01-27-2015 10:49 AM
50 12
tools
  1. DroidUser42's Avatar
    They had me do a special boot that disconnects it from the battery termporarily.
    Do tell.
    teemulehtinen likes this.
    01-19-2015 11:49 PM
  2. pil71's Avatar
    how did you do the special boot?
    01-20-2015 01:47 PM
  3. anonim's Avatar
    After having it register over 9,000 calories burned today (don't I wish?), I finally called support. They had me do a special boot that disconnects it from the battery termporarily. Low and behold, it seems to be accurately reflecting my heart rate again. They set up a follow up call for a few days from now to see if it's still working.

    I will say, support for the Band so far is substantially better than I've experienced with Microsoft on some other issues.
    Can you let us in on the secret method of resetting the band?
    01-20-2015 06:11 PM
  4. cj-m's Avatar
    Just a further addition - this is dcrainmakers response to the FitBit Surge HR Optical sensor, which could be applied to the Band as well.

    Ray Maker replied
    January 21, 2015 at 6:12 am#99
    It mostly comes down to two factors: The sensor itself, and the algorithms behind it.
    If you look at the Scosche, it uses a sensor from Valencell. That company - all they do - is work on optical HR sensors for other companies (such as Scosche). That means that they've taken years to perfect it - both the sensors themselves as well as the algorithms behind it for exercise.
    Meanwhile, companies like Samsung and many others, go and buy off the shelf sensors from companies you've likely never heard of. Though there are some bigger players, for example Texas Instruments. These sensors usually have very little R&D behind them, and are mostly designed for non-moving scenarios. Over the last 8-12 months, we've seen a flurry of entrants in the OEM market with these sensors. They in turn also tend to have almost no algorithms included with them - it's often raw signal (such as TI's case). So it's up to a given company (i.e a Fitbit) to sort that out.
    Now I don't know the arrangements that Fitbit has on their sensor, nor who made it (or whether that company supplied or assisted with the algorithms). I do know it's not Valencell or Mio/Philips (which are the main two players with reputable sensors for sport use).
    01-21-2015 09:26 AM
  5. mkg3's Avatar
    .....It mostly comes down to two factors: The sensor itself, and the algorithms behind it.

    .........These sensors usually have very little R&D behind them, and are mostly designed for non-moving scenarios. Over the last 8-12 months, we've seen a flurry of entrants in the OEM market with these sensors. They in turn also tend to have almost no algorithms included with them - it's often raw signal (such as TI's case). So it's up to a given company (i.e a Fitbit) to sort that out................I do know it's not Valencell or Mio/Philips (which are the main two players with reputable sensors for sport use).
    Clearly, it comes down to sensors and algorithm. That is given.

    The rest of the response from Ray Maker is just an assertion without substantiation. Its would be a mistake, I believe, to say its anything more than his opinion - nothing more.

    Does anyone really think that MS who spent 3 years developing Band spent no R&D into algorithms or finding appropriate sensor vendor? The inconsistent sensing may be due to simply where the state of the art is in optical sensors - which is a different issue than lack of R&D directly.

    While there may be some that slap HW together, most recognized companies are engaged in much work (read R&D) in wearables across the industry (smell of money). And FDA has device regulations to control what is being claimed additionally...
    01-21-2015 11:08 AM
  6. anon(5327127)'s Avatar
    People are people and, as they're people, they do people type things. ALL of us will have the band on differing clasp settings. Some tight, some loose and, as I'm sure that you're aware, your wrist can be different sizes depending on the time of day and what you're doing. This will obviously lead to different results for different people.

    ​Point made :)
    01-21-2015 11:45 AM
  7. gadgetrants's Avatar
    Does anyone really think that MS who spent 3 years developing Band spent no R&D into algorithms or finding appropriate sensor vendor? The inconsistent sensing may be due to simply where the state of the art is in optical sensors - which is a different issue than lack of R&D directly.
    People are people and, as they're people, they do people type things. ALL of us will have the band on differing clasp settings. Some tight, some loose and, as I'm sure that you're aware, your wrist can be different sizes depending on the time of day and what you're doing. This will obviously lead to different results for different people.
    As a quick follow-up to my earlier comment above regarding IR and bloodflow as the channel for reading HR, just wanted to note (based on my caveman level of knowledge) that the main challenge is this:

    The sensor shoots a beam of light into your wrist. Based on (1) the precise time of the beam origin, and (2) the delay in time at reflection and (3) pattern of light reflected back, the system then attempts to estimate whether your vein/artery/whatever is filled (i.e., your heart just pumped a beat...remembering of course that there is a delay between reading at the wrist and at the heart itself). It then repeats this detection process until a second pulse is measured ("filled" vein -> "empty" vein -> "filled" vein), and then it has a single-cycle estimate of HR. As we all know, a single-pulse cycle is a poor estimate, so additional cycles are better.

    Here is where things get tricky, and I can only speculate: apparently the IR recording is akin to a photograph. The problem is that if the sensor moves a few millimeters between readings, it's like shifting the camera to a new location and taking a second photo. Figuring out where the veins are and comparing them to the last reading is not trivial, and in fact some higher-end sensors do various tricks like (1) measure drift of the sensor over the skin surface to eliminate the interference, or (2) measure from two or more locations to "triangulate" the vein.

    tl;dr I'm reasonably convinced that the Band needs to be pretty tight to read HR. Anecdotal reports here are largely consistent with this.

    ​-Matt
    01-21-2015 01:01 PM
  8. Madame_X's Avatar
    Can you let us in on the secret method of resetting the band?
    Hold the power and action button down for 20-30 seconds, ignoring any messages that come up, until it vibrates and the screen goes blank. You will need to engage the charger to power it back on.
    gadgetrants likes this.
    01-21-2015 09:00 PM
  9. DroidUser42's Avatar
    Hold the power and action button down for 20-30 seconds, ignoring any messages that come up, until it vibrates and the screen goes blank. You will need to engage the charger to power it back on.
    Ah, I think that puts the unit back to "fresh out of the box".
    01-21-2015 10:41 PM
  10. jropelato1's Avatar
    Mine seems to work great
    01-21-2015 11:48 PM
  11. Madame_X's Avatar
    Ah, I think that puts the unit back to "fresh out of the box".
    No, it doesn't do a factory reset. Your data is still there. When you power it back on, the Band just thinks that the battery has died and asks you to confirm date and time.

    Sadly, even after resetting, mine seems to have a glitch with the heart rate creeping up and up. Every day I don't reset it, my daily calorie consumption is higher and higher, to ridiculously inaccurate levels. I have a follow up call scheduled with support tonight--I'm hoping that maybe they can help me flash the operating system.

    I really hate to think it might be a hardware issue since I know that getting a replacement would take a while.
    01-22-2015 01:13 PM
  12. mkg3's Avatar
    ..........Every day I don't reset it, my daily calorie consumption is higher and higher, to ridiculously inaccurate levels........
    Perhaps you shouldn't worry so much and burn more calories by increasing your heart rate; hence burning more calories...

    Or if you really mean consumption is up, skip the 9th snack of the day!!

    Sorry, it was too easy...
    gadgetrants likes this.
    01-22-2015 01:46 PM
  13. DroidUser42's Avatar
    The problem is that if the sensor moves a few millimeters between readings, it's like shifting the camera to a new location and taking a second photo.
    "A few millimeters" is quite a bit of movement. I doubt if mine is slipping one millimeter.

    I'm still of the theory that motion alters the blood flow enough to confuse things.
    01-22-2015 09:01 PM
  14. Madame_X's Avatar
    Perhaps you shouldn't worry so much and burn more calories by increasing your heart rate; hence burning more calories...

    Or if you really mean consumption is up, skip the 9th snack of the day!!

    Sorry, it was too easy...
    You got me.

    Sadly, Microsoft thinks I have a faulty sensor so they're doing a swap. So, I will be without my Band for the next week or so.
    01-22-2015 11:34 PM
  15. gadgetrants's Avatar
    "A few millimeters" is quite a bit of movement. I doubt if mine is slipping one millimeter.

    I'm still of the theory that motion alters the blood flow enough to confuse things.
    I wear it *tighter* when I run, but not enough to make my hand turn blue. As I've noted elsewhere, at least in my case when I hold my hand steady while running, the darn sensors responsible for gauging pace (i.e., accelerometer and gyrometer, I suppose) think I've stopped and my estimated pace slows to a crawl (though my little feets are still plenty busy). What's weird is that if I hold my arm steady while walking in the default mode, steps reliably register!

    Oh, I should probably mention that's *indoor* running, which the Band surprisingly gauges reasonably well even with GPS turned off. Perhaps running outside with a steady arm wouldn't "fool" the Band like it does indoors.

    -Matt
    01-23-2015 08:08 AM
  16. novaprime9's Avatar
    Almost certainly not, since the GPS would be on for that (or at least should be...) and then it won't matter too much if your hand/arm is still moving. The band will know it's moving based on location. :)
    01-23-2015 10:12 AM
  17. gadgetrants's Avatar
    Almost certainly not, since the GPS would be on for that (or at least should be...) and then it won't matter too much if your hand/arm is still moving. The band will know it's moving based on location. :)
    I have the exact same intuition. However, once we learned that the Band "uses your GPS-tracked walks/runs outside to calibrate the other sensors that estimate pace" I began to wonder how the so-called calibration algorithm works. Is it *always* collecting data and updating the calibration? Does it only use the first mile, or some other limited time/distance? Do you need to run around outside at different speeds, or can it accurately extrapolate from one pace to a wider range (which I think is mathematically quite difficult). Say you calibrate at a 9 min/mile -- after calibrating, will it estimate your pace during an 8 or a 12 min/mile (with GPS off) accurately?

    I realize this probably sounds like a largely hypothetical exercise, but it's a meaningful one for me as I do a considerable amount of running indoors, and I imagine others may also (e.g. treadmill). What's critical here (and to my amazement) is that after calibrating, you *can* do indoor running and the Band *will* do a respectable job estimating pace even when GPS is off. Though you then have the "dead-arm" problem I mentioned earlier.

    -Matt
    Last edited by gadgetrants; 01-23-2015 at 11:45 AM.
    01-23-2015 11:19 AM
  18. littlebib's Avatar
    Say you calibrate at a 9 min/mile -- after calibrating, will it estimate your pace during an 8 or a 12 min/mile (with GPS off) accurately?

    -Matt
    From the experimenting that I've done, I *think* that it simply looks at your last outdoor run as the baseline for your stride length. When I run faster, I don't necessarily move my feet faster, but I take a longer stride. Since my last GPS run was an average of about 8:15, then when I'm on the treadmill, regardless of whether I'm running at a 9:00 pace or a 7:30 pace, since I still have the same number of steps per minute, the Band thinks I'm running 8:15.
    01-23-2015 12:08 PM
  19. anon(8555314)'s Avatar
    The heart rate that shows on my band during a workout or run shows higher than the results when I look at the session later from the app. Today while walking to work I noticed a heart rate in the 130s, then later looked at it on the app and it showed a max of 129. I noticed the same thing on a recent cardio workout where I saw 190s on the band, then later on the app it showed a high of 177.
    01-23-2015 12:59 PM
  20. AndyCalling's Avatar
    Hmm. Perhaps this is why MS provide a Fitbit band with their premium phones rather than their own model. Still trying to make it work.
    01-23-2015 01:01 PM
  21. gadgetrants's Avatar
    From the experimenting that I've done, I *think* that it simply looks at your last outdoor run as the baseline for your stride length. When I run faster, I don't necessarily move my feet faster, but I take a longer stride. Since my last GPS run was an average of about 8:15, then when I'm on the treadmill, regardless of whether I'm running at a 9:00 pace or a 7:30 pace, since I still have the same number of steps per minute, the Band thinks I'm running 8:15.
    Same thing here. When I run indoors it's spot on if I'm running at my normal outdoor pace, but when I speed up the estimate falls off pretty quickly. I'd definitely add this issue to the list of things on the next firmware wishlist.

    -Matt
    01-23-2015 02:17 PM
  22. willgill's Avatar
    Has anyone seen mention of Microsoft Band allowing input from a heart rate chest strap?

    Their are several Bluetooth 4 LE chest straps on the market. Wouldn't it be nice to have the option of using a strap if you need high accuracy HRM for training?

    Sent from my SAMSUNG-SGH-I337 using Tapatalk
    01-26-2015 12:20 AM
  23. Madame_X's Avatar
    Received my replacement band today and this one is tracking where I would expect it to be. When I view the real-time heart rate, it locks on and stays locked. My first one would sometimes take a very long time to lock my heart rate, and even when it did it would only stay locked for a few seconds at a time. That leads me to believe there was something wrong with it from the beginning.

    Fingers crossed this one works better for me long term...
    gadgetrants likes this.
    01-26-2015 08:46 PM
  24. DroidUser42's Avatar
    Cool! Twofur fix! I wish I could be so lucky.
    01-26-2015 11:13 PM
  25. gadgetrants's Avatar
    That leads me to believe there was something wrong with it from the beginning.

    Fingers crossed this one works better for me long term...
    This leads me to wonder if you're looking at a newer production model (assumption 1), and if in fact they've made any design modifications since the first run (assumption 2).

    ​-Matt
    01-27-2015 10:49 AM
50 12

Similar Threads

  1. Restoring data from 920 to 930
    By Luis Sheppard in forum Ask a Question
    Replies: 8
    Last Post: 02-25-2015, 01:48 PM
  2. Replies: 8
    Last Post: 01-12-2015, 06:38 PM
  3. Replies: 2
    Last Post: 01-10-2015, 10:34 PM
  4. Grab the file to Bitmapimage
    By iice cage in forum Developers Corner
    Replies: 0
    Last Post: 01-10-2015, 08:10 PM
LINK TO POST COPIED TO CLIPBOARD