Reflow of 0.5 pitch with oven

Thread Starter

fsonnichsen

Joined Jun 6, 2013
49
I have been using a T-962 oven recently with excellent results. I bought the "mod kit" just in case, but off the shelf the oven works fine for me. I do boards under 75x75mm. I am doing SOIC, 402, TSSOP all with great results when I look at them under the microscope.
Then I encountered the 100 pin PIC24 in 14x14mm TQFP. I have done them with an iron (drag soldering) but would like to oven them with everything else. I am just not getting it and wondered if any of you have experience here.

>> I have used both a stencil, and a syringe-tube (putting a solid line of paste on the pins).
>> I am using Chipquik T3 paste around 20-30microns, usually leaded.

After I am done I always get several bridges. I would think that I should come up "clean" and avoid this step.
The syringe does go on thicker than the stencil but the stencil puts on a nice thin layer--I still get bridges.

What am I doing wrong here?

Thanks all
Fritz
 

jpanhalt

Joined Jan 18, 2008
11,087
Do you have solder resist between the pins? It sounds like too much paste (details about stencil?) or possibly a design problem with the pads. Can you cure the bridging with solder wick? Are the pads too long under the device?

Disclosure: I have only tested my T-962 oven with waste PCBs and have only hand soldered 0.5 mm pitch pins in the past.
 

Thread Starter

fsonnichsen

Joined Jun 6, 2013
49
Thanks jpanhalt. Good questions-
I use a board fab and they put "mask" between the pins. Where the pins meet the boards they to not bridge. The bridging occurs where the pin leaves the board and travels up to the PIC. And worse-this is a big problem-it seems to get behind the pins and bridge on the back.
I really agree with your "too much paste". That was why I went form a syringe to a stencil. I am using a metal one at 100um. The paste looks pretty sparse using the stencil but due to my "shakiness" I tend to set it all little off (half a pin on one side say) and then "mush" it around a bit. With a 16 pin 0.5 pitch for example I can usually get it right down but the 100 pin always leaves an alignment issue on one side.
Wick (I like Chemtronix) seems a bit of a problem--for some reason I couldn't get a melt with a pointed tip. I went to a "hoof" type bevel and now I am getting the melt to reach the wick. This is getting me through for now but messy. Still hope to go all oven as it is a huge time saver for me.

I am really curious with your comments about the pads ---i just grabbed a stencil from KiCAD and it "looked" right. But as you mention there may be better padding but I am not sure what that would be. Do you have any hints here? I wonder if that would help.

Regarding the T-962--its been here a couple months and made a couple dozen perfect boards with 0.5 pitch (but not 100pin pics!) and 402 caps. About 50 parts/board--boards are typically less than 75mmx75mm. One board at a time. They look (and work) perfect under the microscope.
I heard terrible things about it the T-962 so I bought the upgrade kit from England at the same time but I will just hold on to it. I didn't get any "smell" that was mentioned and I have had no problem with the melt (but have not had time to measure the temps). I believe the temperature is up to spec. Either they fixed their problems of maybe for small single boards the temperature curves are OK. I am only using curve 2 for leaded.
I am a physicist but thanks to grants have to do all my own electronics and boards so the oven is a huge time saver for me.

Thanks again!
Fritz
 

nsaspook

Joined Aug 27, 2009
13,086
IMG_20200811_155436.jpgIMG_20200811_155427.jpg

I've had problems with 100+ pin chips and reflow caused by too much paste under the pins, especially at the rear of the landing. The melted solder is pushed out the back as the surface tension pulls the chip down and sometimes there are gaps between pins and lands. A good video of the process.
 

jpanhalt

Joined Jan 18, 2008
11,087
The only upgrades I did to mine were the software (United Engineering, GitHub free), replaced masking tape with Kapton, and the DS18B20 1-wire temp sensor for the reference temp.

Your stencil thickness is the same as mine. Mine is SS. I am 77 and my hands are not steady either. Since you get the smaller ones positioned correctly, I would build some sort of guide that I could position. Offhand, if the backside is a problem, could you lay down some shelf* just over the pads on that side. Position the chip on the three sides, and remove the shelf? My chip (yet to be done) is only 44 pin and 0.65 mm pitch. I am not too worried about that, but it also has a Molex 0.5 mm ZIP for FPC. I am more worried about the Molex melting, so have planned on removing the solder paste from those pads,

Chemtronics is the only solder wick I use. I have tried others too.


*
1598118267597.png
 

TechWise

Joined Aug 24, 2018
151
I have done 0.5mm pitch in the oven before and it worked out okay. It was a 0.5mm pitch 180 pin HSEC connector but I would think the same principles apply.

One thing I found very important was to secure the PCB you're working on in a sturdy manner. Then, I take some unpopuled PCBs of the same thickness and secure them around the one I'm working on. This gives something for the stencil to sit on and prevents it lifting in the middle. Finally, when applying the paste, I press down hard on the stencil near the narrow pitch pins. Firstly, to make absolutely sure it doesn't move. Secondly, to make sure that paste doesn't work its way under the stencil and lift it up as this allows more paste to get under the stencil and bridge the pads.

As a last step, I press down hard on one edge of the stencil and then lift it from the other side. This makes absolutely sure that you don't smudge the paste at all when the stencil comes up.
 

Thread Starter

fsonnichsen

Joined Jun 6, 2013
49
Thanks all. I agree that proper seating of the stencil is important but I am using techniques similar to Techwise and see good results.
Those youtubes always look so good---like watching a concert violist making everything look easy--until you try it yourself! The devil is in the details and it is interesting that almost nothing detailed is published about soldering. I've scoured the web and amazon and just see the simple basic methods-very little talk about layout, pastes, problems etc--probably too boring to the academics out there. But I'll bet there are more soldered joints in the world than just about any other type of technology.

I think maybe setting the chip is my major problem and I will work on it. Probably would be great opportunity to build some type of frame with my vacuum pen but then--where does all the time go!

thanks
Fritz
 

jpanhalt

Joined Jan 18, 2008
11,087
Stencil aperture and thickness get a lot of discussion. And as expected, little theory and almost all personal experience. I went with the advice of the stencil maker I had chosen:
1598271678535.png

I chose 4 mil. But there are claims that is on the thick size.

Unfortunately, current weather here (NE Ohio) prevents me from seeing how reliable that advice was.
 

trebla

Joined Jun 29, 2019
542
I currently have done PCB with 0.4mm pitch dsPIC33 without any issues of bridging. I use temporary rig for fixing PCB on place, for this rig i use OSH stencils acrylic rigs, taped in place. 100um metal stencil is also taped to this rig. I put paste on stencil like sausage with half inch diameter, then bigger weight of solder paste helps get nicely get thru stencil openings, especially if you have bigger boards. Then i spread paste with squegee about 1cm/s speed. Stencils lifting speed is also important, it must be not too fast or low, i have practicing alot with this. For bigger boards i use framed stencils and stencil printer, tape does not hold stencils on rig with bigger boards until you reach the other end of board wit squegee.

I have noticed if PCB does not warm up enough in reflow owen, then is bigger possibility to form up solder bridges.

And if you use such fine pitch devices you better order PCB-s with ENIG surface, not HASL.
 

Thread Starter

fsonnichsen

Joined Jun 6, 2013
49
Thanks all again. I will focus my efforts in the stencil going forward as I think this is the culprit, especially after reading your comments. I have 3 PIC24 boards to kick out this week--fingers crossed.

Hot weather here on Cape Cod for 6 weeks--make my Scandinavian blood boil----same in NE ohio?

Cheers all
Fritz
 

Thread Starter

fsonnichsen

Joined Jun 6, 2013
49
Here is an addendum to this post. I finally am getting good results soldering 100pin PICs. Thanks to the people on this post. There are 3 things alluded to here that I did:
1) I made sure that the stencil is very secure--I paid a lot of attention to this and built a simple $7 clamp to help out. While I like electrical tape because it is simple, for the high density stuff I find this jig very useful. I am attaching photos of the jig and its clamping the stencil. I use 3 clamps as you may need to lift one to get underneath for complex chips. I chose 2 spacings for most of the stencils I used. I use metal stencils-I found that simple plywood had just the right friction to hold the stencil.

2) After laying down the initial paste I make a "clearing run" with the pasting card, applying a fair amount of pressure. This insures a pretty thin layer. Just like hand soldering I found it was easy to use too much solder when learning. I check the solder with the microscope before laying chips.

3) I did pretty well with the above methods but got even better results re-drawing my footprint. My original was from the PIC datasheet. It had very long traces, these running well underneath the legs. I shortened them around the inner perimeter leaving little space for the solder to creep. Photo attached.

There is a Moores law for transistors but none for joints. Probably there are at least 6 joints for every IC. With all the thousands of volumes out there on ICs I could find none with detailed information on soldering techniques. Luckily there are forums like these.

Thanks all
Fritz
 

Attachments

trebla

Joined Jun 29, 2019
542
From my exprience, the best results for solder paste laying is achieved when PCB and these OSH stencils acrylic rigs are about the same height (thickness) so stencil can touch the PCB and rigs surface evenly. Then you can focus only to stencil lateral fixing stability. If stencil moves slightly laterally then is somewhat difficult to place components precicely because solder paste offset disturbs you visually and can push component mechanically to uncoated side direction as usually very small forces applied to place the component onto PCB.
 
Top