As I reported on Saturday, Apple changed the way the heart rate sensor works on the Apple Watch, changing a technical document that describes the feature, changing the way this feature is described for the Apple Watch.
Apple updated a technical document about how the heart rate sensor works. They are no longer saying that the Apple Watch reads your heart rate every ten minutes; instead, they say this:
Apple Watch attempts to measure your heart rate every 10 minutes, but won’t record it when you’re in motion or your arm is moving.
Here’s how that technical document looked on May 9 (via the Wayback Machine):
It’s also worth noting that Apple’s French version of this document still talks about 10-minute readings.
If you bought the Apple Watch as a fitness tracker, you may have wanted it to record your heart rate regularly in order to present you with a somewhat reliable estimate of the number of calories you’ve burned. Many users posting to a thread on Apple’s forums are quite angry about; this feature was one of the main reasons they bought the watch. The Apple Watch, without this feature, is little more than a fancy step counter, unless you start a workout with the Workout app.
I find this to be a bit ridiculous. As I’ve written here before, either Apple realized that the heart rate monitor simply isn’t reliable (as I’ve seen), or they wanted to save battery life. If the latter was their rationale, they should have an option to allow the user to choose whether they want 10-minute readings. As one user on the Apple forum said:
“I absolutely do not trust my rings to be accurate at all. And if they are not accurate or consistent, how can you even use them to monitor progress? This thing is basically worthless as a fitness device. And that was the main reason I purchased it.”
While knowing your resting heart rate is useful, and is a good way of determining your cardiac fitness, you don’t need to do it every ten minutes when you’re not moving; in fact, that is simply foolish in terms of battery life. You do, however, need to record the heart rate when someone is walking, even if they’re not doing a workout, in order to have an idea how active they are. If someone is walking to work, they’re not going to bother to start a workout; and they shouldn’t. But their activity should be counted as more than just steps. For now, that doesn’t seem to be the case.
I think the problem is in the accuracy of the heart rate sensor. As Apple’s technical document explains, the Apple Watch has two types of heart rate sensors.
Apple Watch uses green LED lights paired with light?sensitive photodiodes to detect the amount of blood flowing through your wrist at any given moment.
These sensors are used when you are in workout mode. But the device also has other sensors:
The heart rate sensor can also use infrared light. This mode is what Apple Watch uses when it measures your heart rate every 10 minutes.
While it’s a bit ridiculous that this part of the document still talks about 10-minute readings, what’s important to note is that these readings, which seem to be less accurate, are using a different sensor. It seems that Apple found these readings to be unreliable, and simply turned them off. (It’s worth noting that even with the 10-minute checks, I still had plenty of battery life at the end of the day, before the 1.0.1 update that turned off this feature.)
To many users, it looks like Apple pulled a bait-and-switch, promising a certain feature and not delivering it. Apple needs to say whether the change is because of faulty heart rate sensors – which means they have a bigger issue – or because of battery life. And if it’s the latter, they should allow users to choose whether or not the Apple Watch checks their heart rate every ten minutes. Let users decide how they want their battery usage to work.
However, if the heart rate sensors are faulty, simply turning them off, after promising this feature, is a mistake. They should fix them, whether through a software update, or by exchanging the devices. They promised a feature, and they can’t simply pretend that they never did so.
Note: one other interesting tidbit that an Apple forum poster pointed out, and that I confirmed on my iPhone’s Health app, is the fact that the Apple Watch doesn’t seem to be recording any heart rate readings below 61 bpm. I’ve had some much lower, which were incorrect, but a few around 57-59 which are probably correct. (Even though I’m not the fittest guy out there, my resting heart rate is fairly low.) Since the 1.0.1 update, I’m seeing that, every day, the lowest recorded heart rate is 61.
I’m endlessly entertained at the expense of the people who bought this ugly, pricey pig-in-a-poke!
I’m endlessly entertained at the expense of the people who bought this ugly, pricey pig-in-a-poke!
As you know I had doubts about this product and took the ‘wait and see’ approach. Boy, am I glad!
As you know I had doubts about this product and took the ‘wait and see’ approach. Boy, am I glad!
Hi Kirk
Thank you for you consistent updates on this issue. I use the activity app very day. However I treat it more like a guide and motivator so the level of Heart Rate precision is not my first priority. Also more importantly when I workout or run I use a Polar Chest strap monitor so the activity app is getting very accurate readings during that time period.
Have you been in contact with Apple Support on the Heart Rate issue? I think its very important that the continue to be aware of it. Have you tried to reach out to Jay Blahnik? He’s the director of Fitness Technology at Apple
Thanks again for the updates. I check your blog everyday
Ray G.
Norwalk CT,
Hi Kirk
Thank you for you consistent updates on this issue. I use the activity app very day. However I treat it more like a guide and motivator so the level of Heart Rate precision is not my first priority. Also more importantly when I workout or run I use a Polar Chest strap monitor so the activity app is getting very accurate readings during that time period.
Have you been in contact with Apple Support on the Heart Rate issue? I think its very important that the continue to be aware of it. Have you tried to reach out to Jay Blahnik? He’s the director of Fitness Technology at Apple
Thanks again for the updates. I check your blog everyday
Ray G.
Norwalk CT,
This is simply an accuracy issue. This is typical for all optical monitors of this type. Variations in the return signal from periodic blood flow need to be larger than variations due to motion of the watch/sensor position relative to the arm. How much motion you can tolerate depends on the strength of your signal, which is a function of transmitter power and receiver collection area and efficiency (this is why Apple uses larger lenses than most manufacturers) and scattering cross section in the tissue (function of diode frequency).
The Apple document makes it clear that it still makes a heart rate measurement using the low power IR mode every 10 minutes, so there is no battery life savings. It just doesn’t record the measurment if the motion sensor indicates that it might be inaccurate.
If you want accurate readings during motion, then you need to use the Workout app which turns on the higher power visible diodes. I presume that Apple will try to automate this feature and also try to guess your workout type from the motion sensors, but not in version 1.
No, it doesn’t record every ten minutes. It only does so if you’re not moving. So, yes, it does save power. But it gives very few readings during the day, which means that using the HR to calculate calories and activity is not possible.
How do you know it is not attempting a measurement every 10 minutes? You sound sure.
Apple’s technical document says that it is not. (It probably checks to see if there’s movement, and then aborts.) And data in the Health app shows that it’s not recording every ten minutes. Even when I’m at rest, it’s not recording every 10 minutes.
But it could be taking a measurement, deciding it is out of range, checking motion, and aborting. It’s hard to guess on exactly what the algorithm is doing.
For example, I’ve been wearing the watch when I sleep. I see that much of the time it does record every 10 minutes or so. Other times it skips up to an hour. Is it because I’m moving or because the watch is not positioned correctly and the reading is poor? I don’t know the answer, I thought maybe you had been given some additional information that wasn’t otherwise known.
If it’s able to take a measurement when you’re not moving, it has to check for movement. My guess is that it’s only the accelerometer that is involved at that stage. If so, the measurement aborts. There’s no point in starting a reading and then checking; that just wastes power.
Even though it says every 10 minutes, it’s not always taking measurements at 10-minute intervals, and this even before the update. There’s something else going on that is preventing measurements from being made. Lots of people are seeing this.
I’m going to stick with my guess that is likely a bad measurement that kicks off the check on the accelerometer for movement. It seems unlikely that I’m moving enough when I’m sleeping to miss several 10 minute intervals yet I see numerous gaps in the data.
You can stay with your guess. Until further information is available, your guess is as good as mine.
This is simply an accuracy issue. This is typical for all optical monitors of this type. Variations in the return signal from periodic blood flow need to be larger than variations due to motion of the watch/sensor position relative to the arm. How much motion you can tolerate depends on the strength of your signal, which is a function of transmitter power and receiver collection area and efficiency (this is why Apple uses larger lenses than most manufacturers) and scattering cross section in the tissue (function of diode frequency).
The Apple document makes it clear that it still makes a heart rate measurement using the low power IR mode every 10 minutes, so there is no battery life savings. It just doesn’t record the measurment if the motion sensor indicates that it might be inaccurate.
If you want accurate readings during motion, then you need to use the Workout app which turns on the higher power visible diodes. I presume that Apple will try to automate this feature and also try to guess your workout type from the motion sensors, but not in version 1.
No, it doesn’t record every ten minutes. It only does so if you’re not moving. So, yes, it does save power. But it gives very few readings during the day, which means that using the HR to calculate calories and activity is not possible.
How do you know it is not attempting a measurement every 10 minutes? You sound sure.
Apple’s technical document says that it is not. (It probably checks to see if there’s movement, and then aborts.) And data in the Health app shows that it’s not recording every ten minutes. Even when I’m at rest, it’s not recording every 10 minutes.
But it could be taking a measurement, deciding it is out of range, checking motion, and aborting. It’s hard to guess on exactly what the algorithm is doing.
For example, I’ve been wearing the watch when I sleep. I see that much of the time it does record every 10 minutes or so. Other times it skips up to an hour. Is it because I’m moving or because the watch is not positioned correctly and the reading is poor? I don’t know the answer, I thought maybe you had been given some additional information that wasn’t otherwise known.
If it’s able to take a measurement when you’re not moving, it has to check for movement. My guess is that it’s only the accelerometer that is involved at that stage. If so, the measurement aborts. There’s no point in starting a reading and then checking; that just wastes power.
Even though it says every 10 minutes, it’s not always taking measurements at 10-minute intervals, and this even before the update. There’s something else going on that is preventing measurements from being made. Lots of people are seeing this.
I’m going to stick with my guess that is likely a bad measurement that kicks off the check on the accelerometer for movement. It seems unlikely that I’m moving enough when I’m sleeping to miss several 10 minute intervals yet I see numerous gaps in the data.
You can stay with your guess. Until further information is available, your guess is as good as mine.
….so when you go to the doctor and they use the traditional bp cuff.. you need to be standing still, and your arm can’t move.. what’s the problem?
You really don’t get it? The reason you don’t move for blood pressure is so your blood pressure doesn’t change, not because of the measurement device. Here, the original point was that the heart rate measurement would help calculate your activity.
….so when you go to the doctor and they use the traditional bp cuff.. you need to be standing still, and your arm can’t move.. what’s the problem?
You really don’t get it? The reason you don’t move for blood pressure is so your blood pressure doesn’t change, not because of the measurement device. Here, the original point was that the heart rate measurement would help calculate your activity.
Sorry but I think that you are overthinking this without getting more details and perhaps longer term plans. I read that it is trying to read every 10 minutes but just trying to avoid cases that could give false readings. Most people will be fine.
And too be honest, if I really need a more accurate and faster monitoring I will be in exercise mode.
I suspect in reality, that this will come to not much at all.
Making a headline that says ‘bait and switch’ – sure you aren’t baiting people for link hits versus accurately investigating and reporting?
If you read the article, and the other articles I’ve written, which I link to, you’ll see that I have researched this quite a bit. The device was supposed to use 10-minute readings to give an accurate calorie count, and activity measurement, even when you’re not in a workout. If it were to be any better than other fitness trackers, it would have to do that. It’s not trying to read every ten minutes, and this is very clear if you read Apple’s document. If you’re moving, it simply doesn’t try to measure.
Sorry but I think that you are overthinking this without getting more details and perhaps longer term plans. I read that it is trying to read every 10 minutes but just trying to avoid cases that could give false readings. Most people will be fine.
And too be honest, if I really need a more accurate and faster monitoring I will be in exercise mode.
I suspect in reality, that this will come to not much at all.
Making a headline that says ‘bait and switch’ – sure you aren’t baiting people for link hits versus accurately investigating and reporting?
If you read the article, and the other articles I’ve written, which I link to, you’ll see that I have researched this quite a bit. The device was supposed to use 10-minute readings to give an accurate calorie count, and activity measurement, even when you’re not in a workout. If it were to be any better than other fitness trackers, it would have to do that. It’s not trying to read every ten minutes, and this is very clear if you read Apple’s document. If you’re moving, it simply doesn’t try to measure.
Sounds like an easy enough solution would be to have the visible/green/more accurate sensors kick on if the motion detectors detect motion, and perhaps have this be an option that those totally uninterested could turn off. I don’t use the watch for fitness tracking but agree that monitoring basic activity separate from a workout should be possible and isn’t being done effectively if your heartrate is only being recorded when at rest. Like a previous poster said, monitoring your heartrate as you walk to work, climb the stairs to your next meeting, etc., shouldn’t require additional effort (i.e., starting Workout).
Sounds like an easy enough solution would be to have the visible/green/more accurate sensors kick on if the motion detectors detect motion, and perhaps have this be an option that those totally uninterested could turn off. I don’t use the watch for fitness tracking but agree that monitoring basic activity separate from a workout should be possible and isn’t being done effectively if your heartrate is only being recorded when at rest. Like a previous poster said, monitoring your heartrate as you walk to work, climb the stairs to your next meeting, etc., shouldn’t require additional effort (i.e., starting Workout).
One other note. Checking my overnight wearing of the watch, I see heart rates lower than 61 bpm so I don’t think your final note is correct. Specifically I see 59 bpm several times.
Interesting, thanks for pointing that out. I see a flat line in the Health app at the lower end, and when I look at the data, 61 is the lowest; as I said in the article, someone on the Apple forum said this too. It seems like an odd limit, but I’ve not worn it overnight to find out. (Perhaps I’ll try tonight; but the idea of sleeping with a watch is odd…)
I was curious about my sleeping heart rate. My resting heart rate during the day seems a bit high but I take some fairly serious blood pressure medicine so that might affect it. I was happy to see it drop frequently overnight into the low 60 and below range.
I’m seeing numbers down to the low 40s (which is correct for me). Not sure why you aren’t seeing anything lower than 61.
You’re not sleeping with a watch. You’re sleeping with a fitness tracker.
One other note. Checking my overnight wearing of the watch, I see heart rates lower than 61 bpm so I don’t think your final note is correct. Specifically I see 59 bpm several times.
Interesting, thanks for pointing that out. I see a flat line in the Health app at the lower end, and when I look at the data, 61 is the lowest; as I said in the article, someone on the Apple forum said this too. It seems like an odd limit, but I’ve not worn it overnight to find out. (Perhaps I’ll try tonight; but the idea of sleeping with a watch is odd…)
I was curious about my sleeping heart rate. My resting heart rate during the day seems a bit high but I take some fairly serious blood pressure medicine so that might affect it. I was happy to see it drop frequently overnight into the low 60 and below range.
I’m seeing numbers down to the low 40s (which is correct for me). Not sure why you aren’t seeing anything lower than 61.
You’re not sleeping with a watch. You’re sleeping with a fitness tracker.
You should be more careful with your “Yellowish” headlines. They could get you into trouble if Apple’s Legal Counsel decides to challenge your “Bait and Switch” explicit accusations.
It not only seems that the company is not engaging in any bait and switch sales tactics but making every effort to find an optimum solution to any issues around the heart meter of the watch.
In any case, if Apple is not able to find a adequate software solution, all buyers that bought the watch under those specifications or even just the expectation of that feature arising from any promotional material from the company, will be given the choice to receive a full refund or have the watch exchanged for an upgraded (hardware) model.
“Bait and Switch” and what I have described in the previous paragraph lay in opposite extremes of Business Ethos and Conduct.
You should be more careful with your “Yellowish” headlines. They could get you into trouble if Apple’s Legal Counsel decides to challenge your “Bait and Switch” explicit accusations.
It not only seems that the company is not engaging in any bait and switch sales tactics but making every effort to find an optimum solution to any issues around the heart meter of the watch.
In any case, if Apple is not able to find a adequate software solution, all buyers that bought the watch under those specifications or even just the expectation of that feature arising from any promotional material from the company, will be given the choice to receive a full refund or have the watch exchanged for an upgraded (hardware) model.
“Bait and Switch” and what I have described in the previous paragraph lay in opposite extremes of Business Ethos and Conduct.
I am reporting on my first day of wearing the watch but have been keeping up with your reports since the primary reason for my purchase if for fitness tracking. I’m not interested in Fitbit like tracking during the day so much but primarily for workouts (running, cycling).
Initial observations:
1) My heart rate is in the low 40’s. The watch is accurately reporting this so I don’t see the issue about going below 60 bpm.
2) I just completed a 6K run while also wearing a strap to compare HRM accuracy. I’m happy to report that they were almost identical throughout the run. In fact, the Wahoo had a few spikes that were obviously not real that the watch didn’t. During the run I was concerned that the watch wasn’t tight enough since I could feel it moving a little but that didn’t seem to cause any problems.
My main complaint is that I want to continue to use my tracking app (iSmoothRun) instead of the built in one. But until a native app is available for the watch I have to use both the included Workout app and iSmoothRun at the same time and I don’t get the heart rate curve when I upload my workout to Runkeeper.
Based on your comments I feared I would be returning the watch and buy a Garmin running watch instead but at first use it doesn’t appear I’ll need to do that. My rationale for the Apple Watch instead of the Garmin was to get my music directly from the watch without an additional device. Ideally I’d like music, HRM and GPS in one but for now the Watch is the only device that I know of to give 2 of the 3 and I can have the GPS when I want by bringing my iphone.
Btw, I’ve gone thru 4 HRM straps in 18 months so another motivation was to never purchase another one. At $75 each x 4 the price of the watch just for HRM alone seems reasonable
Regarding the measurements during non-workout periods I guess I look at consistency as the most important aspect. Just to see that things are improving over time. I don’t buy that any measurement of calories, etc are accurate (everyone’s metabolism, etc are different so it’s all just estimations based on averages anyway). Also, to really do these sorts of calculations more scientifically it would take more data that just heart rate. For every 10 minutes a reading is taken I would think you would also need to log what you were doing at that moment for it to be of much value at that level of granularity.
I am reporting on my first day of wearing the watch but have been keeping up with your reports since the primary reason for my purchase if for fitness tracking. I’m not interested in Fitbit like tracking during the day so much but primarily for workouts (running, cycling).
Initial observations:
1) My heart rate is in the low 40’s. The watch is accurately reporting this so I don’t see the issue about going below 60 bpm.
2) I just completed a 6K run while also wearing a strap to compare HRM accuracy. I’m happy to report that they were almost identical throughout the run. In fact, the Wahoo had a few spikes that were obviously not real that the watch didn’t. During the run I was concerned that the watch wasn’t tight enough since I could feel it moving a little but that didn’t seem to cause any problems.
My main complaint is that I want to continue to use my tracking app (iSmoothRun) instead of the built in one. But until a native app is available for the watch I have to use both the included Workout app and iSmoothRun at the same time and I don’t get the heart rate curve when I upload my workout to Runkeeper.
Based on your comments I feared I would be returning the watch and buy a Garmin running watch instead but at first use it doesn’t appear I’ll need to do that. My rationale for the Apple Watch instead of the Garmin was to get my music directly from the watch without an additional device. Ideally I’d like music, HRM and GPS in one but for now the Watch is the only device that I know of to give 2 of the 3 and I can have the GPS when I want by bringing my iphone.
Btw, I’ve gone thru 4 HRM straps in 18 months so another motivation was to never purchase another one. At $75 each x 4 the price of the watch just for HRM alone seems reasonable
Regarding the measurements during non-workout periods I guess I look at consistency as the most important aspect. Just to see that things are improving over time. I don’t buy that any measurement of calories, etc are accurate (everyone’s metabolism, etc are different so it’s all just estimations based on averages anyway). Also, to really do these sorts of calculations more scientifically it would take more data that just heart rate. For every 10 minutes a reading is taken I would think you would also need to log what you were doing at that moment for it to be of much value at that level of granularity.
The heart-rate-below-60 issue does not exist. I just looked at my watch and the latest reading (5 minutes ago) was 49bpm. This is about right for my resting heart rate. I see other, similar readings in the Health app.
The heart-rate-below-60 issue does not exist. I just looked at my watch and the latest reading (5 minutes ago) was 49bpm. This is about right for my resting heart rate. I see other, similar readings in the Health app.
So, just an update on this as a developer. I too found with WatchOS 1.0.1 that my heart rate was not being measured for sometimes up to 9 hours of a day which is ridiculous. I am also developing a heart rate app at the moment to view graphs of resting heart rates etc and found this to trash my trends! HOWEVER, I am pleased to report that with the latest WatchOS 2 beta 5 released last week, I am finding my watch is measuring a lot more often. It’s seems to be every 10 minutes most of the time. For the times when it isn’t I think two things have happened. They’ve done increased the ‘your moving’ threshold to prevent a recording and they have also (I think) got the watch to try again each minute after a failed attempt rather that waiting the full 10 minutes. I am seeing better battery life with this watchOS than 1.0.1 too so hopefully everything is on track! I hope this is the news u all wanted to hear. Incidentally, if any of you are interested, I plan at the end of this month to run a test flight beta of my app which displays all this data in graphs, it’s a free way for you to test it it you want. If your interested then let me know!
So, just an update on this as a developer. I too found with WatchOS 1.0.1 that my heart rate was not being measured for sometimes up to 9 hours of a day which is ridiculous. I am also developing a heart rate app at the moment to view graphs of resting heart rates etc and found this to trash my trends! HOWEVER, I am pleased to report that with the latest WatchOS 2 beta 5 released last week, I am finding my watch is measuring a lot more often. It’s seems to be every 10 minutes most of the time. For the times when it isn’t I think two things have happened. They’ve done increased the ‘your moving’ threshold to prevent a recording and they have also (I think) got the watch to try again each minute after a failed attempt rather that waiting the full 10 minutes. I am seeing better battery life with this watchOS than 1.0.1 too so hopefully everything is on track! I hope this is the news u all wanted to hear. Incidentally, if any of you are interested, I plan at the end of this month to run a test flight beta of my app which displays all this data in graphs, it’s a free way for you to test it it you want. If your interested then let me know!
Hi,
Thank you for the post. I just wanted to say that the lack of continuous measurements ended for me after the Watch OS 2 update. Right now I have many THOUSANDS of daily measurements, nearly accounting for EVERY minute of my day. In fact, for most minutes I have tons of sequential measurements (10-100!) as well. Rarely do I see the every 10 mins intervals, if I do it’s for an about an hour, then the many measurements per minute period continues for hours.
So all is fixed now, at least on my end. 🙂
Hi,
Thank you for the post. I just wanted to say that the lack of continuous measurements ended for me after the Watch OS 2 update. Right now I have many THOUSANDS of daily measurements, nearly accounting for EVERY minute of my day. In fact, for most minutes I have tons of sequential measurements (10-100!) as well. Rarely do I see the every 10 mins intervals, if I do it’s for an about an hour, then the many measurements per minute period continues for hours.
So all is fixed now, at least on my end. 🙂
on Watch OS2 the Healts App shows results every 10 seconds
on Watch OS2 the Healts App shows results every 10 seconds
I’m fairly certain they enforce a max rate. 215 is the max regularly historically implying some readings are inaccurate (I hope) and that is the max constraint. My min however regularly drops to 40 or even the thirties.
I’m fairly certain they enforce a max rate. 215 is the max regularly historically implying some readings are inaccurate (I hope) and that is the max constraint. My min however regularly drops to 40 or even the thirties.