Stencil Printing Problems

[jnnewton]

I'm having issues with making a pcb that has a 100 pin tqfp part on it. I'm using the manufacturers recommended pad / mask layout for the part. The problem is inconsistency. I, very often, get boards that have both too little solder on some pins and bridging on others. What i think is happening is that when the solder pre-heats and sort of "spreads" out, it touches the paste from adjacent pad. Then during the acutal reflow portion, one of 3 things happens:
1. The paste gets wicked all to one pin, leaving the other with none
2. The paste bridges the two pins
3. It reflows properly.

So, I though the answer would be a thinner stencil, but that created two other issues:
1. The stencil printer drags the paste up out of the apertures if the pressure is too high
2. The stencil printer leaves paste attached to the paste in the apertures on top of the stencil, which causes no-release.
And if the pressure is "just right" I get a combination of the two scenarios.

Can anyone tell me what I'm doing wrong?

 

[Tony Stewart]

I'm having issues with making a pcb that has a 100 pin tqfp part on it. I'm using the manufacturers recommended pad / mask layout for the part. The problem is inconsistency. I, very often, get boards that have both too little solder on some pins and bridging on others. What i think is happening is that when the solder pre-heats and sort of "spreads" out, it touches the paste from adjacent pad. Then during the acutal reflow portion, one of 3 things happens:
1. The paste gets wicked all to one pin, leaving the other with none
2. The paste bridges the two pins
3. It reflows properly.

So, I though the answer would be a thinner stencil, but that created two other issues:
1. The stencil printer drags the paste up out of the apertures if the pressure is too high
2. The stencil printer leaves paste attached to the paste in the apertures on top of the stencil, which causes no-release.
And if the pressure is "just right" I get a combination of the two scenarios.

Can anyone tell me what I'm doing wrong?

recheck thermal profile for time duration above liquidus, type of solder paste used , exposure time of paste out of the fridge , exposure time of the copper out of plastic wrap for oxidation, and relative humidity away from 50% are also is a major factors.

 

[jnnewton]

The reflow profile is set from the paste tech sheet (Kester NXG1) (45-90sec reflow time), I'm at about 60-seconds. So, this is good.
I don't know if the type is ok. It was recommended by a supplier, I use it for every board, fine pitch or not.
Paste comes out of the fridge, gets put in the printer, and the jar goes back in the fridge or in the trash if it's done. How long is too long for the paste on the printer?
The boards are used as needed, so some may sit unsealed, but I get the same results regardless of whether it's the first or the last board. How long from unsealing to use is too long?
I don't really know what I could do about RH, even if i could measure it. It's not very dry or humid where I'm doing this, but other than that, I don't know.

 

[Tony Stewart]

Has anything changed on solder mask design or material type?

The surface tension forces drive solder to those places which are hottest, and that the only way to guarantee a single joint of the right profile is by having both pad and component termination at the same temperature when paste reflow is initiated with the tracks masked to have a cooler substrate.

So imagine you want the leads to be hotter than the pads for solder to wick up but not too far. These are all affected by reflow parameters, so your profile needs to be adjusted , perhaps as you suspected with a lower ramp and steeper peak.

What is the mask material type and size between pads?
Is this forced air or convection reflow??

 

[jnnewton]

The only thing that changed is that the pitch of the parts used went from 0.65 to 0.4mm. I was getting one or two bridges here or there with 0.65, pretty much no disconnects. Now every other board seems to have one or the other or both. As for the mask width between the pads it appears to be 0.04. Now, this was just based on the give pad width (0.20), mask extension: 0.08 all sides, and the pitch: 0.4, so 0.4-0.2-2*0.08 = 0.04mm

 

[Tony Stewart]

Better try to collaborate with solder paste supplier and equipment supplier .

I would try reducing peak times above liquidus to 30 sec.

 

[Tony Stewart]

Do you have test coupons with dummy parts for experimentation with profiler?

 

[Tony Stewart]

I see some are using different printing methods.
Optipad, Sipad and Precision Pad Technology

 

[Tony Stewart]

Not sure if this helps on pad design etc. https://www.ti.com/lit/an/scea038/scea038.pdf Also stencil aperture shape can help with other shapes.

 

[Tony Stewart]

The reflow profile is set from the paste tech sheet (Kester NXG1) (45-90sec reflow time), I'm at about 60-seconds. So, this is good.
I don't know if the type is ok. It was recommended by a supplier, I use it for every board, fine pitch or not.
Paste comes out of the fridge, gets put in the printer, and the jar goes back in the fridge or in the trash if it's done. How long is too long for the paste on the printer?
The boards are used as needed, so some may sit unsealed, but I get the same results regardless of whether it's the first or the last board. How long from unsealing to use is too long?
I don't really know what I could do about RH, even if i could measure it. It's not very dry or humid where I'm doing this, but other than that, I don't know.

50% humidity improves heat transfer in air to reduce temperature differences between component, paste and tracks.

Too much is a problem as well as too little.

 

[Tony Stewart]

The only thing that changed is that the pitch of the parts used went from 0.65 to 0.4mm. I was getting one or two bridges here or there with 0.65, pretty much no disconnects. Now every other board seems to have one or the other or both. As for the mask width between the pads it appears to be 0.04. Now, this was just based on the give pad width (0.20), mask extension: 0.08 all sides, and the pitch: 0.4, so 0.4-0.2-2*0.08 = 0.04mm

Moving to 0.4mm is a big leap in challenges with Pb free solder. (oxymoron)

This may not help, but ... **broken link removed**

 

[jnnewton]

This is a lot of information. I need to break it down, so I can eliminate things that are OK, and focus on the unknowns.
1. For 0.4mm pitch tqfp parts, is there any reason 0.04mm mask would not be ok?
2. Why would I move my reflow time outside of the paste manufacturers tech sheet limits (to 30 seconds?
3. From the time paste is taken out of the refrigerator, how long do I have until it needs to be reflowed?

 

[Tony Stewart]

Yes there are many variables.

Does printer problem correlate with orientation of pads w.r.t. print direction?
What type of stencil ? laser or acid etch on SS?
Polished?
Do bridge void problems occur after print inspection?

I don't have IPC stds for soldermask handy but that can make a difference with gap size and registration.

The forces of surface tension should be strong enough to attract to closest pad and pin.

Can you show layout portion?

 

[Diver300]

I have heard that you should let the paste warm up to room temperature before you open the tub, to avoid getting condensation on the solder paste.

I know that solder paste has a relatively short shelf life, which can be extended with refrigeration, so you should make sure that you buy paste in small enough containers that you don't need to warm them up too many times.

If you use a tub in a week, there really isn't much point in refrigerating it once it has been started. It's not refrigerated while being shipped.

 

[kubeek]

How thick is your stencil? We use 100 or 120 um for fine pitch boards.