Suunto Smart Heart Rate Belt 2

Mff73

@atoponce
For testing, we were advised to wear it, not too tight. Maybe try to loose a bit the belt.

atoponce

@Mff73 I wonder if having oily skin creates a film under the electrodes preventing a strong conductive connection until sweat breaks through. Perhaps an alcohol wipe to degrease the skin and electrode or ultrasound gel would improve things.

martintrail

@atoponce Well, it shouldn’t be rocket science. Just put it on and go. It seems more like a FW issue to me. But we’ll see after a few more runs. I’ve never had any problems with my old Suunto HRM or my Garmin HRM.

Inge Nallsson

@martintrail said:

It seems more like a FW issue to me

I sure hope it is! My replacement unit arrives today because you guys had a good streak with other units… And talking about streaks, the last bad one you had with strong drops in the HR (and ZS pruning) visavis a perfect R-R plot, did you upload the correct run for analysis? The timeline from the SA graph is ca 38 minutes while the R-R data has a timeline of ca 27 minutes.

martintrail

@Inge-Nallsson What I know, is the R-R chart is built from the beats the strap actually sent, so when the strap loses contact, those minutes simply disappear instead of showing as a gap. About 10 minutes are missing, all in the first half of the run.

That matches the HR graph in the Suunto app: during the first 20 minutes there are flat lines and impossible drops (down to 88 bpm while climbing), which look like the watch filling in gaps. After that the HR looks natural, and ZoneSense started showing data around minute 28, roughly 10 minutes after the signal became stable.

The R-R data that did get recorded is very clean, so it’s not noise, it’s dropouts.

Inge Nallsson

@martintrail Thanks for the clarification. I’m not that at home with the workings of R-R, obviously.

atoponce

Today’s run. I first wiped my chest clean with an alcohol wipe, then applied aloe vera gel to the electrodes. Aloe vera should be a good enough water-based stand-in until my electrode-specific gel is delivered.

The first ~8 minutes are noisy in the R-R intervals plot, then it settles. Took a bathroom stop at ~18 minutes, thus the jump there. This seems to be an improvement over the previous plots, but I don’t know if that scatter in the first ~8-10 minutes is normal. Also, I’m tapering for a Saturday marathon, thus the shorter run duration. Comparing apples-to-apples in duration is probably preferred.

Regardless, the average line tells me this is largely physiology. R-R is settling while HR climbs as I warm-up. But the scatter in those first minutes tell me perhaps the contact between the strap and chest isn’t uniform, thus the artifacts. I was kind of hoping the SHRB2 artifact detection in the module would remove this, but maybe that’s looking for something different?

Inge Nallsson

Inge-Nallsson said:

I will definitely initiate a return of the SHRM2. The faith in this unit is lost and questions about the overall hardware/software interactions are raised in my mind.

Alright, a new unit arrived today and I groaned when seeing that the serial number began as “SN: 2625”, which means manufactured in week 25 of this year (middle of june) - exactly the same as the previous bad one.

I proceeded with the pairing in Suunto App (no new firmware), but this time I left everything as default. That is, I did not turn on running dynamics. My reasoning being that I didn’t want anything to possibly interfere with my main interest; good HR data.

After pairing with the Race S I turned off the optical HR option in the Running configurations, as well as the switch for running dynamics registration from the belt, minimising error paths as much as possible.

Turning the watch off and on again (yeah, computer basics) I went out and did the exact same running experiment as before: 1 km on flat asphalt (ca 5 metres Asc/Desc) in about 7 minutes. I do it as 500 metres out and then back again.

Live HR looked just like it should with a slowly rising bpm number from a normal 70-ish standing base. Saving the run, the HR graph on watch looked All OK. The Suunto App had a beautiful curve with a slight dip where I turned the run around; just as expected. I was however slightly worried about the flat section of 124 bpm neighboring the 3 minutes mark seen in the graph below:

So I went to the Runalyze site to scrutinize the R-R data. And it turned out that the 124 bpm section indeed was a period with a stable heart rate, not a loss of signal or some other anomaly. Look at these nice graphs:

Tomorrow I will do a Run/Walk session (still injured with cracked rib and ankle problems), doing 10 * 200 metres running/800 metres walking at a skiing path. A good source of Asc/Desc and HR variations over ca 1 hour 33 minutes.

Here’s hoping this SHRM2 unit will stay sound…

atoponce

Eduardo Mateos reviews the SHRM2 and sees similar spikes and dropouts early in the run, including stuck data paired with different watches, including the Run 2.

Data analysis starts at 9:28: https://www.youtube.com/watch?v=qchbJWGsn08

martintrail

My HRM data wasnt fine today. Again. Spikes.

snow*

@martintrail It happened to me systematically as well. I contacted support, and they advised me to perform a factory reset before sending it in for service. no more issues since then

martintrail

@snow Okay, I’ll give it a try. How do I do it?

snow*

@martintrail Press the button on the transmitter for 20+ seconds until led starts blinking Blue Red Green or so

martintrail

@snow Thanks, all done, so we’ll see how it goes next time.

2b2bff

@snow said:

they advised me to perform a factory reset before sending it in for service. no more issues since then

As often as I see in here that a (factory) reset helps with all sorts of issues it increases my worries about software quality over the whole Suunto ecosystem… Very sad this…

Inge Nallsson

@atoponce said:

Eduardo Mateos reviews the SHRM2 and sees similar spikes and dropouts early in the run, including stuck data paired with different watches, including the Run 2.

From what I could see, he never had the SHRM2 paired with the Run 2, only with devices from competitors. And one of the stuck data segments he specifically pointed out was from the Run 2 using internal optical HR capture. He called it cadence lock, but I’m sceptical since the HR was down at 130 - 140 bpm while he was doing a 5 km run with a fast climbing HR on the other devices (sincerely hope his cadence isn’t that low, in order to salvage the knees).

He blamed the bad data from the SHRM2 on “software”, but considering that all his exercises were done on the 3, 4, 5 and 6 September and my own first (bad) unit was sent the 11:th, while this my second (and seemingly good) unit got dispatched the 29:th, they all most likely have the same firmware version. I know that my two units were manufactured the same week (25) and had the same firmware upon receipt. And no new version is publicly available.

In my mind this raises the question of how fragile the electronic components in the pod are. Could there be Good, Average and Bad units in circulation from the manufacturing process? And can units decline by themselves? As early adopters we do take some risks…

I’m touching wood here while declaring a second exercise as a success with the new unit. There was one minor hiccup in the first 1.5 minutes but the rest seems spot on. I used the Garmin Forerunner 955 as a control (the OHR there usually stays +1-2 bpm higher than HR belts) and Interval alarm (you can program Intervals directly on the watch based on time or distance). Doing 10 * 200 metres running / 800 metres walking with the main part on a skiing path, here’s the first two intervals zoomed in. Running downhill the first 120 metres the, wet and firmly fastened, SHRM2 (connected to the Race S) has the general idea, but not until the run portion is over, after 200m /ca 1:20, does it suddenly jump up to a correct HR. The jump is from 92 to 110 bpm. Suunto:

Garmin has the correct data from start. Note that what seems like a sudden climb in HR reflects the terrain; going downhill and then slightly up:

As I said previously, the rest of the exercise was captured exemplary by the SHRM2. During the whole 1 hour and 30 minutes I frequently, almost manically, glanced at the Race S, counted off two seconds (to give the OHR a chance to catch up) and then looking at the FR955 it would have the same (or being 1 bpm off) Heart Rate value. The HR was, and felt, correct for both terrain and paces. Suunto:

Garmin:

To conclude this report, here is the R-R data from the Runalyze site. Total Anomalies were only 0.4% with the main scatter occurring during the first 1.5 minutes - the bad HR episode:

If things deteriorate with this SHRM2 unit I will report it here, naturally.

Edit: For the whole session, the FR955 had Avg HR 101 bpm and Max 132. The SHRM2 Avg HR was 100 and Max 132.

martintrail

I can confirm that HRM factory reset, removing it from the Suunto app, watch and the bluetooth list helped. Pair again and the data looks fine today.

Inge Nallsson

@martintrail Well, I won’t do a factory reset (yet) on this unit, which I think is fundamentally sound hardware wise. But the algorithm does have problems, so I’m hoping for a firmware update - eventually…

Continued with my Run/Walk sessions yesterday, slightly increasing the run portion; 10 * 300 metres running / 700 metres walking. Garmin FR955 (with own OHR) still as control and Interval alarm. Preparation for the exercise was to spray both chest and strap with a high saline solution (1/2 a teaspoon table salt in 250 millilitre of tap water) and doing a minute of slow running as warmup. Therefore the HR begins at 97 bpm.

Didn’t see it during the run, but the second interval fell victim to some algorithm shenanigans. Zooming in here to show the faulty interval registration between two good ones. The rest of the exercise was OK, even giving good data during and after a fall at the fifth interval at 34 minutes in (bloody knees and palms) showing up only as a few scatter readings in the R-R stream. I was up and running again within 20 seconds…

Algo fault in Suunto:

Reality in Garmin:

Whole exercise, Suunto:

And Garmin:

R-R data Runalyze (total of 0.5% Anomalies):

Whole session: FR955 (OHR) Avg HR 107 bpm, Max 139. SHRM2 Avg HR 107 bpm, Max 139.

snow*

@Inge-Nallsson said:

Well, I won’t do a factory reset (yet) on this unit, which I think is fundamentally sound hardware wise. But the algorithm does have problems, so I’m hoping for a firmware update - eventually

I was having those issues before I performed a factory reset (btw, I have the same sn 2625) and it hasn’t happened since then so I’m not sure if it’s an algorithm-related issue. Let’s see if @martintrail has anything to say about it.

martintrail

@snow My last run after the factory reset looked good. 80-minute was clean: HR follows the terrain naturally, no flat sections, no dropouts. If I’m seeing this right

Since so many people seem to have the same issue, it looks more like a pairing or firmware problem than faulty hardware. Hopefully Suunto can look into it, and in the meantime this might save someone a return. I’ll report back if the dropouts come back.

HR, Pace, Elevation