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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Pick and Place Startup - LQPF100 bridging issues

Views: 4557

#86440

Pick and Place Startup - LQPF100 bridging issues | 23 March, 2021

Hello everyone,

I work for a company that, up until now, has done everything using through-hole components. This has been fine for years, but we have been pushed into the world of surface mount manufacturing by a new product that is far cheaper and faster to manufacture this way. Our setup consists of a manual stencil printer, CHMT36VA desktop pick and place machine, and a DDM Novastar GF-B-HT batch reflow oven. The paste we are using is ChipQuik TS391AX500C.

We have been producing boards for the past 2 weeks and everything is going somewhat alright, but the main issue we are having is the LQFP100 package MCU on our board is highly susceptible to bridging across the pins. For a while I have managed to get around this by using older, dried out paste, which sucks to get through the stencil, but it does not 'slump' (I believe that's what it is called), which is made worse by the fact that the pick and place machine does not place the chip very nicely, requiring it to be pushed into a better place with tweezers. Using the dried paste, pushing the chip around does not smear it, and it will generally solder without bridges. But using new solder and having to push the chip (and smearing the paste) makes a terrible amount of them.

To me the issue here is the pick and place machine, because if we did not have to push the chip then we wouldn't have to smear the paste, and in cases that I have had that happen, the board turns out alright. But I don't think such a cheap machine will be able to place them with repeatable reliability, so are there any other solutions anyone would recommend? Right now it is taking far too long to fix bridges, and makes a rather ugly mess of the board too.

Another issue we have is consistent printing as well. Sometimes paste will stick to the board and not the stencil apertures as it should, other times it will not stick to the board at all, seemingly at random. Any help is greatly appreciated!

reply »

#86443

Pick and Place Startup - LQPF100 bridging issues | 24 March, 2021

Sounds like you need help in several areas. Firstly, using dry paste is never the correct solution...you will achieve poor consistency. Lets go back to basics....check the pad sizes on the pcb are roughly as advised by the device manufacturer...or the IPC standard. Design the aperture so it is slightly smaller than the pcb pad, but making sure you adhere to the >0.66 area aspect ratio...so for example...if the device is 0.5mm pitch QFP, the pcb pad may be 12 thou wide..make the aperture 9 to 10 thou wide...and reduce it's length by 10 % overall compared to the pcb pad.... Depending on device pitch, select appropriate stencil thickness....maybe 5 thou thick for a 0.5pitch QFP is a start. This set up should give you about the right amount of paste in the the right place. When printing, slowly drag the blade across the stencil, just enough down force to wipe the stencil clean after the blade....the paste bead must roll in front of the blade to give consistent aperture fill....DO NOT SQUEEZE PASTE INTO THE APERTURE...you will get poor print definition and more likelihood of bridging. Inspect the print under a microscope...make sure there is no paste bridging or bleeding.....incidentally, even if you get bridging at this stage...a good paste will reflow the bridges away and the joints will be bridge free (unless you have far too much paste on each deposit)...it's one of the properties modern pastes have nowadays....great coalescence properties. In placement...gently push the device into the paste...about 25% of the depth into the paste is a good guide....it is not unusual to see the paste deposits deform after placement...but ideally they will not be bridging...but even if they are....a decent paste will reflow the bridges away. Reflow...most modern pastes have massive reflow process windows nowadays...but if you have a paste that hot slumps you have to be more careful setting the profile. If you are happy that the paste isn't hot slumping, but you still get bridges...you can try to optimise the aperture sizes...or try a different paste...Almit has a paste that should make your life easier.... Hope this helps ? thanks Graham

reply »

#86444

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

Well everything Graham said with some additional points..

You mention manual printing but very little about your stencil or setup. Everything you have said indicates a very limited budget has been expended, however what you have is an extreme example of the fact printing is the most important process to get right. So if you are using an unframed floppy piece of stainless as a stencil, stop and order one in a frame (or for a tensioning system if you have one).

If ordering from China (JLC/PCBWay etc) choose the polished option and you really want to be specifying the thickness. I'd be tempted to suggest for your process that should be 0.1mm (4 thou) just because its likely a manual print with a cheap stencil is depositing too much paste anyway. Bear in mind with stencils that part of being cheap also comes with zero input from the manufacturer, they will cut what you send, so you need to get your stencil layer exactly as it should be. A "proper" stencil manufacturer will take your entire design file and either create the paste layer from scratch or tweak the existing one to make sure it complies with IPC standards as best as possible and can suggest advanced options like stepped stencils where they might help or apply aperture adjustments to certain packages.

On the placement front you have a very small very lightweight machine with no encoders or other features that keep bigger machines accurate and reliable for decades. That means you need to fiddle and tweak everything about how it images and moves when handling your fine pitch part. That will include moving slowly, only moving in one axis at a time, making whatever adjustments you can to get the best possible image and be absolutely sure your PCB isn't moving a tiny bit when the other parts are being placed.

I have said this in many places, Youtube of full of DIY,HomeLab hackers and influencers with videos of manually pasting boards. Its fairly clear none of them have EVER worked in a production environment and they ALL do it badly. The "roll" Graham mentions is important, you are not squishing paste through holes, that will cause the paste to spread under the stencil, giving a rubbish print and spoiling the gasket between stencil and PCB and thereby the next print too.

Lastly, you might find a QFN package rather than a QFP is less prone to shorts as there are no leads to wick up paste - YMMV.

reply »

#86446

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

Your missing the point. The placement machine is the primary issue. Does this toy even say it can handle that component? If you have the ability try slowing the machine down when it is placing that component. Slow X, Y, and Theta axis.

Jerry

reply »

#86447

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

I think between us we've not missed too many points...

reply »

#86448

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

That is all a lot to take in, but thank you all for the recommendations. From what I understand, I'm going to start by ordering a new polished 0.1mm stencil with size reduced apertures and seeing if reduced paste deposit will help.

As for the stencil setup, we are using a hinged holder with framed stencils from PCBWay. I have been using Eagle for PCB design and just using the auto-cream on the pads which just defines cream over the entire pad. I assume you are correct that PCBWay does nothing to adjust and just manufactures the apertures as defined, so I went and followed your recommendations to make the aperture smaller.

Would it be reasonable to think that with the new stencil and reduced paste deposit that even if we still have to adjust the placement of the chip by hand that bridging will be reduced? From what I gather that will greatly reduce or eliminate bridging.

As for the PnP machine, I don't think there is much else I can do to get it to place any more accurately, though my boss has now said he wouldn't be entirely opposed to spending 40k or less on a better machine.

Attachments:

reply »

#86449

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

What is the QFP pitch ? What aperture dimensions have you chosen ? Is the stencil laser cut then polished ?

Your AFTER stencil size looks promising...your BEFORE stencil aperture size looks disastrous...would surely give bridges..

Your pcb pad size looks good (noticeably wider than half the device pitch)...as Stephen says in the post below...his rule of thumb is good advice...not sure I'd go as far as make the pad width 3/4 of the device pitch...maybe 2/3..but the difference is probably not critical...but it's a top tip to help with manufacturing the assembly.

reply »

#86450

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

A couple more things. The PCB. What size are the pads? One time a board shop had difficulty putting masking between the pads of a QFP and the pads suffered as a result. When I told them they didn't have to have masking between the pads, the pads were much better and our bridging problem went away. Some people thought that would be bad because the mask acts as a dam. A rule of thumb I found works for QFPs is make the PCB pads 3/4 of the pitch and the stencil aperature 1/2 the pitch. It might sound counter intuitive to have the pads wider which means closer but you want the solder to be able to go up the side of the lead. If the lead is as wide as the pad then the solder can't do that.

What finish is the PCB? I would not be surprised if it is HASL which can be problematic for fine pitch QFPs.

Something the others probably think is not worth mentioning because everyone knows it is that the SMT parts do a dance in the reflow. Seriously the parts move during reflow, fortunately they tend to move where they should be. If a part has more than half the lead on the pad, you probably don't have to adjust it. And if the leads are 3/4 on the pad then you are not going to make it better by shifting it.

A rule of thumb I found works for QFPs is make the PCB pads 3/4 of the pitch and the stencil aperture 1/2 the pitch. It might sound counter intuitive to have the pads wider which means closer but you want the solder to be able to go up the side of the lead. If the lead is as wide as the pad then the solder can't do that.

reply »

#86451

Pick and Place Startup - LQPF100 bridging issues | 25 March, 2021

The QFP pitch is 0.5mm. Pad size is 1.475 x 0.3mm, and the new aperture size is 1.32 x 0.22mm. I'm not sure about whether or not it is lasered, but on the newly ordered stencil I selected the polishing option. Looks like the pad width is already 2/3 of the pitch so that checks out.

Board finish is HASL w/lead. What different surface finish would you recommend for our situation? Also our reflow oven is drawer based with a glass top, so I have watched parts be sucked into place. When I was using the dry paste the parts clearly did not do this as well as new paste. The PnP machine also seems to have gotten better at placing the chip straighter, so that might even just eliminate the need to move the chip at all.

reply »

#86455

Pick and Place Startup - LQPF100 bridging issues | 26 March, 2021

Leaded HASL on small pads can be quite difficult to print on...the HASL tends to be domed rather than flat..and that can affect print quality...you really need a perfectly flat pcb pad to print paste on when you are assembling 0.5mm QFPs.

If the pcb supplier is making the pads 'flat' with HASL he runs the risk of blowing all the solder off the pads and leaving just tin/copper intermetallic on the pcb pad...it is flat, so that's good...but it's a tricky layer to solder to compared to pure HASL solder.....tin/copper intermetallic has much worse solderability than a good HASL layer...which is of course pure solder....nothing solders better than solder as we say.

Any layer that is flat is a good choice for pcb pad solderable coating....ENIG is popular..and would probably be my first choice...price may be higher from pcb supplier...speak to the pcb supplier about what FLAT coatings they offer for pcb pads.

reply »


Tom

#86574

Pick and Place Startup - LQPF100 bridging issues | 29 April, 2021

Buy a good used PNP machine from a dealer in your area. There is no comparison when you are using these machines for production. Unfortunately the ddm equipment you describe is unlikely to perform any better and the time you spend trying would be better spend installing and learning proper equipment.

reply »

#86575

Pick and Place Startup - LQPF100 bridging issues | 29 April, 2021

I have the same issue but with one customer only. All the other have smaller or more difficult QFPs with really low bridging issues - close to no-one. I think the PCB and PCB design also an issue - other customers are using 32/64/100/144 QFPs and we have clearly zero bridging - and squeegee setup also important. Strange but we never had problems with QFNs and have problems with SSOPs - SSOP solved with using other supplier, seems to be leg making issue or old component.

In the past I over print them a bit, used lower clearane on solderpads to place more solder on legs - avoid voiding or paste-stucking in appertures. Now, we are following the IPC stencil guidelines. First time we used 150um stencil haven't better results and now 130um. 130um have a little bit better results but have 1-4 bridge on each QFP.

Now I found that the squeegee, squeegee pressure and support for PCBs make a big sentense.

May I try a new blade. We have a quite old, 4-5 years old blade but we never use it more than 500 cycle a week.

Lower squeegee pressure. Now we are on 150N, in the past we had 60-70N. With 60-70N we had way better results but I have to recalibrate the squeegee pressure-feedback module before use lower value.

Transport not seems to be fully paralell. One side seems to be lower a little bit - 0.3-1mm - so I have to check this too in printing position.

reply »

Fluid Dispensers

Electronics Equipment Consignment