Antennas are difficult
One of the most difficult problems with designing a wireless solution is antenna design. The math involved is just the start. The equipment needed to measure both on-board signals as well as off-board radiation is prohibitively expensive. Lastly, the cost of iterating is expensive. Even with programmatic monte carlo simulations of antennas, they still need to be manufactured and measured in real-world conditions and then refined.If you’re a small design team just starting out, you literally can’t afford to design and test your own antenna. Luckily, there are plenty of people who specialize in antenna design and have created a number of solutions based on the frequency range and application.Short range protocols like Wifi, Zigbee, and Bluetooth have very low cost antennas because of the limited frequency range and the short distance the waves need to travel to work. Longer range signals like Cellular become more complicated because a more powerful signal is needed for the signal to be meaningful at a useful distance. The design task is complicated greatly by the various frequency bands and regional regulation.
Antenna Form Factors
Antennas come in several different form-factors, depending on your physical design, range, and cost needs.
SMD Chip antennas are the easiest to use, but can often suffer from either being too large, or not transmitting far enough. They are a breeze, though, because they come on reels and can be handled by a pick and place machine. (Image from Taoglas)
Stamped/folded metal is extremely inexpensive, as the tooling isn’t complicated. The designs, however, tend to be bulky. This combined with stamped designs being more susceptible to vibrations limit them to larger, stationary devices such as a router or base station. I’ve also had a high number of failures due to shipping vibrations, so it might take your manufacturer a while to lock down the fabrication process. Unless a fad for oversized electronics emerges, I don’t see these being viable on wearables or solutions that go in your pocket. (Image from Taoglas)
My personal favorite are flexible whip PCBs. While not as cheap as stamped metal, they are relatively flat. If your main PCB is very small, or enclosed in such a way as to block out the intentional radiation, a flexible PCB antenna can move the signal to a better place on the board, as well as the potential to direct it to a different plane than the circuit. Up until a few years ago I would have recommended an SMA connector in order to fit with a majority of lab equipment, but with the expectation of increasingly smaller gadgets, a U.FL connector will shrink your design. The loss of signal from adding a U.FL to SMA connector can be accounted for and subtracted from the signal loss. (Image from Taoglas)
The whip is useful to minimize problems with mechanical tolerance, but for the very smallest solutions, an SMD spring contact will move the signal to the antenna for pennies. This requires the most experimentation because the antenna must deal with interference from the PCB itself. (Image from https://www.ifixit.com/Teardown/Nexus+7+Teardown/9623)
The fastest and cheapest route to a good design is to partner with a testing facility. They will already have experience with your wavelengths, and can connect you with several different antenna vendors and designs. Additionally, they will be able to make recommendations to your PCB design. If you are already working with a contract manufacturer, they may also have some good recommendations, or have a partnership with a testing lab.It may be, however, that you would like to do some low-cost prototyping before sinking thousands of dollars into certification and lab fees. It’s good to know what equipment is out there.There are several big names in antenna design. Taoglass is the 800 pound gorilla in the room. In addition to custom designs, they offer a very wide variety of antennas from surface mount chip antennas, to flexible pcb whip antennas, or stamped/folded metal antennas. Taoglass offers antenna design / certification help. Just like their antennas, their service tend to be on the high side cost-wise, but paying a premium gives you a performance premium. Think about that when you’re very far away from the cell tower and you’re still able to get a signal.Sinbon is a great competitor to Taoglas. They have offered on several designs to perform preliminary tooling and testing on-spec and have drastically lower prices than Taoglas. Additionally, Sinbon offers manufacturing and design solutions.
Antenova have fewer solutions, but offer very inexpensive off the shelf solutions, so they’re worth checking out.
Molex also has an antenna division. If you’re already doing a lot of business with Molex, I recommend getting a quote from them, as you might get some significant cost savings. Their other specialty is molding plastic and they offer integrated solutions for plastic + antenna. The designs I’ve seen from Molex offer up complex geometries to satisfy any industrial designer.
Antenna design is complicated. I find it good to start with an off the shelf solution to and after the development looks promising, to hand the files off to a testing facility to help take you across the finish line.If you’re looking for some places to start, I would check these out; http://www.taoglas.com/antennas/GSM-CDMA-Cellular/Internal_Flexible_PCB_antennas_-_FXP_Series/http://www.molex.com/molex/products/listview.jsp?query=105263&path=cHome%23%23-1%23%23-1~~q313035323633~~ncANTENNAS%23%230%23%233&channel=Productshttp://www.antenova-m2m.com/products/giganova-standard-antennasIf there is vendor or service that you’d like to praise, We’d like to hear about it. The best thing about being on an open hardware design is being able to work in public and get the best advice available.