Fitbit Sense heart rate is way off

[MarkLiberman]

Just got a new Sense. When I exercise to raise my heart rate, 3 independent monitoring methods all show (for example) a rate of 140-145 -- the sensor on the elliptical, the readout from my Orange Theory "Burn" device, and a fingertip blood oxygenation monitor. But the (apparently worthless) Sense says "89".

I've tried multiple times, different wrists, different placements -- essentially the same result.

In contrast, my Ionic was pretty good at this -- the only problem was that the battery basically died, would only last a few hours.  Overall, Fitbit seems to be making crappy devices these days.

Moderator edit: format

[arturmeireles]

@t.parker I'm going to the gym right now to train chest and I'm going to use my Fitbit Sense, Apple Watch 6 and the Polar H10. After I finish I'm going to post here print screens of my workouts on all 3 devices 🥳💪🏻

[t.parker]

@arturmeireles great! 🙂 Lucky you, the gym in your place is open 🙂 In the UK, I'm in Tier 4 (pretty much lockdown) so all the gyms are closed (lucky, I have home mini-gym).

@Mateos  Ionic for me was just horrible not only when weight lifting but in general. Running, cycling - the results most of times were wrong, very often it couldn't read my HR (two dashes). For weight lifting, Ionic was as bad as reported for Sense - too low HR. Another weird thing about Ionic was that when I was walking it was reporting inflated HR (sometimes 20-30bpm more than real one). I'm not trying to defend Sense here. It may be that with my experience, my expectations are just very low, hence results I got surprised me.

Just edit: I've done 5k run but now I can't sync. I don't know whether to laugh or cry here 😄 I'm wondering whether Fitbit will ever fix syncing issues. Last successful sync at 8:33am.

[t.parker]

Finally, I managed to sync Sense (since I stopped using Fitbit such issues were long forgotten). Well, here's some more data.

First of all a word of explanation about PolarH10 results. The drops in the graph are caused by dying battery. Ant+ connection was breaking during my run, however, I decided to use it anyway and apart from those drops, the HR capture is accurate.

Run 5k.

3 different intensities driven by cadence of:

160spm

170spm

180spm

In general, it was pretty much steady-pace run. Due to temperature outside (-1C) it wasn't easy to elevate my HR.

HR was captured by PolarOH1 (connected to PolarBeat app), PolarH10 (connected to Garmin watch) and Fitbit Sense.

PolarOH1 and H10 seem to have pretty much same level of accuracy (despite dying battery in H10):

Purple and green graphs overlap pretty much all the time. The green (OH1) goes a bit beyond as I turned off recording on my phone a bit later.

Now, let's add Sense to the group:

This is kinda surprising. It's a mess however somehow it follows overall trend. The dip between 20:50 - 25:00 is visible (btw, it was interesting experiment for GPS which Sense failed - stepping in place adds distance despite connected GPS). Also, very often Sense HR is a lot higher than the real one. I expect the cadence may have something to do with it:

The cadence is one of the factors that may affect wrist OHR. Despite trying to keep cadence fixed (3 ranges, orange, green, blue) it is visible that there is some noise. If this noise is not being filtered out by the Sense algorithms it may affect the results (steps may be read as heart beats). I have seen that before, too. Ionic was very sensitive to cadence noise.

	Polar OH1	Polar H10	Fitbit Sense
AvgHR	161	162	165
MaxHR	175	175	179

It isn't that far off for this kind of run. I guess, for casual runner it will do. Training with HR zones in mind is a different story though. If I am right about cadence noise, then my run shows better results as I am running with particular cadence rather than pace. The cadence noise in my case would be reduced due to my running style. More "chaotic" runner may experience worse HR readings.

Still, I have other activities to test and the test of the GPS is on the way (surprisingly accurate when it comes to the map but then the distance and pace computation is just wrong - new device, same old story).

[Mladen90]

@t.parker 

Thanks for all the analysis...i think that none should consider an optical sensor for serious/professional training.

I'm far from being an expert but the sense(and your data seems confirming it) seems perfect for a casual person that want some data and is doing some HIIT/running constantly.

[t.parker]

@Mladen90 I think optical sensor like OH1 should do the job (see the graphs, it's overlapping with chest strap). I would say that wrist optical HR is no way to go when it comes to serious training. Wrist is probably worst possible place to wear the sensor. The best for Fitbit would be if the devices received compatibility with external sensors to give users choice. It would be easier to admit that in some cases wrist OHR is not enough. But right now, Fitbit cannot say "yes, weight lifting due to its nature may show more inaccurate readings" because there is no alternative given to the users. I don't think that external sensors compatibility is going to happen though, unless Fitbit will release own branded chest strap.

[Mladen90]

@t.parker 

yeah my bad...i said "optical sensor" but i was thinking more generally about smartwatches or smartbands(used as watch) or as you specified "wrist optical HR".

Thanks again for all the tests 😁

[slowdive]

I need something in my wrist so I avoid going to high HR zones. I'm clear to do all kind of effort (green light from two cardiologists and several exams) but if during a class I stay in high hr zones the next days I get a small pain in my chest. That's why I need something close to my polar h10. Polar h10 is fine when I'm doing exercise by myself because I can keep the phone always on. But I'm in a class it's more complicated. The ideal would be fitbit recognizes polar h20 like many smart watches do.

For example, yesterday, while on my rpm class I sense that my HR was way to high. Looked at the watch and it showed me 100. After ending the class I looked at the polar results and by that that time my HR was about 150. I need something reliable to slowdown when I reach high hr to avoid the small chest pains. With sense I can't. With polar hr10 in classes it's complicated since I have to look all the time to my smartphone

[t.parker]

@slowdive what you need is a watch that can connect to external sensors. Lots of sports watches can do that. Beating chest strap accuracy and response time is very hard for any wristwatch.

[Mateos]

Hi t.sparker, interesting analysis. Would love to see a training with more interval. From what I have seen (and saw in the analysis of reviewers on YT) is that the Sense eventually reaches an elevated HR level but that there is a serious lag. More so than for example with the Charge 4 or other watches that they compared the Sense with. 

I am not technically as savvy as you are (with data esports and so on) but I would be interested to see if you come to the same conclusion. In your training sessions you have gradually increased your HR but I wonder what happens if you introduce more intervals. In the first training you posted there was interval at the end and there Inthink you can see the lag as well. 

[slowdive]

Yes, I know. Been searching for that but apart from sport watches like polar or garmin (either too expensive and made for professionals athletes or cheap with issues and basic smartwatch features) I have no idea if other watches will pair with h10. If only sense could do that. 

I have a small chest pain because sense showed me my HR was 100 when it was around 150. I had no idea on what hr zones I was and now my chest hurts. If it monitored correctly this wouldn't happen. 

Sense got to go.

[t.parker]

For last week I could barely find time to do any serious workout but coming back now. So, this time it's going to be quick. Two types of activities - running (5k, fixed cadence), and very short bodyweight session pretty much to cool down, done right after the run.

The run:

Orange - Fitbit Sense, Blue - PolarH10 (with new battery :))

Bodyweight workout (tracked as "Workout"):

Orange - Fitbit Sense, Blue - PolarH10

As much as the run keeps some trend (well, lets say it does), the "Workout" HR is an utter rubbish. Within those 13 minutes Sense where never correct (maybe only during cool down at the very end). I wasn't expecting it to be that bad. The watch was worn as recommended and since it is so small and light, it stayed at the same position throught entire workout. Despite my previous test, which came up better than I expected, this one is a massive disappointment. The HR indeed is very bad. There is no consistency in the result, it's not like there is a delay and watch has to catch up etc. It just reads invalid data. Mind that my short workout wasn't very intense (it was more TUT oriented). It was:

(20x squats

 20x bulgarian squat L 

 20x bulgarian squat R

 10x squat&jump

 30sec rest) x 3

cooldown static stretching

On the blue graph the 3 peaks are result of squat&jump.

I will try to find time to do proper weight treining, although gyms here are still closed and my equipment at home is kinda limited right now.

For a while, I really thought I would like this watch (especially after Ionic, a big disappointment).

[Mateos]

Again proof that something is very wrong with the HR sensor. How many more examples and user stories does Fitbit need to give a reaction? Many owners have spend more than €400.- for the watch including premium subscription and have ended up with a useless device (when it comes to activity tracking). I think we are at least entitled to a statement by Fitbit on how they think they are going to solve the problem.  

[t.parker]

Back with one more test. This time there are more "contestants".

PolarH10 (chest strap, connected to PolarV650 bike computer)

PolarOH1 (connected to MapMyFitness app)

Garmin Fenix 6 Pro (using wrist HR, usually I don't use WHR but for sake of test)

Fitbit Sense (using wrist HR)

The activity is 30min indoor cycling. On Fitbit Sense tracked as "Spin". There are 3 peaks which include uphill intervals. Most of the time, HR should stay in the low-aerobic zone. 

Colors:

Yellow - PolarH10

Red - PolarOH1

Green - Garmin Fenix 6 Pro

Blue - Fitbit Sense

Firstly, the gold standard of the fitness industry, PolarH10:

 PolarH10 is pretty much my reference for good HR.

Next PolarOH1. Optical but stays very close to H10, lets call it "silver standard":

 It isn't perfect but acceptable, with no unusual drops or spikes. OH1 follows H10 pretty close.

Now Garmin Fenix 6 Pro:

 The beginning isn't too great. I'm not sure why. Most of the time, the result is acceptable. However, this beginning is one of the reasons why I don't use WHR. My watch is paired with H10 or OH1. Keeping HR zones in mind, I would not train using the watch HR.

Now, time for Fitbit Sense:

Similar, 'take-off' to Garmin. Overall, it isn't horribly bad but from all the sensors, Sense is the least accurate.

Just for comparison,

H10 vs Sense:

H10 vs Garmin F6 Pro

And wristwatch stand-off, Sense vs F6Pro:

Final conclusion?

For this kind of "cardio" activity Sense may be "good enough". It follows the trend and despite it lacks the accuracy of competitors, if somebody does cardio to burn some calories I believe this HR data will provide a good approximation. PurePulse needs an upgrade.

[Sam.Gal]

Has there been any kind of address or news given on this shortcoming? There are reviews spanning back from September discussing the faulty HR tracking on the Sense and Customer Support seems to still be in denial or clueless as to the problem. My theory is that they just waiting until after the window to return the device from Christmas closes before they tell everyone they are out of luck and this is a hardware issue. Totally fine if you are a casual user or only do cardio, but a complete embarrassment of a fitness tracker otherwise.

Fitbit owes us answers.

[t.parker]

Here's my last attempt. This time low intensity hike. 5km, 52min. Sense vs PolarH10:

Pink - Sense, Blue - PolarH10

Pink - Fitbit Sense

Blue - PolarH10 (chest strap)

It may come as a surprise that this time Sense registers higher HR than chest strap. I am not surprised though. I have seen this behaviour on Ionic. In fact, Sense behaves no different from Ionic in my case. It may be just me but for low-intensity activities like walking, it captures way higher HR but for high-intensity sports it measures lower HR.

The average HR:

Sense: 112

PolarH10: 100

This results may be more important for a casual Fitbit user. Why is that? I have reasons to believe, that majority of users are into step-based, rather low/medium intensity activities and the main goal is getting into shape (and very often losing weight). It means that average user will care about the energy expenditure and the major daily activity is going to be walking.

It's pretty much it. Sense HR needs improvement. I doubt however it's going to happen. Here's why. I have already seen very same issues when I had Ionic. Now, it's like deja vu. Fitbit had plenty of time to fix HR, yet the new device comes with similar problem. It is probably the last test I've done as I have decided to return the device. Maybe next one will be better.

[DSKFit]

Thank you for confirming that you already tried to restart multiple times. I've gone ahead and sent your information to the Customer Support team for further assistance, so you should be getting a reply soon.

I hope everything goes well, keep me posted.

Put me on the list too. I just bought a Sense 2, took it to the gym for the first time and... what a disappointment 😞 😪