The Dash wireless earbuds cropped up earlier this week on Kickstarter, and you guys have been blowing up our inbox requesting a review. We can see why you’re interested; they’re offering a pair of wireless earbuds with integrated heart rate monitor, bone microphone, touch-sensitive controls, voice feedback, 4GB MP3 player, and more. While a cursory glance might bring up memories of the Smarty Ring, a device with a similarly extreme sizing and technology requirements, there are a few details present in the Dash campaign that make it not so far fetched.
The Dash claims that their battery has a capacity of 100mAh or roughly twice the capacity of a 4th generation iPod Shuffle. 100mAh isn’t a lot when compared to your typical wearable device, but The Dash has some pretty modest claims for battery life. They claim 4 hours of “play time” or 3 hours of “play & track time”. I’m assuming play time means audio playback (Bluetooth or MP3) and track adds optical heart rate measurements. This means that the audio player requires about 25mA and the heart rate sensor draws an additional 8.3mA.
Streaming audio over Bluetooth takes more power than reading it from a flash chip, so let’s focus on that interpretation of the battery life claims. Based on the Bluetooth specs listed on the campaign (CSR aptX codec, Dual Mode Bluetooth 4.0, etc), I’m assuming that the Dash is using the CSR8645 or something very similar. Page 92 of the CSR8645 data sheet lists the power consumption of the device in several different operating modes. Their “Stereo high quality SBC” mode is the highest current draw listed, and it requires 13.3mA with microphones and speakers disconnected. That leaves around 11.7mA (43mW from a 3.7V battery) to power the speaker. Meeting the 10mW output power specification they list would require an audio amplifier with just 25% efficiency which is very reasonable. For comparison, the 4th generation iPod Shuffle uses a 190mWhr battery and has an advertised battery life of 15 hours. That works out to around 13mW to drive two headphones and decode the audio files.
The “track” functionality is performed through optical heart rate monitoring (OHRM). This process consists of shining light into the wearer’s skin and monitoring the light reflected back. Unfortunately, it’s very difficult to obtain accurate power requirement numbers for OHRM. Everyone does OHRM differently whether it’s by placing the sensor on different parts of the body or simply using different wavelengths of light. Texas Instruments recommends Red and IR (like Dash) for their AFE4400, but their spec sheet doesn’t include a recommended LED current (which will make up a majority of the power profile). I also couldn’t find the battery size specifications for any of the optical heart rate monitoring devices currently on the market, so I couldn’t estimate power requirements that way either.
Just to get a ball park figure, the AFE4400 has a maximum LED current setting of 50mA. Each measurement consists of four phases: two LED light measurements (IR and Red) and two matching ambient light measurements with the LEDs turned off. This means that each of the two LEDs can be turned on a maximum of 25% of the time (25% “duty cycle”). The absolute maximum average current draw from the LEDs is therefore 25mA. That’s higher than 8.3mA, but the expectation is that this kind of device will be configured based on the application and not always operated at 100% power capacity. The figures on pages 12-14 of the data sheet indicate that the device can operate with fairly consistent performance down to 5% duty cycle which drops the average down to 2.5mA even if the LEDs are still operating at full brightness.
Even adding the processing power required to analyze the resulting data, I have to conclude that 8.3mA is definitely within the realm of reason for optical heart rate monitoring.
Many times, the only thing separating the gadgets of today from the technology of the future is size. Making a smart-headphone isn’t difficult, but making one small enough to fit entirely in an ear is. It makes sense therefore to really scrutinize the size and shape of the Dash headphones when trying to evaluate their viability. Before digging into the innards, it’s important to verify that the scale of the device shown is accurate. We’ve seen projects like Smarty Ring grossly misrepresent the actual dimensions of their products in their marketing materials, and we want to make sure that backers won’t be surprised by what they get.
Fortunately, Dash provides many specific and consistent details concerning the size of the product including a dimensioned line drawing:
A little bit of Photoshop work indicates that their prototypes shown more or less match this scale (quarter for reference):
I don’t spend too much time looking at my own ears, so I had to grab a quarter myself to verify, but a quick check indicates that this person has normal-sized ears and this earbud is about as big as you think it is. With that out of the way, let’s take a look at what’s inside.
It looks like they’re using a pretty cleverly designed rigid-flex circuit board to pack all of their components into a tiny form factor. This is a fairly common manufacturing method that usually involves connecting multiple rigid component-laden sections of circuit board together with flexible ribbon cable-like segments. These flexible sections are much smaller and more durable than typical wires or connectors, and they allow for some pretty wild form factors. With careful design, PCBs can be assembled flat and then folded up like a complicated origami structure. This image of some other examples might make this methodology a little more clear:
From their x-ray view, it looks like the circuit wraps around the battery. Comparing their line drawings to their x-ray view with a little CAD work, it’s pretty easy to estimate their battery size:
So for battery volume, it looks like it’s around 15 x 15 x 9mm or 2025mm^3. A 100mAh battery will hold about 0.37 watt-hours which gives this battery an energy density of 183 watt-hours per liter. Assuming this device is using a lithium-polymer battery with standard chemistry, this is right on par with expectations. Li-po chemistries can achieve energy densities of around 300WH/L (secondary source), and when you factor in the volume taken up by the battery packaging and swelling, 183WH/L is very reasonable.
After the battery, most of the components are very small. Their radio is most likely a CSR8645 which has a footprint of 5.5×5.5mm and contains battery charging circuitry, an audio codec, noise cancellation, and even a controller for a few LEDs. Heart rate and pulse-oximitry could be handled by an analog front-end such as the Texas Instruments AFE4400 which is only 6x6mm. The largest part besides the battery should be the 4GB memory storage module, but I’ve found a few examples including an offering from Kingston that can be found as small as 12x16mm. The flash part is most likely the large square component in the center of the x-ray view.
I’m genuinely surprised by how plausible their sizing is. All of these components are around 1mm thick, and given the dimensions of the earbuds, stacking them up with a flexible circuit board should allow them all to fit easily. Rigid-flex construction is difficult and expensive, but if they manage to pull it off, there’s no reason in my mind why they shouldn’t be able to fit what they’ve claimed into the form factor they’re demonstrating.
My first reaction looking at their project page was one of incredulity at how this thing was going to determine the wearer’s distance or speed without the use of a cellphone’s GPS:
This vague wording seems to imply that the device can be used to track distance without the use of a cellphone, but it’s only in a foot note farther down the page that they indicate that it indeed requires a cellphone to use this feature.
I personally think that they could be a little more clear about this limitation given that they’re really pushing the “no phone required” angle, but it doesn’t look like they’re trying to trick anyone. With the ridiculous prospect of cramming a GPS radio into a headphone out of the way, let’s look at some of the other challenges.
One of the less obvious issues that will crop up with two independent earbuds is audio synchronization. The human auditory system is incredibly good at detecting time delays as small as 10s of microseconds between audio hitting one ear vs this other. This is one of the methods used to locate the source of a sound. If Dash doesn’t want their users to perceive their headphone audio coming from a random direction, they need to be able to synchronize these devices very well.
Unlike Split which relied on a lack of radio communication to improve battery life, the Dash headphones will be in constant communication which reduces the amount of error introduced by clock drift substantially. As far as these pings go, I found a paper discussing methods of synchronizing Bluetooth sensors down to “a few microseconds”, so it looks like something like this is possible if they have access to some of the low-level features of the Bluetooth chip they’re using.
The feature that I’m most concerned about at this point is the waterproof rating. They claim that it will withstand 1 meter submersion which I’m assuming means IPX7. If I were a mechanical engineer, I could speak more to the specific challenges of waterproofing, but the general rule is that waterproofing gadgets is never an easy thing to do; it gets especially difficult when you have interfaces like speakers, microphones, and cable connections. They haven’t demonstrated the water proof capabilities of their device yet, and they don’t specify when they will be doing waterproof testing in their production plan. They also don’t mention it under their Risks and Challenges.
Can they do it?
I started writing this article thinking that these earbuds were a hoax, but try as I might, I can’t with any good conviction say that they aren’t possible.
That being said, they won’t be easy to make. The design is incredibly complex, and tacking on features like waterproofing adds a whole new level of complexity to an already difficult project. If the team were a group of college kids, I would be more concerned, but a quick glance at the project’s founder’s credentials show that this guy has been around the block a time or two.
This project has many of the hallmarks of a successful Kickstarter project. The product specifications are modest, the dimensions are specific and consistent, the 3D models and progress shots are realistic, the physical prototypes look polished, and the team behind it is competent. The only other thing I could ask for is a product demonstration, but their timeline shows that while they have verified their electronic design, they won’t have working Alpha prototypes until June.
The timeline doesn’t have much wiggle room for unanticipated issues that may crop up in the future, and they don’t have secondary sourcing on many components should a particular vendor fail to deliver on time. There are a number of things that could prevent them from hitting their November 2014 delivery deadline, but if you have backed this project, plan to receive a product with few if any compromises.
As Adrian points out in the comments, The Dash team has actually detailed what looks like a pretty extensive amount of waterproof testing in the FAQ section of their product page. There’s still no guarantee that they won’t run into issues with mass production units, but it’s just about all one could ask for at this point in their development cycle.