Various issues - Software crashing / High Speed Readings / App Usage

Hi all.

Brand new to Zwift. Currently encountering a few issues, which some I gather are ongoing regarding Elite …

(1) Using a brand new Elite Turbo Muin with Misuro B+. I have set it up correctly including the right settings on the Misuro B+. I have also turned the belt tension up many times already based on other suggested fixes. I have correctly allowed Zwift to use the Misuro B+ as a Power source and not speed source, bridged via the Android Zwift Companion App.

Elite have an app called “myETraining” which let’s you setup and pickup the sensors for the first time correctly. It has a “level” training mode where you just pedal away and it shows you a screen of data. This data is pretty much correct (+/- 5%) versus real life in terms of the correlation between watt, RPM and speed (kmh). It also correlates correctly to how I physically feel when riding in real life.

Now comparing the Elite “myETraining” data to Zwift figures. When riding, I look at Zwift and concurrently have the Elite app open in front of me. I am getting around 20% higher Zwift speed readings and 5-10% higher watt readings. The problem is worse when more effort is deployed by me (e.g. standing sprint). For example, the Elite data tells me 110 watts & 70 RPM & 20kmh. Zwift tells me 120 watts & 70 RPM & 30kmh (yes, RPM is pretty much matching between the two).

What makes the Zwift figures more unbelievable is that the bike I have on the Elite Turbo Muin is an old Orbea 8 speed, running compact 50/34 cranks and an 11-28 cassette. The experience I mention above was when I was in the small chainring (34T) and 16T cassette cog. My “real life riding” bike has a mid-compact 52/36 and an 11-28 cassette. I know for a fact that for me to get to 30kmh on the flats, I have to crank much faster than 70 RPM on the 16T cog, and also I am turning over a 36T chainring which has 2 teeth more than the Orbea!

As a result, I am easily overtaking other Zwift riders like I’m high on some special asthma inhalers. Apparently Zwift support have “resolved” this case, but I don’t think it is really resolved…

Any other tips anyone can give me? I think Zwift need to rejig their maths so that some Turbo Trainers data are downsized … I just don’t get how the Elite app data is correct but Zwift is wrong as both software systems are reading from the same source.

(2) I have a Garmin 935 watch. It has an in-built heart rate monitor. It has a feature where I can broadcast my HR using ANT+. The Elite app (that I mentioned above) picks my HR up perfectly fine and is accurate. The Zwift Companion App does not at all. It picks up only the Misuro B+ but nothing else. So, it is clear that my phone picks up the HR data just fine, but Zwift says no.

Has anyone got their smart watch to broadcast the HR successfully to work with Zwift? Any other tips?

(3) Every time I exit a ride, after either selecting “save” or “discard”, the Zwift app crashes and totally exits back to Windows.

Is this normal?

(4) In the Zwift Companion App, is there a way to start your ride, as in actually press the go/ride button? I only see a stop button…

~Thanks everyone. Issue #1 is the biggest problem of course. I just don’t see the point of using Zwift if I am “cheating” all the way.

  1. You can’t make any comparison between the speed your trainer says your doing and Zwift. Speed in Zwift is calculated on power, power to weight vs the gradient and other factors. Speed on a trainer is usually calculated by how fast the flywheel is rotating.

Speed in Zwift is higher than the real world. One reason is no head winds and smooth roads.

  1. Crashing when exiting is not normal. You probably need to do a support ticket.

  2. You start your ride in Zwift as soon as you start pedalling. There is no need for a start button.


Nigel is right.

Results compared between Zwift and a Bike Computer will rarely if ever coincide because watt output is calculated differently in Zwift and takes grade and resistance into account. While you can get remarkably close to real-world statistics and utilize Zwift as a training platform, it’s also a game with a focus on social interaction.

It looks like you’re attempting to connect your Garmin 935 HRM watch to a Windows 8 PC. While we are working on it, currently Zwift doesn’t support Bluetooth connections under Windows. There are two ways to get around that, though.

With Zwift Companion App, you’ll be able to host your Bluetooth connections from your smartphone or tablet over WiFi to your Windows 8 PC. However, even if your Samsung has built-in ANT+ capability, Zwift Companion won’t pair to it properly.

Alternatively, since your Garmin watch can also use ANT+, you could consider an ANT+ USB Dongle for your PC which would allow you to connect directly over ANT+:

When you save and exit Zwift and it disappears, that’s not Zwift crashing. It purposely ends the session to ensure the next session begins from a blank slate. Without exiting Zwift after each session, there’s a chance that persistent session data can cause performance issues, so to get around that, we just insist that Zwift gets a quick reboot between saved rides.

The stop button you noticed in Zwift Companion app doesn’t stop the rider, it stops the event, workout, or freeride you’re participating in and exits entirely. We don’t have a ‘start’ button, as adding one would also require some kind of start prompt from Zwift itself and would require members to dismount their bikes if they don’t run Zwift Companion. Generally, to start riding, all you need to do is begin peddling.

Thx guys for the prompt reply.

Very interesting.

The Elite data also bases speed on Wattage. Which also I mention above, but I would have thought that at least the Watt numbers would be identical when I compare the two when riding live, and not have a 5-10% error all the time. And just to clarify, the data comparison was recorded/compared when riding in Zwift on a flat 0’ gradient.

Is there a way I can give myself a disadvantage in Zwift? I forget if it was during setup, but can I change my rider weight to be heavier so I go slower?

I will look at getting a USB dongle to receive all data (Misuro and Garmin watch)… But will I still be able to use the Zwift Companion App as a “remote control” without having double data being picked up?

Also just on the point of speed, why does Zwift not take into account opposing forces such as air movement and road resistance? Even on a zero wind day and smooth flat road there is still opposing forces. 

Regarding Zwift exiting after every ride, that is really archaic. For a paid subscribed online game, versus today’s online gaming scene, not being able to return to a main menu is like breaking the most basic of requirements. There are more intensely resource hungry online games that have no problems. 

Thx everyone. Apologies for newbie questions.

Something isn’t right if you’re getting different a power reading on 2 or more devices from the same power source. For example I run Trainerroad on my phone and have it controlling my smart trainer. Power numbers come from my power meter. Zwift runs on my PC purely as some some entertainment. Powers numbers are the same on both over time. There’s some slight variation whilst riding due to how the power smoothing is applied and how it’s sampled.

I’ve had best results with Zwift using an ANT+ dongle connecting to the FEC power source rather than the standard power source for my smart trainer. FEC is a superior protocol. (Normally however I use my power meter as the power source but every so often I use the smart trainer’s power just to check its accuracy.)

Elite will be using a different alogrithm to calculate speed than what Zwift uses. 

Yeah, you’ve nailed the problem. I think Zwift is too lenient in it’s formulae. 

My trainer is not a pure smart trainer and has a sensor (Misuro B+) that is not a proper power meter. Therefore Zwift and Elite relies on algorithms to calculate speed… However the wattage should be identical. I will come back with results on using a ANT+ dongle.

BTW the alogrithms Zwift use from my experience aren’t that good. For example I also have my old Cycleops Fluid 2 Trainer. I did a test one day using Zwift’s zpower on my computer and then Trainerroad’s virtual power on my tablet. There was a significant difference between the 2 with Trainerroad being closer to my power meter than Zwift. Zwift was 40 - 50 watts out and Trainerroad around 20 watts out. (Power meter is correct as I’ve since got a set of Assioma dual sided pedals and that agrees with my 4iiii crank based power meter). Theoretically you would expect power to be the same.

This is why many races on Zwift don’t allow zpower users to podium or even exclude them from the results. They can be well off the mark with power readings.

And that is why I am surprised and disappointed.

Zwift has been around for long enough to have a lot of data. Ppl are paying a reasonably high amount of money for subscription. Surely there is enough data to analyse which brands perform how, and enough money to pay someone to make this improvement… and also to upgrade the software so that it doesn’t exit after every ride :neutral_face:

For example, zpower users… Surely they have enough data to then apply a digital handicap if they are consistently overestimating. Elite Muin users, judging by forum dates, have been on Zwift for at least 3 years, and that is a lot of data to analyse.

Problem with zpower is it’s only an approximation. Things like tyre pressure and how much tension you apply make a big difference. It can be both over and under estimating for the same model of trainer.  They do actually restrict some undefined trainers to 400 watts. Also if you put out way too much power you get a warning message.

So long as you’re not racing on Zwift I wouldn’t worry about it. Just enjoy yourself. Even if your power was close to being accurate you would be excluded from many races. If you do want to race then a decent direct drive trainer and or a power meter is best. Wheel on smart trainers can be way off or that’s my experience with a Kickr Snap.

Yeah cool.

My Elite Muin is direct drive, so i was hoping it would be at least within 5% error.

Thx for the chat… Going to test ANT+ dongle…