And welcome to Board Talk with Jim Hall and Phil Zarrow, the Assembly Brothers pick and place.
This comes from G.F.
And it's a really well defined question, giving us lots of data to dig our teeth into.
In other words the question is going to take longer than the answer.
Okay, contract manufacturer is producing power supplies for us with no-clean leaded solder paste, it's a large component, it's half inch square, sit flush on the board. We're having failures after only a few days of operation that looked to be the result of an electrical short. When we mechanically remove these large parts we find a tacky substance, the consistency of rubber cement underneath.
Our initial thought is that this is uncured no-clean flux from the solder paste that cannot escape from under the component during reflow. The no-clean paste manufacturer said that it's impossible for the flux not to be cured after reflow. However, in its uncured state, it would be tacky and would break down under voltage, these parts see 300 volts DC.
So the question, that's the situation, is this uncured residue common when no-clean leaded solder paste is used with large, no clearance parts? What causes this? Can we bake the completed assemblies to ensure the material is fully cured? Does the part need to be changed to one with a clearance gap?
If you have some paste under there, I assume you're creating some sort of solder joint, so you have to have some clearance. The joint, the height of the joint so is there a zero-clearance or why would there be solder paste under there?
So I assume it's short of a QFN of some sort, or BTC Bottom Terminated Component, IPC spec 7093. But first the idea of a large part, there is a very good article that was published several years ago when BTCs first came on the market by one of the cleaning gurus and contamination gurus of our industry, Terry Munson and he did just this.
He pulled off a part after it was soldered at the proper reflow profile and this wasn't even that big, it was a relatively small, bottom terminated component and he found sticky residues underneath and analyzed them and they were in fact uncured no-clean flux from the solder paste.
His analysis stated that the flux around the perimeter of the part exposed directly to your heating environment cured and hardened more quickly than the paste under the center of the component. Creating a barrier, a relatively solid barrier around sealing off the perimeter, preventing the escape of volatiles from the paste remaining under the part.
Remember, no-clean starts out active, if it is properly heated it should be deactivated. Three mechanisms are used by most no-clean formulations. The first is chemical reaction. Take active chemicals and react them into something that is no longer an active chemistry. Encapsulation by the resins of any residual active chemistry, and evaporation. And Terry theorized that this barrier around the perimeter prevented evaporation which creates this uncured, sticky residue that you are seeing.
This is a common problem dealing with low clearance parts assuming this is more of a BTC which will have a very minimal clearance, 2 or 3 mils and there are solder paste formulations that can help with this. Printing particularly, if you have a large ground plane you're printing a lot of paste under there, some sort of window-paning is a common technique that is used.
There's some other venting methodologies as well.
Modifying your reflow profile -- some tests have shown that a long soak section, a longer soak before reflow can minimize these problems.
Another approach has been using solder preform which have a lower flux content -- Indium has been experimenting with that and it's worth looking into.
And Alpha has just published new information on that. Certainly if you have a part that provides a larger gap under there for whatever reason I don't know how you would do that with -- if this is a true BTC, but again we're speculating because as good as your definition is here, if you had described the part a little bit better.
This is for all you future questioners so those are basically the tools that you have available, but certainly if you can use parts better with a higher gap underneath, you can minimize the possibility of corrosion, entrapment and cleaning problems and so forth, whether you are using a no-clean or not.
So I hope we answered your question. This is Phil and Jim, the Assembly Brothers.
Don't solder like my brother.
And don't solder like my brother.
It would be helpful to know the type of component that is giving the issue. We are all speculating based on the component being a BTC/QFN. The gentlemen gave good recommendations for reducing the residues. If none of those approaches work for you, there are solder preform solutions that can fix a joint height and significantly reduce solder residues under you component. Feel free to contact me alphaassembly.com
Jerry Sidone, Alpha Assembly Solutions
Bottom Terminated Components, such as QFNs, can leave a significant level of flux residue under the component termination. The high volume of solder under the component in combination with standoff clearances of less than 2 mils can leave a significant level of tacky flux under the component. |
As the solder in the ground lug reflows, large voids can form due to poor flux outgassing. The flux in the streets accumulate and further block outgassing channels.
Many research studies find that this flux is active and will form leakage currents in the presence of humidity and bias.
Board design options that increase standoff gaps and place thermal vias in the area of the ground lug will help. Both of these design options improve outgassing and reduce ionic constituents in the residue left behind.
Mike Bixenman, Kyzen Corporation
The gummy residue under the QFN is flux trapped by the sealing of the flux at the edge of the QFN perimeter pads. Since the QFN is pulled down to 1.5 to 0.5 mils off the board surface it holds the flux activators and does not allow proper volitization of the flux.|
Two QFNs locations on a board and the one with a QFN placed will have a no clean flux residue level of 180-225 ug/in2 of WOA, where the QFN pad with no component will have a WOA level of 10-20 ug/in2. No Clean flux is only inert if the carrier and activators volitalize and vent properly.
No clean does not mean no residue. If you would like to see dendrites growing through the gooey flux or measured parasitic leakage contact me at www.foresiteinc.com. No the high voltage does not but even low voltage with sensitive circuits will failure or RF circuits between the power pad and ground plane.
Terry Munson, Terry Munson, USA
I'd like to hear you discuss more the failure analysis aspect of the original question. You theorize that the gummy material is no-clean flux residue, and possibly a process indicator. |
But shouldn't it be inert? If the natural evaporation process of the interior flux compounds was inhibited by a perimeter barrier of flux, would that render the resulting residue active? Does the high voltage increase the sensitivity of the circuit to current leakage?
Alan C.,PSI, USA