How Spark Microsystems Makes Short-Range Wireless Deterministic

Your product can have a world-class cloud stack and a blazing-fast 5G link, then lose the whole experience in the last half meter. That’s the “last meter” problem, and it’s why we sat down with Dr. Frederic Nabki Co-founder and Chief Technology Officer Spark Microsystems, to talk about ultra-wideband wireless that targets wirelike responsiveness instead of “good enough” latency.

We dig into where Bluetooth and Wi‑Fi still shine and where they hit real limitations for deterministic wireless, ultra-low latency, and interference-heavy environments. Frederick explains why Spark’s approach uses impulse radio UWB, how sub-nanosecond-scale pulses change the game for multipath and coexistence, and how wide UWB spectrum enables frequency agility when the airwaves get crowded. If you’ve ever been in a trade show hall where microphones and earbuds fall apart, you’ll recognize why interference robustness is no longer optional for industrial IoT, medical devices, wearables, and robotics.

The examples get concrete: a gaming mouse that targets about 150 microseconds end-to-end latency, robots that need fast control loops to avoid collisions, and brain-computer interface systems where cables create infection risk and power budgets are unforgiving. We also cover Spark’s go-to-market details, including transceiver silicon, an SDK, reference designs, antenna guidance for FR4 PCBs, and why modules can simplify certification.

If you care about ultra-wideband, UWB data communication, ultra-low power wireless, and real-time connectivity, hit play, then subscribe, share the episode, and leave a review so more builders can find it.

Transcript

Why Ultra-Low Latency Matters

SPEAKER_00 0:02

Hey everybody, really excited for a conversation today about pushing the boundary of wireless far beyond the usual Bluetooth and Wi-Fi conversation for ultra-low latency and ultra-low power performance and industrial IoT and related applications today with Spark Microsystems. Frederick, how are you? Great. Thanks for having me. Well, thanks for being here. I've been in the wireless space a long time. Was involved in the first uh Bluetooth single chip radio uh 30 plus years ago. But we've come a long way since then. So I'm really excited for an update on the bleeding edge here.

Meet Spark Microsystems And The Mission

SPEAKER_00 0:42

Before that, maybe introduce yourself. And for those hearing about Spark for the first time, how do you describe the company and the mission?

SPEAKER_01 0:50

Sure. So I'm the co-founder and CTO of Spark Microsystems. My background is radio frequency integrated circuits. I started right about when people are starting to port Wi-Fi chipsets to CMOS, right? And that was a big inflection point. And the company we founded uh you know almost a decade ago now, but not quite yet, is uh is essentially a company that tries to revisit short range connectivity to provide new performance metrics to the table when engineers want to design wireless uh connectivity for their peripherals for any type of short range type application. So in which way, and I'll I'll quickly enumerate it, very low latency, sub-millisecond latency, very low power consumption. So, you know, single-digit uh two, one, two nano joule per bit for the data transport and receive, very good throughput, up to 20 megabit per second payload, and also interference robustness so that we can coexist with Wi-Fi and Bluetooth and not be you know um degraded by their presence because we need to coexist in this wireless world, as you know.

SPEAKER_00 1:57

Indeed. Yeah, and of course, Bluetooth and Wi-Fi, huge successes, and they continue to evolve with the special interest groups and standard bodies and uh amazing progress.

Where Bluetooth And Wi-Fi Hit Limits

SPEAKER_00 2:10

Where do they work well today? And where do they start hitting you know limitations?

SPEAKER_01 2:16

Well, look, Bluetooth and Wi-Fi, no question, great success stories, right? You mentioned it 30 years and going. I don't see them going away anytime soon. They are doing great. Uh and I've been evolving. Of course, the pace of evolution in these standards are a little bit more incremental nowadays than they were perhaps uh 25 years ago. Um, and there's new classes of devices, I'd say, that are pushing the limits of what Bluetooth and Wi-Fi can achieve, right? They're both addressing, I'd say, separate markets, right? Wi-Fi is more performance driven, Bluetooth is more power-driven, and of course they're trying to they're coalescing as we see the protocols evolving both sides. In our case, you know, what we're trying to target is applications where most of the time a wire would have done the job and they just have a hard time cutting that wire. Why why is that? Well, they want deterministic wireless transport. What does that mean? It means that packets come in uh at a very well-known time interval. Uh, the data flow is very constant, there's not a lot of uh undulation on the data rate. Of course, Wi-Fi, we know, can be quite uh uh difficult with that. Uh they want a system that coin exists because they have multiple radios in their device and they want all the radios to be leveraged fully, right? More and more. There's more bandwidth requirements, there's more uh latency requirements, so delay of communication is really becoming an issue for all of the applications that are piped into your sensors, right? Nowadays, people are talking, smart glasses, uh, you know, all these interactive uh haptics, things where your senses are meant to be fully in tune with the device, right? So that it reacts to you. And we all know our senses are pretty sensitive to latency, right? Usually uh 10 milliseconds, you're already at the edge of what people feel are acceptable. So that's where I feel, and Spark feels that Bluetooth and Wi-Fi might be a little bit more limited in what they can do. They weren't built for this initially. Of course, they're doing a great job trying to push the technologies further to try to address that, but that's where ultra-wide band technology comes in, where it's it's homed in for this from day one. Uh, at least our technology was built for that purpose. So obviously, when you build something from scratch to do a particular set of capabilities, which I listed earlier, you're always going to end up with something that is much more optimized for those kind of use cases, which are starting to appear in a lot of different markets.

SPEAKER_00 4:42

Indeed, super interesting

Ultra-Wideband Beyond Location Tracking

SPEAKER_00 4:44

time. And most people like myself, you know, we think of ultra-wideband as like location tracking or positioning technology, but how is your approach, your technology, different?

SPEAKER_01 4:56

Well, if you remember Bluetooth, since you you've been there from the very beginning, uh the good old Bluetooth headset was kind of its beginning, right? And everybody associated Bluetooth to that ear-worn device. I don't think nowadays you'll see a lot of people with that kind of device uh with Bluetooth, right? They're using Bluetooth for everything but. So Ultra Ride Band, you know, was the first wireless radio to be incorporated within a smartphone in a long time, right? It's been cellular, Wi-Fi, Bluetooth for 20 years. Then we all know what happened uh a few years ago where the phone makers started to deploy this technology. And the first use case was positioning. Great use case, and we we we we all know that success. So um our technology really tries to take ultra-ide band and widen its applicability to just more than ranging, right? No wireless technology should be a one-trip pony. You're not gonna have the lasting power in the market and the ecosystem just because of one capability. So, what Spark brought to the table is how do you use the ultra-right band spectrum and create a radio that's tailored towards data communication, but at super low latencies, very high data rates comparatively to other low-power radios, and as a as a at a power profile that's extremely low for that kind of capability, while enabling the use of other radios at the same time, while taking care of Wi-Fi, being able to communicate at the same time, while having high concurrency capability with different radios, right? So the the ultra band spectrum is a spectrum, but it can be used for more than just ranging. That's kind of where Spark is going with its technology to try to push that use case beyond just ranging. And I I feel it's similar to history about with the first company that decided to go from the head, head, you know, head head headpiece to to to God knows what was the first uh you know Bluetooth application beyond that that that uh hands-free headset, right?

SPEAKER_00 6:46

Yeah, no, it's been quite a journey. And uh RF design is is kind of a black art, a black magic, uh, it seems. And uh lots of trade-offs always from a design standpoint, you know, latency versus power consumption versus throughput versus cost, and on and on.

Impulse Radio And The Spark Gap

SPEAKER_00 7:05

How do you think about those trade-offs with your technology one against the other?

SPEAKER_01 7:10

Well, you know, it's interesting that you mentioned that because you know, Bluetooth Wi-Fi are narrowband radios. They're using a modulated carrier, Bluetooth is starting to go up with more complex modulation as it's trying to get closer to where Wi-Fi can go, and Wi-Fi is trying to come down to more power efficient modulations and modes, and they're kind of gonna meet somewhere in the middle. And I'm not sure what happens when they meet. We'll see. Uh, for sure, uh it's gonna be an interesting thing to watch. Um, and uh, I'd say spark technology is very different, it's an impulse radio. So impulse radios don't use per se carriers that are you know amplitude modulated and so forth, they use electromagnetic impulses, much like Marconi did, you know, 100 years ago with gaps. Legend. Spark gaps. Obviously, a spark gap. Now you're getting a hint as to why the company is called Spark Microsystems. Ah, I've got it. The Spark Gap radio was technically an impulse radio. Obviously, it wasn't very uh effective at long range and channelization was a problem, and that's why, you know, not much after uh the big ships back in Titanic years, you know, we moved on to good old FMAM, and we all know where we are today. So, but because it's an impulse radio, though, it does have inherent uh capabilities when it comes to multi-pad robustness, uh, fair very low latency because the impulse in our radio is on the order of two nanoseconds long in duration. So when that um that elemental symbol, right, RF symbol is so different than an AM modulated or FM modulated or quam modulated sine waves, there's a lot of stuff you can do, not just with regards to the robustness of the connectivity and as a channel, but also regards to the uh electronics that you built to create and receive that waveform. And that's where the company uh pushed hard to innovate, to really take the radio, strip it out, start from a blank page, say if we had to build the best impulse radio to do what I just mentioned earlier, how would it look like, right? From the software to drive it all the way to the transistor level. And that's where we ended up with the products we have today.

Wirelike Responsiveness In Real Products

SPEAKER_00 9:17

Amazing. And you talk about actual wire-like responsiveness. That's intriguing. What does that look like for the user in a real-world application? I imagine a lot of mission-critical applications too, like medical devices, robotics, automotive, real-time responsiveness. Uh you can't get that wrong.

SPEAKER_01 9:36

Yeah, I'll give you a very mundane application just to start with something. Everybody's using a good old wireless mouse, right? A wireless mouse. Oh, okay, it's simple. Uh you know, it's a handheld device, it's a pretty high performance sensor, needs to be very accurate. You're not going to probably tolerate a lot of data packet losses, right? For obvious reasons. Uh, and the gamers are pushing those mice to have extremely low latency because it's a it's one of those, you know, uh esports, and uh every millisecond, every microsecond counts for those guys. So even though the application is quite mundane, right? It's a mouse, it turns out that uh the the best class of technology to do a gaming mouse is a USB wire. That's the one you're gonna get about I think 70 microseconds of latency overall into and USB to USB, right? Uh the issue is you have a wire dangling from your mouse, and these gamers they really don't want anything to dangle from the from the mouse. Uh so wireless technology came in quickly into wireless mice, but people realized quickly that the latencies were more in the order of you know a millisecond, five milliseconds, uh millisecond scale. Uh whereas Spark today we offer a wireless solution, uh, you know, a reference design to use our chipsets and gaming mice that gives you 150 microsecond latency.

unknown 10:54

Right.

SPEAKER_01 10:54

End to end, USB to USB, right? From the from the electronic signal that says I clicked a button all the way to the USB triggering on the USB port of the PC, 140 microseconds, which is uh, you know, we we brand it wirelike because the next best thing is a real wire. And as I told you, it's about 70 microseconds. But of course, there's no wireless there. So that's a very mundane example. Other examples include one that I would say is is uh is a bit further in the future, it's brain computer interfaces. Um, you know, we we've had customers come to us because they want to offload a lot of data coming from their probes, okay, because they do, frankly, you know, incredible things with that data, right? To to to to render uh uh some you know brain diseases obsolete. But anyway, uh they have a lot of data, they have extremely stringent uh latency requirements because obviously it's these are are are using your thoughts, right, to to react to, and you can't uh afford latency there. But then you have power requirements that are extremely rigid since it's in, you know, it's literally in your cranium. So what do you do? Well, you you drill a hole and you put a cable sticking out of the head, right? Nobody wants that infection risk and so on. And this is a more niche. I'm just going from the most mundane to the more, you know, and niche, niche, but important.

SPEAKER_00 12:12

How important is a life for a patient in that case?

SPEAKER_01 12:15

Well, with this technology now, they can replace that cable. That's an infection risk. That of course nobody wants something sticking out of their head. Nobody needs to, I don't need to explain to you why. So these are two extremes. But then you've got the middle ground where you have robotics, robots that need to move at high speed and not collide. So obviously, the more latency you have in the control of those robots, the the wider away from each other you're gonna have to have them so they don't collide. So you have less dense systems, you can be less efficient. And obviously, the best case scenario would be a wire, but wiring up a hundred robots in some kind of a warehouse setting is not really an option, right? That wouldn't make any sense. So that's another kind of middle ground application. Um that gives you perspective uh with these three examples of where low latency, high throughput, low power can really help. And of course, the key of wirelike is interfere interference robust. I can give you everything I just said, but then it breaks down the moment you turn on a Wi-Fi router somewhere, then it's all for naught, right? Uh, and uh, that's what we strive to achieve uh high performance with robustness to interference.

Coexistence In Crowded Wireless Spaces

SPEAKER_00 13:22

Well, let's talk about robustness uh to interference. Uh I was at like Pick a Show Him, CES this year, and virtually unusable. A lot of the mics I was using, the headphones, even AirPods. Yeah, you know, you think everything is seamless, so crowded, it was a nightmare. And I think everyone experienced that. And this is not just cell phones. I'm I'm talking private wireless, Bluetooth, Wi-Fi. It was just uh difficult. And of course, you know, low stakes, this is a trade show, but I imagine a hospital or a factory floor, uh, this this is becoming a real challenge. How do you think about coexistence now compared to only a couple years ago? And it's only gonna get worse, right, with new bands and new frequencies coming on stream.

SPEAKER_01 14:10

For sure. Wireless, you know, was always a big thing since I think Bluetooth and Wi-Fi came about, but it's just exploding right now, you know, and uh wearables are not gonna make it better, they're gonna make it worse. Uh, more and more people will have compute on different parts of their body, right? It's gonna mean you're gonna have maybe, you know, a dozen wireless radios on you at some point in time between your watch and your glass and so on. So obviously uh you need something that can handle that. So there's a few reasons why ultra band is is is great for that. First, the spectrum on ultra eight band uh in the US is from uh 3.1 to 10.6 gigahertz. You have a lot of bandwidth to do frequency agility, and that's one thing Spark prides itself at. We're using that spectrum within the radio to be extremely agile. You know, how many radios can can in real time change their channel uh at several gigahertz interrupting? That doesn't exist. Uh even the ranging ultra-head band radios don't leverage that capability, and that's fine. That wasn't their goal when they made the positioning stuff. Our goal was to leverage that spectrum to our advantage, to bring the most robust radio possible. And when you have that kind of bandwidth available, uh and you're a wireless guy, right? You you know what you can do if you have the radio be intelligent and agile to leverage that spectrum. Then there's the fact that this is an impulse. So the fact that it's an impulse, it's extremely narrow in time. There's a few things you can do, secret sauce things, to kind of detect it better than if it's next to a narrowband signal, which has very different energy signatures than an impulse, right? Uh fine wave impulse. You can, you're, you're, you're, you, you know, it's it's it's there's things you can do. So these two things, the fact that we're processing impulses and not continuous continuous wave signals, and the fact that we have access to such a breadth of bandwidth, uh allows us to be quite robust to interference. And what's important too is ultra-right band emissions are regulated across the world to be very low on average. So if you look at a USB port on your computer, uh depending on the quality of who designed it, right, it might have more emissions than an ultra-right band transmitter. Right. And uh so if we're gonna worry about ultra-right band transmitters, uh, we're gonna have to have bigger problems with all the USB ports we have dangling everywhere, right? So um so that's kind of where I feel the technology is is well positioned for robustness, but also well positioned to not interfere, which I think is as important. Uh we, you know, it's it's we we cannot be selfish in the wireless world, right? We cannot be bad neighbors, otherwise everything collapses.

SPEAKER_00 16:39

Oh, such a great point.

Chips SDK Reference Designs And Antennas

SPEAKER_00 16:41

Um to talk about uh Spark, your your technology and go to market. Uh I imagine you you provide the silicon or modules and some software stack to go with it. How do you describe how you're packaging your technology today?

SPEAKER_01 16:56

So we're a chip company, we sell wireless transceivers, uh, and we have a software development kit. Uh and of course, we have reference designs for some uh token applications that we've done quite a bit of work on so that customers can take those reference designs and quickly, you know, spin up their products. So the software development kit is more of a tool where you want to do something more custom that's a bit away from our reference design. It's fully featured and documented on our website. Of course, the chipsets we have two generations shipping now. The first generation product uh started shipping about four years ago. Now the second generation product uh has been in production for about 12 months. Now we're gonna have you know third gen products and so on in the future. And the idea is always to bring higher performance metrics to what I said earlier on those uh value, value you know, capabilities that we bring. Um, and of course, soon we're gonna have modules as well, which can be easier for customers that may not want to work on neuroscertifying uh devices. Uh, but I do want to say our team also has a lot of antenna reference design. And antennas for 2.4 gig radios like Bluetooth or Wi-Fi, that's you know, that's run in the mill. But for ultraide band, it's not quite. And uh, we offer a lot of very good antenna IP that we provide you know freely to customers that use our reference designs or SDK so that you can simply put a PCB trace on there, which is the cheapest thing out there, right? And get yourself a very high efficiency ultra band antenna. And you don't need Rogers, PCB, only anything esoteric there. Just go to FR4 and you'll be off to go. So it's not a really uh expensive technology when you use a Spark Radio R transceiver, okay, a few passes, I think it's about 10 capacitors, one coil, and a PCB antenna, right? So it's uh relatively low cost, uh, even compared to BLE, which is the king of low cost and good enough. So I think that's important too. We wouldn't be in mice if we we were you know an expensive technology, obviously.

SPEAKER_00 18:55

Very cool. And uh, you know, designed in North America, you must be servicing supporting customers, OEMs around the world. That must be exciting, uh, seeing the whole dynamic there play out everywhere around the world.

SPEAKER_01 19:08

Oh, absolutely. You know, we have presence in Asia, Taiwan, uh China, France, uh, Japan, of course, the US, uh, Europe. Uh smaller company than the big guys, but we have the same, you know, worldwide presence. Uh obviously, it's always uh great to go see our teams and have them come over to HQ in Montreal, Canada. And uh we have also a very well-set uh uh you know network of disciples and so forth. And the channel is is, I'd say, uh uh getting pretty robust by now. Uh obviously, we're a new technology, so it's always more challenging to bring something new to the table. But uh consider all considered we're doing pretty well.

SPEAKER_00 19:49

Pretty

The Last Meter Problem And 6G

SPEAKER_00 19:50

well indeed. And uh you have your eye on 6G as a CTO. Do you early to talk about 6G? Uh or is that something you maybe do in your spare time?

SPEAKER_01 20:00

Uh 6G is an interesting technology. I'd say I know at least in Canada, 5G has just started to really become mainstream, believe it or not. Uh, you know, Canada is a wider uh geography with uh density of population, so I feel 5G took a long time to make it. So I'm pretty sure 6G is not coming in soon. But what I like about the promise of 5G and 6G, it's a lot more than just pure data rate. It's about latency, right? As you know, there's a lot of stuff about latency. And when you think about it, imagine this. It's kind of an interesting thought experiment. You have a wonderful data center that costs you, I don't know how many billions nowadays, right? To do your AI processing. You have this amazing 5 or 6G infrastructure that costs also a fortune that brings the wireless signal through your cellular at single digit millisecond, maybe even sub-millisecond. But then you've got, you know, some Bluetooth headset that gives you a millisecond, 150 millisecond latency for that last half a meter from your pocket to your ears, and you just blew up the whole latency budget just for an audio signal. I'm just giving you that thought experiment. Isn't that kind of strange that you spend billions on a cellar architecture on a data center to be low latency processing, and you can't close that link at low latency? So, of course, Spark brings that to the table, right? How do you close sub-millisecond latency at that last, you know, uh at last meter or last couple meters?

SPEAKER_00 21:26

Yes, that's a great thought experiment. The last meter. So congratulations on all the success. Really interesting technology, and uh it's great to see the commercial uh values so apparent as well. Good luck, onwards and upwards. Thank you. Thank you, Evan.

Wrap Up And Where To Follow

SPEAKER_00 21:41

Thank you, and thanks everyone for listening and watching. Also check out our TV show in addition to the podcast Tech Impact.tv on Bloomberg and Fox Business. Thanks, everyone. Thanks, Frederick. Thank you very much for having me.