSP/uBGA devices and now am looking at the X-Ray process. It is my opinion that this should be done offline - can you tell me if Im right or do I need to put one inline? | | Also anyone who is mounting CSP/uBGA devices, can you suggest good X-Ray equipment. What do you use? Does it do a good job? | | Any advice appreciated. | | Steve |

Steve,

There are many systems available in the market like: HP. Nicolet, Faxitron, Glenbrook etc. , But there are some key characterstics required for a better X-ray m/c. 1. Focal Spot: M/c are available in the market from a focal spot of 250 microns to 8 microns. More u go lower on the focal spot , More better the image is (Less blurr on the sides) and more u would be able to see the voids clearly. (ofcourse it also depends on the X-ray KV), But more costlier the system gets.

2. Anode votage.

You can see the shorts, misregistration, missing balls etc. with teh lower cost systems also, but for seeing the voids clearly u have to have a bettr voltage and focal spot.

3. Voltage blooming: The effect of voltage on the size of the voids. System should be such that increasing the voltage doesnot increase the relatve size of the void in the image.

Hope this would help in your selection.

Upinder ======

reply »

Hi Steve,

What Earl says is true. You need a stand alone failure analysis/process verification system before you go inline. No matter what brand you choose, it is important to compare a few features, across the board: - X-ray source kV, mA, and focal spot size - Imaging device field of view - Board handling device, i.e. 5 axis maneuverability - Imaging software capabilities, measurement etc. - After sales service - Price, price, price.

Vikram Butani V.J. Technologies | | | | Hello All, | | | | I am setting up a line to mount CSP/uBGA devices and now am looking at the X-Ray process. It is my opinion that this should be done offline - can you tell me if Im right or do I need to put one inline? | | | | Also anyone who is mounting CSP/uBGA devices, can you suggest good X-Ray equipment. What do you use? Does it do a good job? | | | | Any advice appreciated. | | | | Steve | | | Steve, | | You don't need in line x-ray. Use the process to verify acceptable quality during prototyping. Then, use it on a random sample basis. | | I'm using an 8000 pound HP x-ray that does a farily good job. Me and everyone else at HP, would rather use something like a Nicolet. The laminography is great but the machine is NT based but runs the graphics in DOS, if you can believe that. | | The stuff being advertised on this forum might be worth a shot. I know nothing about them. | | Best resolution and views still are obtained using very high resolution black and white systems. Laminography and speed are not their strong suits. | | In all cases, I would x-ray to confirm first there is no shorting. Then, I would look at insufficient solder, voids, etc. Then, you can correlate luck with skill (bounce findings back to profiles, paste, design rules, other). | | Earl Moon |

reply »