Page 83 - Building A Succesful Board-Test Strategy
P. 83

Test Methods  69

 ov ~1                                             Brake Input
 T-5V



 +10V
  ov                                               One of Three Inputs
                                                   (i.e. PH_A)
          400ns {*—
 10V

  6V                                               One of Three Outputs
                                                   (i.e. OUT_A)
 1.5V

  OV                                               AFTM Voltage Measurement
           t    t    t    t    t    t              at Each Arrow

 Figure 2-12  An in-circuit test for the component in Figure 2-11. Testing each output
 requires six analog voltage measurements. (Durickas, Daniel A. 1992. "228X AFTM
 Applications," GenRad, Concord, Massachusetts.)


    Vacuum or mechanical force pulls the board under test down onto the fixture,
 pushing the probes out through the platen to make contact. Spring-probe com-
 pression exerts about 4 to 8 oz. of force at the test point, ensuring a clean electri-
 cal connection.
    Some fixtures forego the "rat's nest" of wires from receiver to probes in favor
 of a printed-circuit board. This choice improves speed and reliability while mini-
 mizing crosstalk and other types of interference. Such "wireless fixtures" are less
 flexible than their more traditional counterparts. In addition, lead times to manu-
 facture the circuit board may make delivering the fixture in time to meet early pro-
 duction targets more difficult.
    There are four basic types of bed-of-nails fixtures. A conventional, or dedi-
 cated, fixture generally permits testing only one board type, although a large fixture
 sometimes accommodates several related small boards.
    A universal fixture includes enough pins on a standard grid to accommodate
 an entire family of boards. The specific board under test rests on a "personality
 plate" that masks some pins and passes others. Universal fixtures generally cost
 two to three times as much as their conventional counterparts. Pins also require
 more frequent replacement because although testing a particular board may
 involve only some pins, all pins are being compressed. Therefore, if a pin permits
 1,000,000 repetitions, that number must include occasions where the pin merely
 pushes against the bottom of the personality plate. The high pin population on
 universal fixtures makes troubleshooting faulty pins and other maintenance much
 more difficult than in the conventional type.
    Surface-mount fixtures also attach to a standard 100-mil receiver grid. In this
 case, however, a series of translator pins pass through a center fixture platen so
   78   79   80   81   82   83   84   85   86   87   88