Hackaday Prize Entry: The Chocometer

Diabetes affects almost 400 million people worldwide, and complications due to diabetes – blindness, cardiovascular disease, and kidney problems – can be reduced by regular monitoring of blood glucose. The usual way of measuring blood glucose is with a pin prick of blood and a small test strip that costs about $0.30. That’s a lot of test strips and blood used by 400 million diabetics every day. Wouldn’t it be better if there was a less invasive way of measuring blood glucose?

[marcelclaro]’s project for The Hackaday Prize aims to do just that. Instead of measuring blood directly, his project will measure blood glucose by shining light through a finger or an earlobe. Using light to detect blood glucose is something that has been studied in the lab, but so far, there aren’t any products on the market that use this technique.

There are two major problems [marcel] needs to overcome to turn this project into reality. The first is simply raw data for calibration. For [marcel], this is easy; he has Type 1 diabetes, and takes four glucose measurements a day. Patient heal thyself, or something.

The second problem is getting a photosensor that’s sensitive enough. By using an InGaAs PIN diode, a current-controlled oscillator, and a digital counter, [marcel] should have a sensor that’s good enough, with electronics that are cheap enough, to create some tech that is truly game changing for a few hundred million people around the world.


The 2015 Hackaday Prize is sponsored by:

44 thoughts on “Hackaday Prize Entry: The Chocometer

    1. I think some of the software defined radio kits cover up to ~5GHz. Add a custom antenna, and I bet you could do proof of concept experiments. A quick Google shows some interesting glucose absorption features out past 3000nm, though picking glucose out of the soup of compounds in blood sounds “fun”. (raman spectroscopy might also work if a glucose signature could be found)

      1. I already think about Raman, but the portable spectrometer is expensive and low resolution. Formally I working in a lab of mid infrared, I can make sensors of any wavelenght between 1 and 15um, but this sensor need criogenic temperature and is very expensive. 9,3um could be nice for this, but, only if implanted or made with a material which do not need liquid nitrogen

  1. > Wouldn’t it be better if there was a less invasive way of measuring blood glucose?
    This is not the only problem: finger pricking delivers one result at a time. It does not tell you whether your sugar is level, going up or going down. Currently, continuous blood glucose measurement is also invasive, and very expensive, but it is the only way to get a better idea of what your blood sugar is doing over time.
    As a Type 1 diabetic myself, I will be following this hack entry closely. Good luck!

  2. Medical science has been doing research on how to detect blood sugar in blood via noninvasive optical methods for decades. By now there are tons of books and papers on this. So far nobody has been able to make this work even as a proof of concept.

    Blood sugar level simply does not alter the optical properties of blood enough while many other unrelated factors do. As far as we know today it’s pretty much impossible to isolate the blood sugar information from these other factors without mapping a whole lot of other bodily functions at the same time. This would require invasive techniques, rendering the whole approach moot.

    I don’t want to discourage marcelclaro, but the chances that he succeeds where highly specialised teams around the world have failed again and again is very slim IMHO.

    1. I know nothing about diabetics, so these questions are either really ignorant and stupid, or hopefully will trigger some ingenious burst of inspiration in someone else…

      1) Sugar in urine? What about sweat? Saliva? I seem to recall (myth/urban legend?) in days of old diabetes was detected by the taste of urine (*shudder* am very glad not to be a doctor and that science has improved). If true, could one make a sensor you pee’d on to test sugar level? and then if sugar is in urine, would it be present in other bodily fluids? (from first aid training years ago – I recall apple cider smelling breath being a sign of blood sugar issues).

      2) How does blood sugar affect other bodily processes – can levels be inferred by measuring changes in skin resistance, strength of neural signals (low salt levels can manifest as fatigue, maybe sugar affect neural transmissions as well?)

      1. Ignorant, yes. But your questions aren’t stupid. Glucose gets spilled into the urine only when your blood levels are really out of control. You want to be able to measure the blood glucose levels in the normal to near normal range. You don’t ever want them so high that you’re able to measure glucose in your urine. I would guess the same about sweat. You’re right about the apple cider breath but same issue, your levels are way too high by that point.

        Like you, I’m ignorant on your question 2. Blood sugars affect a lot of things but I’m guessing those things aren’t easily measured noninvasively or have the same problem as your question No. 1.

      2. I cal help a little bit.

        1) Glucose in blood and to a lesser degree, subcutaneous tissues, has a concentration between 3 and 8 mmol/L. Up to 30 mmol/L in disease. Contrast this with sodium, ranging 130 – 145 mmol/L.

        Thus, glucose is easier to detect in urine. Diabetes Melitus basically means “pees a lot of sweet urine”. It also froths.

        2) However this and similar measurements (ketones, etc.) are delayed quite a lot, and tend to occur with pathology rather than before it. In the urine’s case, after the glucose has ripped up the filtration mechanism.

        The difficulty comes from keeping values normal, not just detecting abnormalities

      3. 1) Yes, you can measure glucose in the urine, when it gets above roughly 9 mmol/l (~160mg/dl) in the blood. You can buy strips that change colour in the presence of glucose in the urine (called Gluostix). The major problem with this is the time lag. Urine collects over (usually) several hours, so if you test your urine for glucose you’re really just seeing an average since you last urinated.

        This isn’t very useful if you’ve spent most of the last hour with a blood sugar of 15, but your insulin has now kicked in and it has dropped it a lot, and you now want to know if your blood sugar is 3.0 mmol/l (which would require treatment with glucose) or 5.0 mmol/l (which wouldn’t). As both of these values are below the renal threshold, they would both just show up as ‘0 glucose’ on the glucostix.

        The sweat smelling breath is (probably) a result of lack of insulin. The body requires insulin to use the sugar in the blood. If you don’t have enough insulin, it switches to burning fat. A byproduct of using fat for energy is ketones. When these accumulate enough, it can be possible to smell them on someones breath. Ketones are acidic and have a serious impact on the pH of the blood. If the buildup gets too high, the pH drops too low, eventually reaching a level that is “incompatible with life” if nothing is done to remedy the situation. Another thing to note with high blood sugar is that it dehydrates you (because when sugar is spilled in to the urine, it drags fluid with it), which further increases the concentration of the ketones. Low levels of ketones can be handled by the body (most people have them first thing in the morning, or when they’ve fasted for > 8 hours or so), but without the ability to produce insulin, things can go downhill very quickly.

        I’m not sure about whether sugar is present in sweat, but I wouldn’t like to have to be constantly sweaty to get the volume of fluid required for an accurate test. Saliva would be tricky due to trace amounts of sugar left after eating. You also have amylase in your mouth, which would break down any starchy food left hanging around in to glucose, which would then alter the test result.

    1. Yes, let’s put all the success of this project on me, and not the guy who is actually building it.

      Also, let’s link to a self-published book on noninvasive glucose measurement written by a freelance writer, and ignore the evidence published in a peer-reviewed journal saying it will work.

      1. The disease was known since antiquity but they could not do much about it.
        Aretaeus a Greek physician who tried to treat diabetes described life with it as short disgusting and painful.

    1. Oh dear. Where to start?

      I presume you’re talking about type 2 diabetes which is associated – amongst other things – with high carbohydrate diets.
      But actually we only knew the difference between insulin and non-insulin sensitive diabetes since 1936; so the disease is only ~80 years, not 200 years old.
      But then, the syndrome of diabetes has been known since antiquity. It was probably described by the Egyptians 3500 years ago, and definitely by the Indians 2600 years ago. So this existed more than 200 years ago.
      And in addition to the occidental, the Greeks, and the ancient Indians, it was know by the Persians, Chinese, Korean and Japanese.
      Type 1, of course, is uniformly fatal within a few years without the use of insulin.
      I helped treat a child, once, whose parents thought that vegetarianism (well, vegan ism) was the cure, and decided to manage themselves.
      After the child almost died (a second time), started losing feeling in their feet, dropped below the first centile weight-for-age and -for-height, not to mention a family court order for supervised injections at the risk of losing custody they realised that their child was happier when not starving to death (on the inside), slept better without urinating five times a night, and the doctors and nurses started working with their concerns about over medicalisation of their child rather than trying to take it off them.

      Also, vegetarian diets may be inherently low in cholesterol and fats, but carbohydrates such as starches (potatoes, flour/bread, etc…) and low fibre fruits (like almost anything selected from the fruit section of your supermarket) will cause sugar spikes, worsening diabetic control.

      Also take some B12 and Iron supplements otherwise your brain will rot and your bone marrow run dry. Just saying.

      1. A child, with type 2 diabetes? Wow, how does that work, lol? Yes, type 2 diabetes, where individuals choose a sedendary lifestyle, fake foods, and expect a pill or some other quick fix to solve their prblems, like all other modern, Western diseases that don’t exist in places like India and other developing nations.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.