Solder Paste Blues

A little background...

thinking I was the SMD whizz, I got 4 layer boards made recently, and tried soldering on some 28pin SSOP devices, using my hakko soldering iron, some lead free solder and a flux pen.
I've done similar devices before, with no issue, on 2 layer boards, however, I don't think I was quite so ready to deal with 4 layer boards (something tells me that the ground plane was pulling the heat away from the soldering point).
Anyway, long story short, I managed to burn off some of the pads on the boards. (I had 2 boards made locally for $180...). I'm going to try and rescue those boards, as I soldered the other SMD devices on there perfectly (or significantly better). I suppose I'll directly wire the pins without pads to their destination, and hope like hell.

Anyway, I want to avoid this in the future, so I'd like to get some solder paste and try the skillet method.

However, I'm stuck on what kind of solder paste to buy. SRA solder has loads of different ones at:
http://store.sra-solder.com/section.php/10/1/568720496c2ee22287d056385401ccf8

I'm a little stuck on which one to buy.

for RoHs compliance, it needs to be Lead free, but all the ones without "Pb" are very high temperature, and have a nasty flux that needs washing off with hot water and detergent! (according to that website). The temptation of using lead in the solder is very strong, but these products, once i'm in low quantity production could find themselves being shipped to Europe.

Any advice on this would be very much appreciated.

There are lots of "holes" in the ROHS fiasco...  learn how to use the rules to your advantage..  ;D

Generally the plating materials and the heat curves can tell you what kind of solder you want to use.  Are you stuffing these yourself or having a board stuffer do it for you?  If it's the latter I would ask them what kind that they use.  If it's the first then you might call a stuffer and ask them about their heat curves and solder choices as you would normally ask when scouting for a proper stuffer.

As a rule of thumb, ROHS solder will need higher temps to flow.  The rosin/flux will also be a bit harsher because the lead-free crap needs a lot of help to wet to the platings.

If you are stuffing by hand with an iron, I would NOT suggest using paste solder.  The problem is that the paste is actually small spheres of solder which will lodge themselves under/between pins and will cause shorts if you don't fully flow the board.

As a rule, use an iron that is MUCH HOTTER than you think you need.  This allows you to heat the pin/pad much faster than you would with a lower power iron.  While most think that this will heat the board too much, it actually doesn't, it's the opposite.  Holding a lower heat on the PCB for longer is much worse than a higher heat for a short period of time.

Svart said:

As a rule, use an iron that is MUCH HOTTER than you think you need.  This allows you to heat the pin/pad much faster than you would with a lower power iron.  While most think that this will heat the board too much, it actually doesn't, it's the opposite.  Holding a lower heat on the PCB for longer is much worse than a higher heat for a short period of time.

Click to expand...

+1. You can also use a special hot plate to pre-heat the board while you're soldering.

Rochey said:

However, I'm stuck on what kind of solder paste to buy.

Click to expand...

Take one step at a time. Buy the 'easiest' paste to get experience and for your own protos; this is likely gong to be leaded. The pastes you link to are cheap enough to do some experimenting. Use blank boards first; if you don't have enough of your own designs SparkFun sells dud boards which are great to get your feet wet. Once you're confident enough migrate to Pb-free paste.

On cleaning: if all your parts can be water-cleaned, it can be useful to invest in an ultrasonic cleaner such as those used to clean jewelry; those go for 40-50$. You will still need to finish up with polypropylene alcohol and water.

JDB.

don't I stand a higher chance of melting the traces off the board with a higher temp? (sorry if that seems dumb)

Rochey said:

don't I stand a higher chance of melting the traces off the board with a higher temp? (sorry if that seems dumb)

Click to expand...

Sure, any board will delaminate if you cook it long and hot enough. Time and temperature are factors, and I believe the IPC have extensive data on what a board is expected to deal with. Having said that, I suspect very few board manufacturers will sell you a board that can't handle a standard lead-free soldering temperature profile (plus some tolerance).

Get a test board, put some hi-temp paste on it, put it on your skillet, see how much time elapses between the solder melts/flows and the board starts to delaminate. If this amount of time is negative, call your board house and ask for suggestions.

JDB.

It's thermodynamics vs. time.  Once you bond the metals the heat starts to move to areas of lesser heat.  So if you heat fast, the area of high heat is very small and then diminishes quickly as the heat dissipates through all of the other traces/surfaces/into the air/etc so the amount of time and mass of high heat is actually very short and small.  Holding lower heat on a pin for longer periods allows the heat to spread a lot more and heat a lot more area.  This is called heatsoaking.  Once an area is heatsoaked, it can't heatsink any more heat away and you'll start to severely stress the materials.  It'll take a lot longer to dissipate heat from a heatsoaked area since there is technically more heat stored due to the amount of mass that has been heated.

Think of it in terms of current vs. time.  A MOSFET generally has 2 current ratings, one for continuous draw and one for current draw during a given duty cycle, usually around 50%.  The current rating for the 50% duty switching FET is usually much higher because you aren't heating the part as much for a large increase in current because the waste heat is averaging about the same as the continuous part.

> don't I stand a higher chance of melting the traces off the board with a higher temp?

Sure, if you are asleep.

What we are telling you is: VERY hot, VERY quick. You want heat flowing from iron to pad much faster than from pad to board, lift the iron the instant the -top- of the pad has full melting. At this point the bottom of the pad may still be "cool" (below glue-melt temp).

I don't do sub-visible parts. But on DIPs on very cheezy half-baked PCB stuff, I used a 45W UN-regulated iron that glowed in the dark, was often in-and-out in 4 seconds, the board barely warm.

Hello PRR,

I've done a bit of smd on 2 layer boards before, with few issues. (Boards were from goldphoenix pcb). These new boards were 4 layer from sunstone.

The soldering experience with the ssop packages were completely different. 2/2 on this package was a disaster.

I'll post pictures later this week.

/R

hmm I've been soldering to 2, 4, 6 and 8 layer boards for years and have found no real difference in how they solder.  I'm not sure what is going on in your case but you shouldn't be having that much trouble with them.

I routinely do leadless carrier quads(VQFN) on 4 layer boards.  I also do tssop/mssop and even BGAs courtesy of a heatgun, solderpaste and lots of rosin.

What iron/hotplate/hotair are you using?

it's a hakko 936, some lead free solder, and an MG Chemicals flux pen.

I haven't tried the skillet thing yet although i did order some paste etc from SRA yesterday.

I think i'll try it on one of my cheaper 2 layer boards first

try using real rosin, not flux.  this makes life easier, especially when dealing with lead-free stuff.

I just tried the skillet reflow thing.. all I have to say is WOWOWOWOWOWOWOWOW
So I had a simple board, a few 1206 resistors, and two OPA2134's.
I tried putting the solder paste down by hand from a syringe... TAKES A LOT OF WORK!

In other words, I did kinda  messy job of it, a big blob here, a smaller blob there etc.

So I threw is on my $20 skillet (electric frying pan for Europeans) and turned up the heat (about half way for the board to warm up, then 3/4's of the way). The paste looked like it got wetter, but still dull colored.

I started thinking "aaah sh*t, I've wasted more money *sigh*"

Then, all of a sudden, the solder turned shiny, and the parts started to move by themselves, they centered themselves.
My IC's lined up perfectly on the pads, the resistors lined up nicely 'n all. Amazing.

I will post more to my blog soon. I just wanted to share my joy  ;D

can you tell I'm excited?

Now that sounds cost effective...  good job.

I wonder if those hot air stations are the magic... should be gentle on the PCB traces and still allow the the surface tension of the solder to self center the parts...

JR

+1 on the hotter iron.  Works well if you're having melty wire problems too.

Skillet as a substitute for an expensive pre-warmer is cool !!

I suspect that the vias to the planes were not done with thermal relief pads.  Is that so? A surprisingly common mistake which is most vexing when desoldering through-hole stuff.

I hope people get over their SMT shyness and start working on DIY solutions to this kind of production, issues of SMT component quality vs through-hole aside.

BTW, the skillet should be good for rework too.  Just keep the heat a little lower and then use hot air or a rework iron on the part in question.  Probably a good idea to get familiar with the dwell time vs temp for the part as well.

Hot air will blow the parts around if you aren't careful.  When doing things that need lots of heat I routinely use a hot, large tipped iron and a hot air pencil or simply two hot irons but most things I've done really don't need more than one iron.  I just turn my JBC up to 800F and go.  I rarely turn it down unless I need to draw solder long distances like for bridging resistor pads and the like.

I have the blues again.

Here's the story... fired up the skillet tonight, and got out the solder paste syringe.
Lots of huffing and puffing later, virtually no solder paste comes out... then all of a sudden, all the flux that's mixed with the paste starts to come out.

So I take the nozzle off, get a leg of a resistor, and start to prod into teh syringe, and it appears that most of the syringe inside is solid. (very disapointing).

So, I got some of the paste out, mixed it with some of the flux that had already come out of the syringe to make a workable paste, and put the paste on pad by pad myself.

The paste has been safely in the fridge, only taken out when it needs to be used. is that what I did wrong?

I tend to keep mine sitting on the bench.. but then again I HATE using paste manually just for these sorts of things.  :

I mix a little extra rosin or flux with the paste depending on what I'm doing with it too.  I like to keep it rather on the soupy side and less solid than most people care to use.

SVART.. you've mentioned doing BGA with a heat gun, do you just heat the part from the top, or heat the PCB from bottom too?

I'm not brave enough to try to MacGuyver a BGA on my bench but I have one oddball SMT part with a heat sink pad on the bottom of the chip...  I put a via there so I should be able to heat from the bottom with an iron if needed.

I've got some 0603 caps in this design but they don't scare me as much as the tight lead spacing on the ICs, but ask me again later after I actually melt some solder.  Hopefully i can get some surface tension self alignment going on with the ICs.

JR

just a brief update. I've decided from here on out to mix my own solder paste (to get the ideal consistency).

I did a little bit of shopping around - seems that a website called "dealextreme" is a pretty good place to buy things like solder from.

Unfortunately, the solder still has lead in it. But for prototyping, it should be fine.

I got a pot of solder for about $5, and some flux to soften it up a little for about $3.

I'll report back on when they arrive.

/R

Well I've melted my first solder paste and all I can say is be not afraid...  I didn't spring for a stencil and if I was going to do any number of boards then I surely would. Laying down the paste by hand on very tight layouts is too difficult for me to do well enough for production. For now I just loaded up one PCB since every value on it is subject to change...

FWIW it took me twice as long to clean up the solder bridges, but I used some pretty fine pitch devices and resistor arrays, etc.  The individual parts, even 0603 caps were relatively easy... however on the ICs tight lead spacing was prone to bridge, but again my paste application was probably far from optimal. In hindsight less would be better than too much, because you can always touch up joints that don't hold.

Fun part was cooking the board in my kitchen oven... several minute preheat at 300' to get it happy then crank it up to 450' until the paste melts... A little like baking.. but at higher temps.

I just stuck it under my cheap USB microscope to confirm I actually cleared all the bridges before I apply electricity. I found using the back light on the microscope was helpful to see that it was clear between pads, better than looking from above.. the dynamic range from the crappy  camera chip in the microscope sucks, so the glare from the shiny solder saturates the camera before you can see everything, but it's a lot better than my tired old eyeballs. 

I wouldn't try to do this for production, but I am no longer apprehensive about prototyping SMT. With a stencil even BGA should be doable.

JR