Exercise HRM WAY too low, if auto-started, anyone else?

[PeteG-1]

I've only had my Ionic for about a week, moving from a Charge 2, my first.

For whatever reason, my Ionic seems to be doing an absolutely ridiculous job of tracking my rides.

I'm an ex-racer, fairly high intensity, although recovering from a surgery, so maybe a hair slower than a couple of months ago.

That said, I'm out doing rides at pretty high-intensity (rebuilding base), 80-90%, steady, and my HR is about 40 beats too low, the entire time (except for one small spike, at the end of both rides, and today's ride had me descending to the end, about 1/2 mi, so my HR should have been low at that point).

I pulled up a couple of my Charge 2 logs, some of the same exact rides from past weeks, and yep, it's ranging from 40-65 BPM different.  
The only comical part of the whole thing, is that since having my Fitbit (last spring), it's always telling me I'm riding too fast, about 90-95% of the time, for "fat burning".  Now it's telling me I'm riding too slow, hmm... 

Edit: I went back, and I had a couple of GPS/manual-started rides, and those are looking good, or close to my Charge2 values. 

If you compare the auto-start to manual, it's borderline ridiculous, 30-45 BPM less, on average for the same ride (close, same distance, routes, speeds, within 5% or so). 

I wonder how I attained TDF level VO2 levels, in a day ;-]

[PeteG-1]

Bumping my own thread, hoping someone else has gotten somewhere, anywhere, or perhaps one of the Fitbit support people will chime in.

I tested this, yet again, same deal, 96 vs 122 BMP, same 1:40 ride, same 17.5-17.8 avg (slow winter miles), why is the auto-start ride 26 BPM lower?  

I've seen even bigger margins, going for a walk, I can't get much above 80BPM, if auto-started, but if manually I get "believable" readings, say in the 100-105 range...

[bbarrera]

@PeteG-1 good luck, sounds like putting Ionic into cycling mode is the workaround for now. I don't have any issues with a chest strap and still use that for various reasons (like tracking Pwr:HR). If I did have issues with a chest strap, would use Scosche Rhythm+ or the new Wahoo Tickr Fit that was launched this week. Did you read the DCRainmaker review of Ionic? His testing on road bike was done with older firmware, but just reaffirms my belief that optical HRM on the wrist isn't ready for prime time if you want accurate data. Over time I've probably compared close to 100 rides, optical HRM versus chest strap. In my experience (Blaze, Apple Watch), its more hit than miss, but enough misses versus chest strap that its just not worth the aggravation of looking at obviously wrong data - and Murphy's law, when I most interested in HR, thats when optical HRM loses the beat for a few minutes!

Last year's base building was done using CTS plans from Strava (premium) - worked well and I held good fitness level thru Sept. My road bike turned two a month ago, haven't really ridden since teen years. This year I decided to try a more structured approach and signed up for TrainerRoad. On week 4 of the high volume sweet spot base I plan, which is 9+ hours this week. Most of the rides are 88-94% ftp (sweet spot), had first Tempo ride yesterday at 75-90%. Really like TrainerRoad so far, nice steady fitness ramp at 'easier' intensities to ensure I get those physiological adaptations without having to ride 12-20 hours a week. 

[PeteG-1]

Yeah, that's been my experience too, the hit-miss, but not THIS far off.

I'm not racing anymore, and I just want to track my general HR, for recovery tracking and such, plus to see how my HR matches, say to sprint-sections on a ride (how far into the red am I going ;-]).

I pulled out my old Polar, when I ran into some similar issues with my Charge2, but after a few times, I decided it just wasn't worth it, I didn't need another "full device", I'd just be good with the 95%-ish accuracy of my Fitbit tracking, while pedaling fast.

I'm not very structured, anymore, I mostly do 2-4 rides/week with a pretty fast group (mix of active 3-4-5 racers, and the other half like me, no longer racing or taking time off), so I meter that with one or two easy rides or recovery walks, and a longer weekend ride (once springtime swings around).

My HRM is handy though, say if I go out for 2 really fast weekday rides, 40-60mi/ea, and I notice on my recovery day that I'm "overtrained", by keeping an eye on my Fitbit, I can back off for whatever I'm doing that day, and make it a "better/actual" recovery.

I had to have part of my hip fused recently, and I'm just easing back into it now, after a forced 8 weeks off, I'm up to 30-40mi rides, and about 85-90% intensity, max.  I hope to be up to 30-50 and 85-100, by EOM, depending on how much skiing...

[bbarrera]

well in my experience, Fitbit (Surge and Blaze) is worse than Apple on tracking HR on sprints. Maybe Ionic is good enough. I think "95% accurate" is misleading because the interesting HR stuff is the stuff that Fitbit historically has missed (lag). We aren't talking about HR average for a ride.... Also speaking from my own personal experience (not yours of course), tracking RHR has been a poor indicator of "overtraining" because by the time my RHR jumps I'm overcooked and common sense (completely exhausted by 8pm, and a couple other things) told me that a day or two before. But everyone is different, maybe RHR tracking works for you.... Tracking HRV when I wake up seems to be better indicator of increasing training stress, but Fitbit doesn't provide that. After following CTL ramp rates and the TSS for each ride over the last year on TrainingPeaks, I can pretty much dial in recovery days and training intensities based on current cumulative load, and a look at upcoming group rides (Wed night hammerfest, weekend climbing rides, and once a month century rides). 

[PeteG-1]

Yeah, I don't tend to use my RHR for "overtraining" as an indicator, I use my current HR, when going out for a recovery ride, or fast walk.  

IME, it's usually pretty obvious, if you're overtrained, if you're "spinning easy" on the flats, and up at 88%, you need some serious recovery ;-]

My training "regimen" is loose, but pretty solid, I could probably drop-in to a cat4 race still, during the summer months, and at least hang on, possibly be somewhat effective (not really training for that anymore, but I ride with fast groups, mostly).  I figure that's good enough.

I did some Randonneur stuff awhile back, and boy was that much trickier to train around (in particular, recovery, I think most serious rando guys are almost always overtrained - a friend who qualified for RAM was always like that, 500-800mi weekends, yeah right...).

[bbarrera]

@PeteG-1 Same here, I could drop into a cat4 race and probably hang on (no hills, too heavy!). The local Wed night has a collection of Cat 3, Cat 2, and cyclocross racers, so its usually a hammerfest!

[PeteG-1]

Yeah, not sure I'd be in there, with the Cat2 riders. 

I rode with a similar group (mostly 2, a few 3s) back in the mid-90s, when I was racing 4s.  Things were usually do-able, as long as they weren't really doing a fast-tempo ride, or a particularly hilly one (that's where cat2 riders being to seem super-human, IME).  It was fun, but I don't think I'd attempt it nowadays.

My group of mostly ex 3-4-5s, and some active, and then a few older, retired pros, is nice.  We're safe, things are plenty fast enough (we have sprint sections and the like), and summers we break into a full-on-hammerfest group and a group2, so there are choices to really bump it up, if desired (some of the summer riders are of the ability of cat 1-2 riders, national-ranked rando-riders, etc).

Fortunately (or not) with this group, a lot of us are aging together, the core anyway.  Then spring comes, and we pick up 5 new active, 20-something racers, and things get "frisky" until we separate.

I'm debating, if I should call Fitbit, and try to get them to look at my side-by-side logging, for rides that were auto/manually started, and consistently have the huge delta.  Seems like something they should be fixing, although my inclination is that they're aware.

[SimonRJ]

Based on my perceived level of experience you guys hold in the cycling realm, would it be safe to say that you already knew that your Fitbit wouldn't be as accurate as your chest strap?

I certainly did and held no high expectations.   Of note, while I'm on my trainer (wahoo kickr with zwift and trainer road, my ionic is almost spot on with my tickr HRM.   As soon as I ride outside, accuracy seems to suffer.   

Based on similar observations made while jogging, I'm led to believe there are still some shortcomings with the Fitbit algorithm that processes heart rate while the gyro sensors are active trying to filter out interference from arm movent.   

At the end of the day, I'll always trust my strap that I always wear on a ride anyway.  

[bbarrera]

@SimonRJ yes, primarily from DCRainmaker reviews I'm aware that all interesting wrist wearables have problems tracking HR while cycling outside. In addition I've also done a lot of my own testing, comparing chest strap to Surge, Blaze, and Apple Watch. The wrist is one of the worst spots to track HR, but best spots for quickly interacting with a wearable. And while the accuracy of optical HRM on the wrist has gotten better, its still annoying when its wrong.

Lack of support for chest strap and/or HR/GPS import is the reason the Force was my last Fitbit tracker (my iPhone is current step tracker). The Surge was a disaster in my personal opinion, the Ionic has promise but at current pace of software it may take years (and new hardware with a speaker/mic). I'm going to buy & try the Ionic sometime this year, once it moves past the current beta quality firmware phase.

[PeteG-1]

---

@SimonRJ wrote:

Based on my perceived level of experience you guys hold in the cycling realm, would it be safe to say that you already knew that your Fitbit wouldn't be as accurate as your chest strap?

I certainly did and held no high expectations.   Of note, while I'm on my trainer (wahoo kickr with zwift and trainer road, my ionic is almost spot on with my tickr HRM.   As soon as I ride outside, accuracy seems to suffer.   

Based on similar observations made while jogging, I'm led to believe there are still some shortcomings with the Fitbit algorithm that processes heart rate while the gyro sensors are active trying to filter out interference from arm movent.   

At the end of the day, I'll always trust my strap that I always wear on a ride anyway.  

---

Yep, 100% on that, I knew this going in, but I was good with the "known issues", from my Charge 2.

My MUCH bigger issue is that if I let the watch auto-start for a ride (or whatever exercise), it's WAY under, on my average HR.

Why would my same ride, same time (obviously the same approximate speeds, give-or-take), be 96BPM average, when the week before on my charge 2, or if I manually start on my Ionic, be the (close to correct) 142BPM average?

There's something really hose with auto-start, and I have multiple logs now that reflect that, it's more than obvious, at least in my simple test set.  I honestly don't know how this made it past test, and out-the-door, coming from a s/w engineering and tech-PM background.  Did they just decide that auto-start wasn't a necessity for RTM?

[bbarrera]

---

@PeteG-1 wrote:
Did they just decide that auto-start wasn't a necessity for RTM?

---

From where I sit, SmartTrack testing was focused on running and walking - Fitbit's target demographic in general and for the Ionic. Putting on my Fitbit PM hat, cycling is 'taken care of' with Strava integration. Since the launch of Surge, cycling has been treated as a check in the box feature and not strongly supported. 

[PeteG-1]

---

@bbarrera wrote:

---

@PeteG-1 wrote:
Did they just decide that auto-start wasn't a necessity for RTM?

---

From where I sit, SmartTrack testing was focused on running and walking - Fitbit's target demographic in general and for the Ionic. Putting on my Fitbit PM hat, cycling is 'taken care of' with Strava integration. Since the launch of Surge, cycling has been treated as a check in the box feature and not strongly supported. 

---

@bbarreraI'm not sure I'm in agreement on that one.  Feature-parity with previous releases, is always a target (with rare exceptions), unless the feature has been deprecated (I still see it, same as for the Charge 2 timeframe).

Yeah, I could see where you might think some/most *real* cycling people would go with Strava (I use my Suunto data for this), but on the other hand, there are TON of "casual cyclists" out there, mixing up their weekly walks, runs, swims, etc.  This would presume that those aren't important, which IMHO is counter to the whole idea of "Fit"bit?  For this crew, the auto-start failing is a big deal, or should be, because it means their rides are going to be grossly under-estimated.

I get that I'm not their target demographic, but if you're going to include an activity still, make it *least* as good as it was (or nuke it, so users can look elsewhere).  I would've pitched a huge fit, as a PM, over this at RTM, partially because I'm a cyclist, but mainly because feature-parity is ingrained in my development experience background (even if it's not part of the Fitbit goals).

[bbarrera]

The date drove the release, so what features are you going to compromise? Keep in mind everything is new (new OS, new HW, must support apps, etc). Heck, from some evidence in a coutlet threads it seems both walking and running use step count to estimate mileage even when using GPS. And you are launching without some key features. 

[PeteG-1]

Yep, I agree, it was definitely a date-driven release, which is almost always a disaster, IME (I've done both, but avoid cycles now that are tied to a certain feature-set AND date-driven).  Vista is a perfect example, that was date-driven (I bailed on Windev for awhile after that one), I won't even go into it...

I just think they should've pulled the music, last-minute, saying they'd push an update maybe in March or something, way better than the current mess, people are going to abandon the music feature, and likely not return.

At least disable the Music UI, worse-case, if it's tricky to pull the modules.

I've not gone back and tried the GPS while walking, although I did notice one "handy" feature.  I go on sort of fast-walks, on my cycling recovery days, it helps with lactic build-up, and before I'd just used my default stride length, with one adjustment, decided it didn't really matter "that much" if I was a few hundred yards off on total distance.

But, the Ionic seems to have pushed the stride-length back to the app, and now even when I leave the GPS off for a walk (so I have multi-day battery, between charges, and I don't need a GPS map of all my walks), it uses my adjusted stride-length, pretty cool feature.

It'd be pretty easy to test though, if you were a runner (can't run anymore, too much SI nerve damage from a bad ski crash as a kid, caught up with me a few years back).  Just set different stride lengths for both, and start a walk-with-GPS, but run instead.  If it's using GPS, it should just "adjust", and record the actual length of the route, vs the "steps length", which would be a lot shorter.

[Camilla_R]

I have also often problems with the heart rate monitoring when cycling outside.

It is around 30-40 bmp too low during intervals and sometimes the exercise stops by it self and is not recorded afterwards. Has happened several times the last weeks. I always start the workout manually.

[Astrid-Marie]

Hey there

I experienced the same thing under different circumstances. When I ride my bike I have no problem hitting a pulse of the 150's and 160's. But when I run (whether or not it's on a treadmill or outdoors), I get a pulse that is way off. Running at 12 km/h i get a pulse of 115-140, but then when I slow down to a walk, after about 40 seconds my pulse will spike to 165 BPM. I know that my pulse is at least 180 BPM and sometimes getting into the 190's when I run at that speed, since I have measured with other methods simultaneously (No, I'm not in the best shape either).

Did you find a fix for this? I am heavily considering returning my FitBit to the store and go back to the Charge series, which I have previously had.

A. 

[PeteG-1]

I "solved it" by switching to Garmin, unfortunately.

It's MUCH closer to my HR (checked against a chest strap), most of the time, although in the winter it still struggles in the cold-cycling (making sure my long-sleeve layer and any jacket or similar is covering the watch helps a lot for the cold). 

I'd say the Garmin is maybe 85-95% accurate, depending on the temperature, whereas my FB (both my Ionic, and to a little lesser degree, my older Charge2) was WAY off, sometimes hardly even registering my ride at all, or only a few parts, as @Astrid-Marie mentioned.  

My FB devices were in the 50-80% accurate range, depending on the cold.

My older FB Charge performed closer to my Garmin, it was a ton more reliable in this regard, than the two newer models.

I wish I could've stayed with FB, I really like their s/w model, and some other features, but the accuracy was sort of making it ridiculous...