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

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Wave Soldering Thick PCB's

JohnW

#18694

Wave Soldering Thick PCB's | 21 January, 2002

Here's a question for you to mull over. I've got a PCB that is 4 millimeters thick, that's about 160thou of USA money if you like it that way. The question is IPC reg's for hole fill say it should be a min of 75% hole fill and 50% on ground planes. All well and good, but these tandards weren't designed for today's thicker PCB's. So I'm wondering what other folk's work too when it comes to thing s lke backplanes and so forth. I'd be interested to hear your thoughts.

JohnW

reply »

#18704

Wave Soldering Thick PCB's | 22 January, 2002

Oooo, thanks for the conversion. Being metrically challenged, az a skratchin ma haid thar fer a mite. We do a G-10 160 thou and have similar issues.

Now, we�ve talked about barrel fill [ie, slowing belt speed, increasing pre-heat temps, etc] before. And this aint one-a-doz, because we already tried all that stuff a couple months ago. It�s a: * Bin blanked by the designer. * Exhausting the flux before hitting the wave, which is not a good thing ... pikes, bridges, etc.... * So, how much higher than 500�F [260�C] should I go? * Setting the wave depth so low, I could crack a walnut betwixt my cheeks, while watching the board go deep into the wave. In other words, it's a DEEP mamajama!!! And that�s with stiffeners added.

Then try these tricks ... * Turn on the chip wave. * Try a foam fluxer. * Flux, preheat, then flux, preheat, then wave. * Paint flux on the primary side before loading the machine.

Next, you�ve got to convince yourself that these are not process problems by: * X-raying the connections. * Doing destructive analysis of the solder connections.

When all else fails and we can�t follow IPC 75% hole fill rule, we waive the rule, based on: * Theory #1: If 50% of a 0.060" thick board (0.030") is OK, then a 25% hole fill of a 0.150" thick board (0.0375") would also be OK. * Theory #2: For most applications, the 75% rule is overkill. There are tens of thousands of single-sided PCB [= 0% fill] in under-the-hood automotive applications [the worst loading environment for the most common applications] that, unless there are some other dumb design features, last the life of the car.

Be interesting to see which theory your customer bites on, eh?

reply »

Peterson

#18730

Wave Soldering Thick PCB's | 28 January, 2002

Being in the backplane biz, we too, encounter similar problems...especially when our customer reads the IPC as Biblical (you will achieve 75% fill or be turned into a pillar of salt). Anyway, recently, one of our process engineers suggested paste-hole for large caps and large connectors. He would like to design a stencil and roll 'em right through reflow. Any thoughts on this approach? We have been mildly successful with the previously mentioned methods (especially preheating the heck out of the board and manually fluxing before wave. I would prefer a faster process however. Is Paste in Hole the answer?

Thanks for your comments.

reply »

JohnW

#18735

Wave Soldering Thick PCB's | 28 January, 2002

Sorry I keep forgetting you lot are backwards in the old metric system :0)

anyhoo, yes IPC, the bastions of right well as long as your board is only 1.6mm thick (sorry 62mils)say that you need 75% min general and 50% on ground planes. I'd like to see em try!

We have created a correlation between our 5DX process and some boards we sectioned, the customer was well impressed but said well at least now you can check you pass IPC before you give us the boards!...

Anyhoo for me I generally solder at 240deg c, that's 464F and get good results. On this one particular board I'm at 260 but getting dam close to secondary reflow. The problem is you can have things as hot or cold as you like but when your via's go all the way thro and arn't tented you tend to transfer heat straight up so again problems. Poorly designed boards for selective wave also creates problems with masking so life is fun! In term's of the flux the words for the week are Sublimation and sublim. Basically the flux goes liquid to gas in 1 easy move. Different fluxes have different rates of sublimation that you need to work out (flux a cold board and weigh it & plot on a graph, up the temp and repeat , do this till you get to zero flux and you should produce your curve, typically it's exponetial) Anyhoo you can actually cheat a little, you can up the flux knowing your gong to sublimate (that's burn off to you and me)so you still have soem left for the wave and get good results and very little in the way of tackiness. Now onthe point of foam fluxers, well a couple of things cause us problems here. The 1st is we're a no clean site, aint got non. and 2nd with the heat you have to put into the cards you need a heck of a lot of pallets to allow them to cool down properly before they hit the foam and collaps it, besides I'm never entirly sure that you'll get teh hole fill you need but I'm open to opposing views on that. Paintig the flux on, we did think about but decided it wasn't controlable or repeatable so scored it off the list but it would help a little. Adding the chip wave will help thro' solder up the barrel and add a little extra heat to give you a boost, but will help kill the flux hence add a little more. We've also seem that lead length is a help / hinder, most of the leads don't come thro the bottom of the card so you have less thermal transfer up the lead / barrel, getting longer leads helps this. Remember lead protrusion should be 1.1x the diameter of the anulus. The biggest thing is design the board right and don't connect everyting to the ground planes, and definately not to all 4!!!!!! count em 4!!!! I've actually been thinking about working out the mechanical stregths of the joints but figured my head wouldn't take it...

Anyhoo I'm calm now.....

We looked at intrusive reflow for a few things but the problem is not all components are suitable for this, connectors have to be specifically spec'd for that and I've personally never tried PTH cap's and not sure I would for fear of a large black singed area on the PCB where the cap used to be, but let me know how you do. I guess the biggest hurdle is getting enough paste couple of methods, stepped stencils or use a combination of squeegee types 1 metal 1 rubber and do a 2 stroke run

anyhoo back to my nice warm solder machine and thick boards....

reply »

#18741

Wave Soldering Thick PCB's | 28 January, 2002

Thanks for cutting me some slack on that metric thang, pardner. ;-)

ANALYSIS: What??? The 5DX is useful for something!!! Shut ma mouth!!! [Don't you say a thing. I think I know what you're thinking.]

FOAM FLUXERS: Your correct, not having a foam fluxer / doing NC flux and using pallets are good reasons that foam fluxers will not work. Yes, sometimes you can get good push with a foamer, and sometime you can�t. But either way [sprayer or foamer], you need to be getting foam on the top side of the board, which sometimes is hard to do with thick boards. Upper barrel flux is: * Toughest to get in-place. * Thinnest material, so it burns-up [sublimates, ooo coo] the before lower portions of the barrel.

PAINTING FLUX: Yes, painting NC flux likely would leave some nasty res.

CHIP WAVE ISSUES: Yes. you�re correct. And chips seem to grow dross as fast the main, yet it�s 1/3 the size.

CONNECTING BARRELS TO GROUND PLANES: The connecting of ground planes to through holes part is OK. It�s the method that�s used to make the connection that causes the problem.

MECHANICAL STRENGTHS OF THE JOINTS: Even if you do know the mechanical strength of the connections, it could be tough to relate those numbers to the end-use of the products. So, it�s a good thing that you decided to save your brain.

reply »

#18742

Wave Soldering Thick PCB's | 28 January, 2002

Well, reflow certainly would increase your preheat, wouldn�t it? On the other hand, some PTH [and SMT] components cannot bear the heat of a reflow oven.

Some boards will never see proper barrel fill, because of either: * Design of the board. * Slipshod fabrication methods.

To better understand the possibilities of pasting PTH parts, consider: * Squirting some paste on a bare board, reflowing it, and assessing your hole fill. * Toasting some components to better understand Tc.

Search the fine SMTnet Archives to develop background on the area.

As John says, it�s tough with pin & paste to get enough material on the board to get hole fill, especially with: * Thick boards * Connectors with more than two rows of pins.

reply »

reflow oven profiler