Surendar Magar still wonders about what to call the business he started in 2008. The chief executive of biosensor start-up HMicro, he said in a recent interview that while the company is trying to change how hospitals gather data from patients, he still does what he has always done: make chips.
Magar’s biography is dyed in semiconductors. After finishing a doctorate in digital signal processing, he joined Texas Instruments and helped develop the TMS320 microprocessor released in 1983. He later started the Wi-Fi chipmaker Athena Semiconductors, which he sold to Broadcom in 2005. HMicro has also devised its own unique chips, but it is difficult to leave healthcare out of the equation.
“It’s probably a hybrid,” Magar said about the business. The company’s first product, a wireless chip that measures a patient’s vital signs and relays that information to a smartphone or dongle plugged into a patient monitor. HMicro will not supply chips directly to hospitals, he said, but to equipment makers that can build wearable sensors around them.
The Fremont, Calif. chipmaker has raised around $30 million in funding since 2008. Last month, it revealed its first chip, HC1100, which has already entered “high-volume production,” the company said in a statement. But it has not yet announced any OEM partners.
Lying at the heart of its first product is the company’s unique chip design, which it started laying out with STMicroelectronics in 2011. The WiPoint architecture includes an array of sensors for monitoring heart rate and respiration, and three different radios so that the chip is never interrupted while streaming data. Squeezed onto the same chip are the power management, processor, and circuits for connecting motion and audio sensors.
The goal of the WiPoint chip is to replace the billions of sensors that hospitals purchase every year, so that doctors can constantly monitor patients without wiring them down to hospital beds. Patients can leave the hospital wearing the device and simply throw it out once the coin-cell battery runs out of juice.
Magar suggested that part of the reason that hospitals have been slow to employ wireless monitoring is the lack of hardware. The market for consumer fitness trackers from companies like Fitbit and Jawbone has vacuumed attention from medical-grade devices that monitor health and medical conditions, he said.
“In general, the medical world is in the stone ages,” Magar said. “And if you really want semiconductors to revolutionize something, you need a very high volume application.”
As the dimensions and cost of semiconductors have shrunk, other companies have taken that view. Vital Connect, which was founded by former Marvell Semiconductor engineers, created a custom chip for wireless patches that monitor vital signs. MC10, a healthcare startup based in Boston, makes wireless biosensors for medical research that stretch and conform to the human body.
General Electric and Phillips – two of the largest makers patient monitors in hospitals – are also testing patches to remotely analyze sweat and check vitals. In 2012, they were among the most vocal supporters of a Federal Communications Commission ruling that opened 40 MHz of spectrum for medical body area networks, or MBANs.
And government regulators have not turned a blind eye. In 2013, the Food and Drug Administration passed guidelines for wireless hospital devices, advising companies to build secure wireless products that can’t be cross-contaminated with Wi-Fi networks used by consumers.
Magar was guided by similar concerns: “In an emergency room, there may be 50 patients wearing the patches,” he said. “They are continually transmitting to a patient monitor, so how do you connect 50 channels in a room without missing a beat?”
Hmicro’s engineers were able to merge three different radios into a single WiPoint chip, so that the device always had an open passage for transmitting data. It operates primarily over Wi-Fi, but uses the medical band and ultra-wideband channels as backups.
The chipmaker also created an optimized version of Wi-Fi that consumes fewer than 10 milliamps of power from the coin-cell battery. The company decided against using Bluetooth – like Vital Connect, for instance – because it was not efficient enough for constantly streaming data.
One lingering question for the company is how the price will compare to wired sensors. Doug Linquest, HMicro’s chief marketing officer, said in an interview that the chips would be slightly more expensive, but that they would become cost-competitive after increasing production. He added that the chips could further cut costs related to disinfecting reusable wires.
Magar suggested that the benefits of wireless monitoring would outweigh pricing concerns. And he fell back on a familiar chip industry metaphor to explain why: “It’s a little like what Wi-Fi did for Ethernet cables. It got rid of the wires and we were able to create an entirely new infrastructure.”
Correction November 3rd, 2016: An earlier version of this article misstated the power consumption of the chip's Wi-Fi radio. It draws fewer than 10 milliamps of power from the coin-cell battery, not fewer than 200 milliamps.