I figured I tore down one of the Wilocity WiGig antennas, may as well tear down one of the Intel ones.
So this antenna is a 2x10 array, its not on a ceramic board, but instead seems to be some form of fiberglass, It also has soldermask and silkscreen.
I had some difficulty getting the shield off and some components got bumped/toasted. They used some underfil on the main RF IC. This made it very hard to remove. The solder mask also makes it impossible to see the track from the X.fl connector. I suspect the interface is similar to the Wilocity Antenna. There are 2 baluns next to the chip, wonder if it takes in a differential signal from coax?
You can see some of the underfill was left behind. It also looks like the underfill pulled up the soldermask where it came up with the chip. I cant discern any kind of pattern on this chip that would indicate which balls the RF comes from.
The bottom of the chip is covered in the underfill so we can see any of the pretty silicon. If anyone has any advice on how to remove this stuff without hurting the silicon below let me know.
A bit chipped off the side and you can see the silicon underneath.
Above are an assortment of WiGig modules I've been able to find on the internet. There are Wilocity chips, as well as Intel chips. The top left only contains the Wilocity WiGig chip, It's from a Dell Wireless docking station. The top right is the corresponding WiGig module from a Dell Laptop.
The Bottom left and bottom right modules are different form factor Intel WiGig modules.
The center module is a smaller form factor M.2 E & A keyed Wilocity WiGig module from a Lenovo think pad.
Driver support for Linux seems to exist for the Wilocity chips but not the Intel. I have yet to actually test any of the modules.
None of these modules produce the 60 GHz directly. I am not sure what the output signal is exactly, but it goes to an ceramic antenna module that is a combined 60 GHz Up/Down converter and 16 Element Beam Forming Antenna.
The RF connectors are not U.fl or even H.fl. They are teeny tiny X.fl connectors.
You can't buy the Wilocity antennas directly, I've had to harvest them from Dell Latitude laptop screen backs, and wireless docks. I have however found the Intel antennas available directly on Newegg. They even came with X.fl terminated coax.
I accidentally broke one of the antennas in half when I was removing it from the wireless docking station which means you get to see some awesome RF Voodo.
In this photo the shield is removed. You can see where I cracked it about 1/3rd of the image from the right. Just above the X.lf connector there is some wonderful microwave vodoo. The middle IC does all the 60GHz beam forming magic. Looking at the previous photo you can see the 2x8 element array. Seems like most of the steering happens horizontally from the way the antenna was mounted on the screen panel and the dock.
You can see some interesting details around the footprint for the Magic 60 GHz die. The trace leading to it is surrounded by via stitched ground. Its clear they are using via in pad from the lack of any top layer traces on this footprint. Also you can see the sides of this chip have a ground guard trace. Then there are the vias on the sides that do not correspond to balls on the IC. There are 8 of them on each side, so I'm guessing there is one for each of the elements. There are 17 balls on each of the left right sides of the IC. It seems like there is a pair for each antenna element, and one extra ball.
Here is a shot of the Die itself. Its a WLCSP. You can see the sides contain roughly identical drive sections for each of the 16 antenna elements. It looks like the two balls for each element might be directly connected rather than driven differential. I'm wondering if they are just x-bit phase shifters. Best guess for the architecture is the baseband signal comes in at some IF, then its upconverted with the 60GHz LO and sent to each of the shifters. Incoming RF is similarly shifted on the way in, down converted to some IF and sent back out. Also I suspect the decisions on how to steer the antenna are done on the digital side, so some kind of circuitry to pick up antenna steering commands and set the phase shifts.
And here is a close up of the RF Input side. I'm guessing the tap off from the incoming signal is for power supply.
It implies that the interface between the antenna and the module carries the LO/IF/Control and power. Also it states the antenna gain is 12dBi, but no information on how steerable it is and to what steering directions that gain applies to.