Cyclical heart rate inaccuracy

Yet more problems.

Heart rate suddenly dropped from around 170 (around threshold) to around 140 (steady run) at 15 km despite no dramatic change in running pace during a half-marathon.

Any sensible suggestions as to why this might be?

Does the firmware dial down the led brightness after a period of time to extend battery life and if so can this be disabled?

I tried to upload an image of the heart rate but this seems to have failed with both png and jpg formats, even when reduced to 320x240 (The forum website seems rather clunky to me)

Dreuw said:

Any sensible suggestions as to why this might be?

Sounds like cadence lock

My cadence would be around 180 so I think that very unlikely.

Do you know how to upload an image, (size, format) as for some reason I can't get it to work?

I don't really get these algorithms. After a period of learning it should know what is a physiologically plausible heart rate for a given running power for a user and should be able to discard a signal (presented along with another) that is outside a chosen p value even if it is of dominant amplitude. If that is too hard to process procedurally could it not be programmed into the PLC of the sensor?

Optical HR seems to have come of age so the performance of the Instinct 2 X is a disappointment. Although bulky it is comparatively light so this surprises me somewhat.

Suddenly losing heart rate during a competition is a big fail.

P.S. btw I now suspect the cyclical heart rate variation occurs when doing a structured workout and navigating. I wonder if the processor is too slow? (chosen for battery life?) to cope. It has quite a small memory too, perhaps for the same reason.

Dreuw said:

My cadence would be around 180 so I think that very unlikely.

I'd tell the opposite. If you tell the watch was showing 170 bpm, then I'd tell it is rather clear it was cadence-locked, and once you warmed up a bit, and the veins started to pump the blood stronger through the veins, it dropped to the real HR.

BTW, the cadence lock does not necessarily mean that the HR will be exactly equal to the cadence. The sensor analyses the optical signal which is the result of an interference of lot of waves - besides the HR, it is the signal noise of the steps, but also diverse parasitic signals due to the moving of the arm, shaking of the watch on the wrist, and pushing of the watch against the veins.

Dreuw said:

Do you know how to upload an image, (size, format) as for some reason I can't get it to work?

Copy and paste works for me.

Otherwise, personally I do not experience any cyclical HR problems with my watch, and without being able to compare your HR graph from the watch with an alternative data source, such as a HRM strap, or a lab ECG, it is hard to make any comments, since we do not know what your true HR was. I recommend doing a comparative test in a lab, or at least recording the HR with a HRM chest strap simultaneously, and comparing the results then. You can get one for a few bucks in every sport store. Get a one with BT connection, so that you can connect it directly to the phone, and log the data for example with the Strava app.

Dreuw said:

My forerunner 255 does not behave this way

Wear both watches on your next run and post the graphs please. 
FWIW I get these dips at irregular intervals with strap and watch and have had them for many years. Conclusion is that it’s physiological in my case. 

Today I can upload the graphs as I think you got the wrong end of the stick. There is a sudden drop out of HR at around 15k during a half-marathon.  The heart rate before this is exactly as expected for me while running at 1/2M pace, around 170, my cadence slightly over 180, (so disambiguated OK).

In any event, I think you missed one of my points. The HR after the drop-out is not consistent with that expected for this workload (flat pace or power or gradient adjusted pace - whatever you like) so the watch algorithm should, after getting to know me a bit, dismiss this signal component and consider only the others.

It is difficult for me to envisage a damped dynamic system where there is a substantial component of resonance at 140 per minute which is greater than the driving frequency of just over 180 but my physics is fading and I would need to dig out a book.

I am running a marathon on 24 Nov and will revert to using my FR255 which seems more reliable (though the battery life is a little limited). I may have a go with both watches, one linked to a chest strap, but only after this event.

P.S. hope you don't mind a small correction: an English person would say "I'd say" rather than "i'd tell". I am no linguist but I think tell is an imperative form, an instruction, for example "tell me your name". I am trying to learn French but finding it hard!

I will have a go after my marathon on 24 Nov. 

That is a curious thing. Have you worn an ECG to document them? Is it an arrhythmia like AF, sick-sinus or a heart block?

It is a shame these devices don't measure within activity heart rate variability. That might be really interesting!

Dreuw said:

I may have a go with both watches, one linked to a chest strap, but only after this event.

If you have a chest strap, then why not using it? If you can show a comparative graph of the same activity recorded simultaneously by your Instinct 2, and by another device connected to the HRM strap, it avoids all speculations about what the true HR really was. And if you have such data, you can send it to the Support directly - it may help them finding and fixing the problem, if there really is one.

Dreuw said:

P.S. hope you don't mind a small correction:

No, not at all. Many thanks for the remark! I'll try to remember it.

Dreuw said:

I am trying to learn French but finding it hard!

Me too. Worse than English.

agreed!

I've worn ECG during exercise while undertaking my research and have no problems. It just 'is' for me apparently.