The pogo alternative
For many years the most effective test socket contact design for BGA devices was the "pogo pin", a pin design that has been in use since the 1950s when it was invented by Everett Charles Technologies. At that time, its primary use was to test PWBs. Over the years the basic pogo design was adapted to accommodate the testing of BGA devices.
First there was a "four-piece" design, followed by a design that reduced the pieces required to three: (Figures 1 and 2). Both designs provide long contact life (typically 500,000 cycles) and a relatively short electrical path of 5.72mm (0.225"). They can be configured to cover a variety of sizes and pitches as well as a range of compliances, and are sold individually.
These designs have several drawbacks. They are expensive, they require a lot of set-up time, and they typically cannot be adapted for applications where the device pitch is less than 0.4mm.
Figure 1. (left): Four-piece with press fit; Figure 2. (right): Three-piece with spacing
The four-piece design (Figure 3) consists of two moving probe tips, one spring, and one tube (barrel) and is usually longer and narrower than the three-piece version. When incorporated into a socket design, the pins of the four-piece are pressed into a receptacle or captured between two counter-drills, to a given height. Both the top and bottom tips protrude, making contact with the PWB and the device under test.
The three-piece design (Figure 4) consists of a top tip, which also acts as the tube or barrel of the probe; a moving bottom tip and a spring. It can be made much shorter than the four-piece, but is usually somewhat wider. The middle of the barrel of the three-piece contains an annular ring that keeps the probe in the receptacle. This ring is captured between two opposing counter-drills, so it floats inside the receptacle.
However, this ring precludes the use of the three-piece design in tight pitch applications. The way around this dilemma is to use a very small, screw-machined barrel, a product which is difficult to manufacture. It is also expensive to plate, due to the small, deep hole through the center of the barrel. Probes with these small machined barrels can cost as much as $6.00 each.
For test, the bottom probe tip is preloaded when the socket is mounted to the PWB, because the angular ring is pushed against the top counter-drill. The upper tip is engaged when the DUT is inserted for test. The current flow is through the barrel and out the probe bottom.
Figure 3. (left): Typical four-piece design shows two moving probe tips, one spring, and one tube (barrel); Figure 4. (right): Standard three-piece design: top tip acts as tube, moving bottom tip and a spring.
Alternative solution – probe and spring
IC devices today frequently have pitches as small as 0.20mm. While many are still terminated in BGA balls, many are now just wafers, leadless, or formed lead devices. These devices still need to be tested, so the demand for reliable, cost-effective test sockets for small-pitch devices has grown. The traditional pogo pin designs have proven incapable of providing a test socket solution.
A new contact pin approach designated “Probe and Spring” addresses the disadvantages of the pogo pin designs by providing lower cost, one consistent height, and shorter set-up times.
The patented “Probe and Spring” contact has only two components (Figure 5). The probes and springs are contained in an “interposer set” that is machined to match the device footprint. This provides equally good life of 500,000 cycles and up, while significantly reducing the electrical path to 1.95mm (0.077").
Fewer components result in better cost economies, and pitches as small as 0.2mm can be accommodated (Figure 5). The new design also offers both “crown shaped probes” (for BGA devices) and “sharp tip probes” (for leadless or formed leaded applications).
The spring protrudes from the interposer set bottom, while the probe protrudes from the top. The interposer set is then contained in a molded housing. There are five different sized molded housings, and the size of the device determines which housing is used.
Figure 5. Probe and spring options
Here is how it works
The “interposer set” consists of a top interposer and a bottom interposer. Both interposers are drilled to the exact footprint of the customer’s device.
When the Aries socket is manufactured, the top interposer is turned upside down. A probe is placed into each of the holes. Each hole is “tapered”. The probe fits in the hole, but it does not fall out because the bottom of the hole is smaller than the “shoulder” of the probe.
Next, a spring is placed over the bottom of each of the probes. Next, the “bottom interposer”(turned upside down) is positioned over the top of all the springs. The holes in the bottom interposer are tapered. The bottom of each spring projects out of the interposer hole. The spring does not fall out of the bottom interposer because the bottom of the hole is smaller than the diameter of the wide middle of the tapered spring.
Once this assembly process is complete, an “interposer sandwich” has been created. The completed interposer sandwich is turned over, so it is right-side up (with probes projecting from the top and springs projecting from the bottom), and is positioned into the appropriate “standardized housing”. (NOTE: the “sandwich” interposer set is replaceable, in the event the customer exceeds 500,000 cycles and wears out the original interposer set.)
Finally, a “pressure pad” is positioned into the lid of the “standardized housing”. The pressure pad can be very thick, or very thin, or in between. The thickness depends on the thickness of the device that is being tested (if the device is thin, the pressure pad is thick; if the device is thick, the pressure pad is thin – regardless, the thickness of the pressure pad is designed to insure that every device lead mates properly with its corresponding probe).
Once the completed socket is bolted to the test board, the bolting process causes the springs coming out of the bottom of the bottom interposer to mate with the pads of the customer’s test board via “pressure mount”. The customer places the device into the Aries test socket and closes the lid. The closing of the lid causes each device lead to mate with the top of its corresponding probe.
The closing of the lid also causes each spring to “tilt”, resulting in a very short signal path, since the signal path is from the device lead to the probe head to the (tilted) spring to the test board pad.
The spring design also balances the "force conundrum"-providing enough force to break through any solder oxide that may exist, while holding the lid-closing force to a minimum. The higher the pin count, the more important this trade-off becomes, since higher pin counts mean more probes, which equates to more force required to close the lid. Different wire diameters used in the same spring design are an important element of resolving this problem.
The same spring design is offered in several diameters, and the same pin design is available in five sizes; thus, one height is used for all sockets, resulting in a socket design that can be "standardized" at the same working height. This consideration minimizes handler set-up time and change-out kit design time.
The tradeoffs to the probe and spring design are that individual contacts are not replaceable, and a minimum pad size of 0.23mm (0.009") is required. If one probe is damaged, the entire probe receptacle can be replaced by the customer in minutes by inserting a replacement interposer set, or refurbished by the factory.
Regardless of these limitations, this new technology reduces downtime, since the entire interposer set is easily replaced.
Figure 6. Current path for probe and spring
Why it works
Coiled springs bow when they are compressed, unless they are restrained by a captive barrel. The Probe and Spring contact actually takes advantage of this bowing and uses it to reduce the electrical path (Figure 7).
Within the new design, the current path is through the probe head and out the bottom of the spring - avoiding the spring as the shortest current path, and eliminating the barrel and bottom plunger.
Another unconventional design feature of the Probe and Spring is a variation of coiled diameter within each individual spring. The opening of each individual spring decreases from the top of the spring to the bottom.
The wider top more readily accepts the probe, while the narrower bottom reduces to a diameter that is just larger than the bottom of the plunger. This "diameter reduction" is achieved while maintaining a closed coil design of the wire to form a "solid tube" when in the closed position - not unlike a barrel. The coils are not compressed to the solid height of the spring.
As the DUT makes contact with the top of the probe, the probe is inserted in a way that rubs against the narrow, closed coil portion of the spring, where it makes contact due to the bowing of the spring. The current path now flows through the plunger to the closed portion of the spring, or through the center of the contact.
User requirements, such as lower pin-to-pin pitch on BGA and leadless devices, have driven many of the advances in the test socket industry. Although ideal for many traditional applications, three- and four-piece pogo pins do not offer the ease of installation and maintenance, or the added cost reductions, of the new two-piece probe and spring contacts.