Checking connections and for bridges on SMT devices on adapter boards.

[JoeN]

For soldering fine pitch SMT devices such as QFPs and QFNs to SMT adapter boards, which I use quite extensively, I have developed the following approach and I wanted to ask if this makes sense, and also if it is safe.

It seems to me that the two big faults you can run into are pins that seem soldered but are actually detached and bridged pins.   I look at the pins under a microscope, but I have missed very fine bridges before even under the scope, usually ones that are riding high at the top of the pins.  And the microscope is not nearly as useful for QFN devices anyway.  So I have adopted this method in addition to visual inspection.

To check for bridged pins, after soldering the device all I do is check pin x against pin x+1 for each "side" of the device with a continuity tester, taking into account sometimes there are banks of pins for ground or VCC that are bridged on purpose.  If a pin is bridged with it's neighbor the continuity tester sees this.  A pin cannot be easily bridged with any other pin other than its neighbor so I only have to go down each side once for adjoining pins.

To check for detached pins I have found that on pretty much every IC I have ever seen, every pin has some resistance to the ground pin.  The resistance is never infinite.  So I put it on the resistance measurement and check from each pin to ground.  An open circuit indicates the pin needs to be touched up.

So obviously for both these tests I am putting a bit of current through the IC.  It's being done with a very modern multimeter - Agilent 34410A.  Is that universally safe to do?  Is there a better way to test this out? 

And I know I get a benefit here using the adapter board, it has built-in test points that a normal PCB doesn't have.  I like taking advantage of that. 

By adapter, I mean adapters like this one, though not always this many pins:

[T3sl4co1l]

ICs with protection diodes will almost always tolerate 10mA into/out of the pins, just fine (i.e., diode-checking with respect to GND or VCC, whatever VCC domain that might happen to be).  A resistance measurement (on the lowest ohms range) or diode check is usually fractional mA, and voltage limited to 2V or so.

Tim

[JoeN]

ICs with protection diodes will almost always tolerate 10mA into/out of the pins, just fine (i.e., diode-checking with respect to GND or VCC, whatever VCC domain that might happen to be).  A resistance measurement (on the lowest ohms range) or diode check is usually fractional mA, and voltage limited to 2V or so.

Tim

Lowest ohms range?  What is that?      My multimeter decides that for me.  Maybe I should grab the spec sheet and see if it has a value for most current used in a resistance measurement...