I purchased a Fitbit Charge HR this morning. Although this was somewhat of an impulse purchase, I had been thinking of acquiring one for a few months now. The purchase was primarily motivated by my partner’s latest pursuit for improved fitness, but I guess any excuse for self-quantification is a good one for a statistician like me.

I paid a little more for the Charge HR (over the Charge), as I wanted the heartrate monitoring to appease my quantitative self, and after some initial confusion at Argos, where I received the inferior Fitbit Flex inside an already opened Charge HR box (a scam presumably relying on the member of staff processing the return not knowing the considerable difference between the two items), I took my new toy home for charging before use.

The device charges via USB and so I took this opportunity to perform the initial setup via my desktop. You can do this via a mobile device by downloading the app, but I figured that I may as well use the desktop seeing as it was already plugged in to charge. Although it turns out all configuration and synchronization with the device is performed over Bluetooth not USB.

Once you’ve downloaded the installer, I had to make the difficult decision to connect with Facebook, Google or neither. I reluctantly chose Facebook as I figured it would make finding and sharing statistics with friends simpler, though I was not spared from having to faff with creating an online account.

During initial configuration, I downloaded the companion Android App to my phone and logged in via my new account. It was interesting to find that I did not need to pair my phone to the Fitbit, as I had already paired the Fitbit to my online account during the desktop installation.

Impatiently, I equipped my left wrist with the half charged Fitbit. The strap is sturdy but quite difficult to adjust (which I’ve needed to do a few times to maintain comfort) and unfortunately catches my arm hairs. However, I was happy to see it was stronger than the clip-on strap on the Flex that I was briefly acquainted with earlier. Despite terminating the charging cycle early, there should still be a few days of charge remaining.

The mobile application is pretty, friendly and easy-to-use. I input my height, weight, a weight-loss goal, a step goal and two morning alarms (for weekdays and weekends). I’m interested to see how effective being woken by the device vibrating is, against a typical audible alarm.

The Charge HR has a small screen for displaying the time and various stats (steps walked, current heartrate, calories burned etc.) which can be cycled by pushing the Fitbit’s only button. Additionally, one can configure a default display mode for double tapping, and turning the screen towards your face. I chose to display my current heart rate and the time, respectively. I was impressed to find detection of the latter motion quite reliable.

I was disappointed to see that the heart rate monitor did not appear to be synchronizing with the mobile app, but I later discovered that it appears to need to collect a few hours of data to estimate your resting heart rate first.

I expect the Fitbit is most useful when you log information on calorie intake and exercise, so the design of the companion application was a driving factor for my purchase. Logging is quick and simple; for food intake one can scan a barcode or search an extensive database to fetch predefined information on nutritional values¹. Frequent items and itemsets (i.e. meals) are enumerated for quick access. Activities can be logged in a similarly easy fashion; you may select from a list of predefined exercises with a predetermined per-minute calorie burn. I like that walking, running and hiking specifically allow you to use your phone’s GPS to more accurately track those particular activities.

As a fan of numbers and graphs, I’m looking forward to using the application when more data has been collected. I’ll report back then.

Today Vic and I had the opportunity to try out taking some micrographs for ourselves. As additional steps are necessary for pre-processing biological samples (primarily the time consuming step of drying and fixing), we decided to find something interesting but not alive that the lab might not have had seen before. We settled on bringing along a bag of various 3D-printing filaments that each possess some interesting property that we hoped would lead to interesting images too.

Sample Preperation

Samples of a few millimetres in length were stuck to the sticky and conductive surface of specimen bolts. Non-conductive samples must be coated in a thin layer of some conductive material so as to be interactable and thus detectable by the scope. The bolts were loaded in to an argon vacuum sputter coater and coated in a layer of platinum-palladium alloy.

Above, samples mounted on specimen bolts (or stubs) ready for coating.
Below, samples tripping the light fantastique.

After coating, samples were immediately ready for loading in to the Hitachi S-4700 field emission scanning electron microscope.

Pseudo-Wood

First up; Lulzbot Laywoo-D3, a “wood based” filament designed to look like cherry wood when extruded¹. Indeed the exterior demonstrated by the first image below seems to have a bark like appearance. Below you can see a cross-section of the filament, a rugged exterior protects a sponge-like fibrous interior.

I wasn’t exactly sure what to expect but had initially envisaged being able to see a more wood-like structure. Of course the product is designed to be both flexible and safely extruded without blocking or damaging the printer head and thus it is possible the wood content is milled and uniformly mixed with the ABS.

Pseudo-Brick

Lulzbot also offer Laybrick, embedding “minerals” to offer a stone-like appearance. On the surface, the resulting images are similar to that of the wood-based counterpart, featuring a textured exterior. Appearing almost smooth at 45x, the coating reveals a more fine texture at 180x. The interior appears more compact and rigid.

I had expected to see more of a crystalline structure due to the minerals. But again, I imagine this would cause trouble during extruding. It should be noted the products are designed to look like wood and brick and achieve this job well enough with their exteriors. There isn’t a pressing need to emulate the materials under the surface!

Glow in the Dark ABS

We also took a look at Lulzbot Glow in the Dark ABS, curious to see whether the glowing property altered the structure in an interesting fashion. From what we could see, it does not. However, this offers a useful comparison to our previous samples: note in particular the smooth exterior and the “cliff-face” at the cutting point where the blade slid across the surface. The interior appears rather solid, densely compacted and neither fibrous or porous.

Flex

Leaving the best and by far the most interesting (to me at least) until last, the Lulzbot Ninja Flex. The material is fascinatingly flexible with strong and durable elasticity. Unsurprisingly you could probably predict these properties from the microscopy images:

Above: the surface coated with obvious fibres. Below, Vic captures a ~10-15µm thick stray strand of flex fibre, followed by a cross-section of the cut sites. Note the “landslide”-like result of the blade passing through the sample, dragging external fibres inward and bringing some of the surface with it.

Above and below, the exterior surface of the “bottom” of the sample. The top image, taken at just 40x offers incredible detail of the complex waves of fibres that afford the material its flexibility, as well as Vic’s first failed attempt at cutting the sample. Below I focus on the few hundred µm chasm created by the failed cut at 110x, here you can see curling similar to that of the glowing ABS dotted with snapped strands caught by the blade.

Above, I capture the top surface of the sample, once again demonstrating the dense fibrous structure. Like the glow in the dark ABS, the interior appears solid rather than sponge-like. Fibres tightly overlap and interweave giving the appearance of a choppy sea. Longer and thicker strands are visible towards the background.

Below my own attempt to capture what I hope is a ~20µm wide stray strand at 2,000x.

Final Words

The imaging process was significantly more fiddly than I had anticipated, with each movement of the beam to another part of the sample often requiring re-calibrating even to get a reasonable idea of what you were now looking at. The microscope offered a low and high power mode; low power was limited on zoom but the latter required an additional axis of calibration² and gave poorer visual feedback on the current position. Focus was particularly hard to get right, even a good looking image turned out to be “just off” when loaded in to the seperate viewing software.

Towards the end I feel like we got the hang of things and am looking forward to returning once we think of some other interesting samples.

You can check out all the images via my Flickr album. Vic’s Flickr album has additional images (that also featured in this post) too.

tl;dr

• Microscopy is very cool.
• Microscopy is difficult.