Tundra Labs is releasing a new SteamVR Tracking hardware development kit. Featuring a tiny new module which condenses the necessary control, processing, and communication components into a mere 16mm × 10mm, the hardware development kit is smaller and cheaper, and aims to streamline the process of designing VR controllers and peripherals that use SteamVR Tracking.

Available for pre-order today, the new Tundra Labs SteamVR Tracking General Purpose HDK (TL448K6D-GP-HDK) is the next step in a continued evolution of smaller, simpler, and cheaper SteamVR Tracking hardware.

“Many companies and developers have recognized that SteamVR Tracking is excellent tracking technology,” said Reid Wender of Triad Semiconductor, a collaborator on the HDK. “Unfortunately, integrating SteamVR Tracking into a working product is a non-trivial task that involves a multidisciplinary team to pull off a polished product. Customers need SteamVR Tracking expertise, electronics prototyping, tracked object mechanical/tracking design, and last but certainly not least, SteamVR-specific firmware development. Even large companies have experienced difficulty pulling all of this expertise together under one roof.”

SteamVR Tracking is Valve’s room-scale tracking system which makes use of ‘base stations’ that sweep lasers across the room. Objects tracked by the system need two primary things: embedded sensors which detect the laser sweeps, and components which can process the data and communicate with a host device which will make use of the information.

The Tundra Labs SteamVR Tracking HDK aims to streamline things across the board by providing a cheaper and easier starting point for companies that want to build products with SteamVR Tracking.

Image courtesy Tundra Labs

The HDK includes a new module which condenses many of the necessary components into a tiny ‘system in package’ (SIP) measuring just 16mm × 10mm. The module contains the SteamVR Tracking firmware and handles all of the necessary data processing for the tracked object before it gets sent to the host device.

On the sensor side, the Tundra Labs HDK features 25 of the latest TS4112 sensors from Triad Semiconductor, another company which has worked closely with Valve on SteamVR Tracking. Triad says that its latest sensors are smaller, more power efficient, and offer improved tracking performance.

Flex cables tipped with SteamVR Tracking sensors allow for prototyping tracked objects of various shapes. | Image courtesy Tundra Labs

With shipping expected to start next month, the $130 Tundra Labs SteamVR Tracking HDK is significantly cheaper than a similar HDK previously offered by Triad for $350. Its core module is also much smaller, enabling new form-factors, like Logitech’s VR stylus.

The tiny Tundra Labs SteamVR Tracking SIP module | Image courtesy Tundra Labs

Tundra Labs founder Luke Beno believes the cost and simplicity of the HDK opens the door for SteamVR Tracking development to go beyond large companies with expansive resources.

“The low cost reference designs also greatly lower the barrier of entry for startups, students, makers, and hobbyists to experiment with the SteamVR Tracking technology and share their ideas. Tundra Labs has created online communities at Hackaday.io and Hackster.io where innovative minds can experiment with new designs, share them and potentially connect these concepts and ideas with manufacturers.”

Beyond its SteamVR Tracking HDK, Tundra Labs also offers custom design services for companies who want to build products with SteamVR Tracking, and sells the SIP module for companies who need the component for mass production of tracked peripherals.

“The Tundra Labs [SIP] was created to make the design and manufacturing of SteamVR Tracked devices easier, lower cost and with significantly faster time to market,” Tundra Labs writes on its website. “For the first time these modules offer a fully scalable production solution to satisfy any volume. This scale spans from niche devices built in the 10s of units to mass market consumer devices with volumes of 100,000 or more.”


Valve Updated SteamVR Tracking Because 'Beat Saber' Players Were Too Fast

While the latest Tundra Labs SteamVR Tracking HDK is built for general purpose use-cases, the company also plans to release application-specific reference designs in the future.

“These could be used for body tracking or alternative controller form factors, ultimately the target use will be determined by customer interest and feedback,” said Beno. “The goal of all reference designs is to leverage the extensibility of [SteamVR Tracking] and the [SIP] to enable a diverse set of accessories into the ecosystem for both consumer and enterprise purposes.”

Newsletter graphic

This article may contain affiliate links. If you click an affiliate link and buy a product we may receive a small commission which helps support the publication. More information.

Ben is the world's most senior professional analyst solely dedicated to the XR industry, having founded Road to VR in 2011—a year before the Oculus Kickstarter sparked a resurgence that led to the modern XR landscape. He has authored more than 3,000 articles chronicling the evolution of the XR industry over more than a decade. With that unique perspective, Ben has been consistently recognized as one of the most influential voices in XR, giving keynotes and joining panel and podcast discussions at key industry events. He is a self-described "journalist and analyst, not evangelist."
  • Ad

    SteamVR tracking has fewer limitations and more potential than any other solution, so I hope it continues to get cheaper so it doesn’t have any issues competing. Valve also needs to push it a little more like by selling body trackers with the index kit.

    • Inside out has more potential in my opinion. All that holds it back being perfect is accuracy and that is, and can be, improved with software. It also paves the way for mobile AR making it more profitable.

      • brandon9271

        Every inside out tracker has a limited FOV and problems focusing on things that are too close. I yet to see a system work any where close to SteamVR or even Rift CV1

        • Lachlan Sleight

          The reality is that right now, there are basically three user groups for VR: Prosumers, VR Enthusiasts and Enterprise. The only people who are willing to deal with inconvenience and a higher price in exchange for a marginal improvement in tracking quality are the hardcore VR enthusiasts.

          Enterprise wants simplicity and reliability over everything else, and prosumers (who often don’t have their VR equipment set up 100% of the time) just want to be able to take the headset out, plug it in, and get into VR.

          Outside-in will always have a place for specific enthusiast / specialized enterprise applications, but inside-out is the way of the future for the majority of VR use-cases – especially once VR hardware merges with AR hardware.

        • mirak

          Yet, and I agree and still choose steamvr but someday the controllers will track themselves with cameras, like the wiimote did.

          Therefore there will be no occlusion, and even be better than lighthouse, except maybe for precision.

          • brandon9271

            That would work well and probably still not cost as much as a SteamVR setup.. I guess you’d need to put some LEDS on the HMD as a reference point in order to fuse the coordinates from both controllers… or actually, the HMD would already have cameras of it’s own so just adding a few LEDs to the controllers maybe.

        • kakek

          Have you used a Rift S recently (with latest software updates ) ? Those problem have already been lessened A LOT.
          Proximity is only a problem at a few centimeters from the headset. So much I am more bothered by my controller touching the headset than loosing tracking anyway. I would argue that in this case, it’s already resolved.
          Sensor FOV can be extended by using more cameras. But with 5 cameras it already pretty much cover all you might need in normal use. By that, I mean that I was not able to tell where I lose tracking without having a mirror for my avatar, and doing specific movement I knew would make it happen.

          For now, precision remain slightly inferior, and occlusion more likely to happen with inside out tracking.
          Though precision is already pretty satisfying, and can increase with future generations ( all you need is higher def / faster cameras).
          And oclusion can also happen with outside trackers, and can ALSO be improved with improvement of inside out tracking.

          The only downside I could see is potential full body tracking. I don’t see a way to do that without external cameras.
          Multiple trackers on the body not being a great solution.

          I’d say on the long term, inside out has more potential.

      • Those systems are for diffrent usecases. Vive trackers, when the companies wil already offset the developing costs, can cost single dollars to produce, just in diffrent form factor.

      • Ad

        Of course AR glasses will use inside out, I consider that a separate issue. But SteamVR tracking is more modular, supports additional peripherals, and generally has fewer limitations.

  • Jag

    I remember seeing almost an exact replica of this article for the 2.0 Base Tracking Stations. They spoke of having half the amount of internal components (which they practically did) and significantly lowering the price of each station for consumer units. Well guess what. Today they are TWICE as much as the 1.0 stations and I’m not gonna fall for hype again. I’m keepin’ me money!

    • Well, they are $150 instead of $100, which is not twice as much, but your point still stands. It’s ridiculous. Why did they claim it was going to be cheaper if they had no intention of making it so?

      • Max

        Because they weren’t expecting to be sold out for two straight months and counting on half their products? The prices will come down when people stop paying them.

        • Trip

          That sounds good, but to the best of my knowledge they weren’t sold out for at least a year leading up to this and the price was higher right on through. Either things didn’t go as expected, or someone decided not to pass the cost savings on to the consumer as was originally implied.

          • Immersive Computing

            From my understanding things didn’t go as planned and failure rates of 20-30% at factory (Flex @ Buffalo Grove Illinois – assembled in USA using foreign sourced components) meant production is off schedule. Some redditors report waiting from before Christmas for replacement 2.0 base stations with no ETA, and a number of these reports were for 2.0 dead on arrival or failure soon after powering up.

            Speaking to developer at event recently using Vive Pro, he said they’d had over a dozen 2.0 base stations fail within a year, compared to 1 failure of 1.0 over 2 year prior. There have been revisions of 2.0 base stations from units supplied to HTC and perhaps ongoing revisions for Valve’s units for Index?


          • Rogue Transfer

            So, the question then is, why should they have failures when the technology used is less of the same parts in Lighthouse 2.0. One harddrive motor, rather than 2; two diode lasers and an integrated controller chip. All long-established standard off-the-shelf parts.

            There’s really very little that should go wrong with such technology. It’s surprising if they have a worse failure rate than 1.0, which had all the same tech, but double the motors(and vibration).

          • Christian Schildwaechter

            Pure speculation not based on any knowledge of the actual 2.0 failures:

            The reduction of parts may be the actual cause of the problems by increasing the load on the remaining components that now have to also do at least part of the job of the ones removed. The initial tracking stations may have been intentionally over-engineered, e.g. using two motors because they weren’t sure one could handle the load.

            Cost reduction is often achieved by building towards lower requirements after a couple of years taught you which of your safety margins were wasteful. Over time the motors could have shown to be sufficiently resistent and/or firmware updates made the second motor unnecessary. So 2.0 used only one, reducing cost and complexity, but in real live usage the single motor now tends to fail more often than expected.

          • coromd

            Reducing component count and improving accuracy+range also requires everything to be working with much tighter tolerances. I don’t know much about how much that effects price, but it could be a reason as to why the 2.0 stations haven’t come down in price.

      • Jag

        Even trackers are still a ridiculous $100 EACH. If anybody wants to jump into VRchat with feet and knee tracking, that’s fucking $400. If they want elbows, that’s $600+.

    • Sofian

      They meant they would make more benefits not lower the price.

      • Trip

        Actually no, they specifically said they would be less expensive.

    • This wave of VR is still in it’s infancy and there’s a lot of figuring out to do regarding the establishment of stable reliable platforms and products that are viable for businesses and consumers.

      This is an excellent tool/platform to help explore new devices and use cases for this medium and will be worth every penny!

      • Jag

        “Will” is the operative word. Not right now.

  • Trip

    I think we need an affordable pair of wearable controllers. A VR glove, but without any super fancy features. Simply a direct replacement for the VR hand controllers we have now, that doesn’t interfere with you being able to pick up other objects.

    Grab a beverage, answer the phone, use a joystick, and pick up a replica rifle stock for games like Onward and Pavlov without having to stop and put down (or in the case of Knuckles take off) the VR controllers, or to build stocks that can mate up to a controller, making them a bit awkward to pick up and put down for throwing grenades etc.

    Do whatever you need to do with your hands, and they are still always there and ready to be used.

    With lighthouse tracking and flex sensors this could easily be achieved right now. I just wish I had the resources and skills to do it myself. Instead I’m looking at SensoryX VRFree and other glove projects that are less practical IMO.

    • CMDR Megashot7727

      If hand tracking improves, that could work you you

      • Trip

        I don’t see it getting accurate or reliable enough to be an enjoyable experience in the very near future. With a glove using lighthouse tracking and flex sensors this could be viable today. There have been a few attempts, but AFAIK none of them committed to using Lighthouse and so positional tracking was lousy. I think it was Noitom talked about optionally strapping on a Vive Tracker but it never got properly supported to the best of my knowledge.

        What we need is a basic no-frills glove specifically for SteamVR and designed to act as a direct replacement for the existing hand controllers. I think nearly all of the effort required for development would be software side. It seems like a market ripe for the picking, especially given the success of the VR shooter and VR Simulator markets.

        The thumbstick functionality would be the only difficult part I think.

        • Immersive Computing

          At this point I’d like a Touch CV1 clone with steamVR tracking. Something simpler and more reliable than Index controllers?

          (I’m on my 5th pair of Index controllers since launch)

          • Trip

            I’ve not had trouble with my knuckles yet but I’ve been too busy with simulator building to put many hours on them. I do agree that the original touch controllers were excellent, superior to Knuckles except for the excellent idea of strapping the controllers to the hand. The newer inside out touch controllers are not nearly as ergonomic for me as are the originals.

          • Immersive Computing

            Touch CV1 controllers and Xbox controller designed by ‘Carbon’.

            My latest… playing Pistol Whip perhaps as I’d been waiting weeks for replacement left controller so just using right.

            Valve have replaced no fuss, though slow to respond and ship


        • G-man

          the problem is you cant have a felxible object being tracked using steam vr tracking. you need to triangulate the position of the object to know where it is. if you bend a finger then the sensors will ping as it is sweeped by the laser but it wont know what part of the laser hit it, unless a second sensor a known distance away pings a certain time after it. the difference in time allows for triangulating the position of the object. so every finger would need to have at least two sensors on every rigid section of the finger and two sensors would have to be in line of sight of a lighthouse to be tracked. its not really viable. or at least would take a lot of intelligent guesswork in the software to get anything usable out of it. add to that the probem of designing a piece of hardware that fits everyones hand. it would be a nightmare.

    • If you had the right resources and skills, you’d realise that it can’t be easily achieved right!!

      • Trip

        Fair enough! Nothing is ever as easy as it seems, especially if you aren’t experienced in dealing with whatever the goal is.

        I was going to say that for my use case all I need is some software made and I could use Vive Tracker’s but that reminded me that while I’d much prefer the gloves were Lighthouse based, I decided occlusion from the switch panels would be an issue. Time to write this dev kit off as a possible solution I guess, at least for now.

    • david vincent

      I doubt we’ll ever see mainstream VR gloves (even with haptics). Too much friction, no buttons, no sticks, not enough added value compared to Knuckles or Touch…

  • Rogue Transfer

    Some misunderstandings in the article. SteamVR tracking objects are not tracked by the system; objects track themselves from the coded IR light from the IR Lighthouses. The processing can happen either onboard or off-loaded to a computer.

    So, the correct terminology is to say: ‘tracking object’, not ‘object tracked’. This is a crucial difference, as it explains that the objects are independently tracking themselves and are not perceived/recorded by an outside tracking system. A SteamVR tracking object can work completely privately.

  • Baldrickk

    With any luck, this will lead to cheaper body trackers.
    I’d be perfectly happy with wired trackers, connected via USB to the headset, thus forgoing the need for batteries and wireless radios on the trackers and the associated radio dongles on the PC end.

    Personally, I’d be fine with having wires running down my leg, with a strap to keep it from flapping around.

    In theory, this would allow a lot of the cost to be sidestepped, as you need only the sensors and this “all in one” chip.

  • Interesting kit, really interesting. For makers it is a joy for sure