Low-g stimulus accuracy

Low-g accelerometers are typically stimulated in a range of -1…+1 g. In practice it means the DUT is turned in different orientation and Earth gravity is benefitted. In a way it sounds simple. Having a DUT on a level +1 g stimulus is applied and when upside down, -1 g.  Afore guarantees the total accuracy to be better than 0,2 degrees. However, there are several issues which have an impact to the stimulus accuracy.

First, it is a matter of the accuracy of  the test handler
Is upside down exactly upside down or just close to it?  At Afore, enhanced servo systems and encoders are used for reaching high turning accuracy.

Second is the the alignment of the DUT
How to guarantee the sensor axis are parallel with handler axis ?  The common way for us is to perform the testing in strip level or with carriers. Even a very small clearance in a socket can cause a major angular error of DUT. Naturally the smaller the DUT is the worse situation becomes. In reality, the socket cannot be very tight, if the handling is done with P&P equipment. The problem can be solved by using aligning sockets, which remove the clearance when closing, but then there is a danger an extra force is induced to the DUT and that may have a harmful effect to the MEMS structure. It is definitely better to let the DUT be in a “free state”. The use of strips or precision carriers allows positioning clearance without big impact to the angular accuracy. The picture below shows the benefit compared to testing of DUT in a socket.

 


Figure. An example how 0,05 clearance can affect to the angular positioning, single DUT vs. Strip.


In case of WLCSP, the wafer is placed on the tape and diced on it. Dicing has only a minor effect to angular position. When aligning the whole wafer to the handler axis, the angular accuracy of single DUT is excellent.

Third issue concerning stimulus accuracy is noise.
The result of good turning accuracy and precise alignment can be destroyed if the handler produces mechanical or electric noise. A sturdy structure is a good starting point. A stiff, low deformation structure, a typical feature to Afore’s handlers, dampens vibrations well.
Electric noise is minimized by good design, including choice and routing of measuring and handler cables.

  Click here for more information