If you want to catch Riot’s attention, mention wireless charging and IoT devices in the same breath and you’ll have done the job. To this end, Atmosic has caught our eye in the past few months, emerging from seemingly nowhere to launch two new SoCs that promise to provide huge performance leaps for IoT developers to harness.
We spoke to Srinivas Pattamatta, VP Business Development, for a proper introduction to the company, who explained that the company was established in 2016, with all the founders stemming from Atheros – the chip design house that Qualcomm bought for $3.6bn in 2011 that specialized in wireless chipsets, ‘back when a billion dollars meant something’ quipped Pattamatta.
These founders, Pattamatta said, have been in the low-power wireless game for twenty years, and are now guided by a strong advisory board that houses Alphabet chairman John Hennessy and Atheros founder Teresa Mang. Having just completed a second round of funding, Atmosic has raised close to $50mn to further its ambitions. Pattamatta said that the funding rounds were both oversubscribed, and the company’s 50 or so staff are now focusing on growing Atmosic’s revenues.
Atmosic makes Bluetooth 5.0 chips and sells them to companies that want to get more out of a battery. Pattamatta explained that there were three core areas. The first is focused on very low power devices that claims to be able to achieve five-times lower active power consumption than rival implementations, on an apples-to-apples basis, and will extend the battery life by three to five times. Pattamatta said that Atmosic was sampling these chips in October 2018, and now has several customers, and is shipping in low volumes.
The second core thrust are chips that provide ‘on-demand receive’ capabilities. Pattamatta explained that this is a second auxiliary receiver that acts like a pair of ears that never sleep, allowing the main device hardware to properly sleep. This receiver is listening for a specially coded Bluetooth packet, which will wake the device. Atmosic claims that this can provide anywhere between 10 and 100 times the power savings, compared to what its rivals can manage.
The third element is what initially caught Riot’s attention – managed energy harvesting. Atmosic can harvest energy from RF, light, thermal, and motion. There are two thrusts here, with the first being to create completely battery-free devices, and the second being a ‘Forever Battery’ that will help to trickle charge a battery in a device.
Pattamatta said that in consumer applications, the user doesn’t typically mind changing a battery, but in business deployments, once you factor in the truck roll costs, the Forever Battery proposition becomes very attractive. Here, Pattamatta says that the technology will extend battery life by 7 or 8 years, meaning that most customers will be upgrading the application itself before they need to change the battery.
It was at this point that we asked about the general reputation of energy harvesting in the industry, pointing to Ossia’s gradually declining claims and the fact that the EnOcean has been plugging away at this for so long that a company like Atmosic had enough room in the market to emerge and launch commercial products.
Pattamatta agreed that energy harvesting has had a bad reputation in the past ten years, but pointed to Atheros’ experience in WiFi as a possible comparison. He noted that 12 years ago, Atheros struggled to achieve 54 Mbps in demonstrations that had devices just two feet apart. Now, WiFi can hit multi-gigabit speeds in commercial deployments, and to this end, Atmosic things RF energy harvesting is in the same early day difficult period that WiFi survived.
To this end, Atmosic is choosing to target devices that run on AA batteries, and is deliberately avoiding those at the higher end of the power consumption tables. Pattamatta said that the trick is not in the actual harvesting, because due to the laws of physics, a great amount of any emitted signal’s energy is lost during transit. Rather, the trick is in cutting the amount of energy needed during operations to the absolute lowest amount.
Pattamatta likened this to two marathon runners. Both are able to complete the marathon, but between races, the Ossia approach must consume vast amounts of calories to recuperate, while the Atmosic runner will get by on just a power bar, because it consumed less energy during the actual race.
To this end, the technology would allow for devices to provide new kinds of applications to their users. The example Pattamatta used was a coin-cell powered asset tag in a hospital, where the current capability might be a six-month battery life on one message per day – which is quite limited from an asset tracking perspective. The Atmosic pitch is that this would instead be a continuous operation with a two-hour latency, without the need to employ a full-time staffer to go around changing batteries.
As for customers, there are two main segments – consumer electronics, and industrial and enterprise. On the consumer side, the main current targets are remote controls, computer peripherals, door locks, and a few smart home applications like switches. Pattamatta said that he could not disclosed names currently, but there are a couple of live products in the US, a couple in Japan, and a few in China. On the industrial side, Bluetooth beacon applications are set to ship within the next six to nine months.
Pattamatta mentioned Atmosic’s recent partnership with Tonly, a division of Chinese electronics giant TCL, which is the number-one wireless module producer in China. Atmosic believes this partnership positions it well to attack the Chinese market, and it’s hard to fault the logic there.
We asked whether Atmosic could put its technologies to use in other wireless protocols, like Zigbee or Z-Wave, or whether there was some inherent benefit in the Bluetooth protocol that necessitated Atmosic’s focus. Pattamatta said that the bigger issue for chip vendors is the silicon design, and to this end, he said that the rampant M&A-driven consolidation in the semiconductor industry has slowed innovation.
However, Pattamatta thinks others will catch up with Atmosic’s performance, pointing to Nordic Semiconductor’s 9mA design that was then beaten by Dialog’s 4mA, but are now both some way behind Atmosic’s claims 0.9mA usage. As for the protocol side of things, Pattamatta noted that EnOcean is fantastic, but that it had focused on the protocol rather than its power usage, returning again to the marathon analogy.
Bluetooth does have slightly lower power consumption than the other protocols, Pattamatta said, but thinks that Atmosic could probably manage to get equivalent performance in Zigbee and Z-Wave if it put its mind to it. The current focus on Bluetooth 5.0 remains, thanks to its new support for long-range (for a PAN) communications, as well as the mesh networking features.