Welcome Guest,Register Now
Log In

ANT Forum

Welcome guest, please Login or Register

   

Commercially available sensors with low rpm sensitivity

Rank

Total Posts: 3

Joined 2020-06-11

PM

Hi,

I am using ANT+ sensors linked with Unity3D for a rehabilitation project and have tried several commercially available sensors who all seem to share the same fundamental problem (in the context of the project) i.e. that anything below 25 rpm will not activate the sensor. Between 25rpm and 50rpm there is significant data loss, which is evident from the frequent null-readings even though connection is established. I have tried both cadence and speed sensors: Wahoo Speed, Wahoo Cadence, Coospo cadence/speed.

Most producers have indicated an upper measurement extremum (about 300rpm and 120km/h), but not really a lower cutoff.

Does anyone know if this is hardware standards, or are aware of some sensors with high sensitivity to slower changes in velocity or revolutions? users will frequently have empaired motor skills, so detecting small changes is paramount

any help will be greatly appreciated! smile      
RankRankRankRank

Total Posts: 175

Joined 2012-06-27

PM

You are referring specifically to ANT+ bike speed, cadence, and speed & cadence sensors?

What do you mean by "frequent null-readings"?

Some sensors implement detection of "motion" that says you have stopped peddling/rolling if the time between rotations drops below a threshold. You may observe this in page 5 for some speed sensors or cadence sensors. This feature is not available for speed and cadence sensors.

A limitation for all 3 sensor types is that at best they can report only the duration of a whole rotation. If you are interested in observing changes in rotational velocity within a single rotation, then ANT+ sensors are not suitable unless you attach them to a rotation that is geared from the true rotation.

     

Signature

Ian Haigh

Rank

Total Posts: 3

Joined 2020-06-11

PM

Hi Ian

Thanks for your reply (and sorry for the delay on my end)

- I am referring to both ANT+ speed sensors and ANT+ cadence sensors. I have tried both with similar result

- "Frequent null readings". Connection is established, and "connected" but the signal returns "0" instead of anything above "0".

"Some sensors implement detection of "motion[..]".
I am interested in knowing what that "threshold might be, as I assume this may be the issue. Are you familiar with sensors that DOES NOT implement this detection?

what do you mean by "in page 5"?

Thanks again!

/Emil

     
RankRankRankRank

Total Posts: 175

Joined 2012-06-27

PM

When you say the signal is 0, the signal is RPM? How are you decoding the broadcast from the sensor? The ANT+ profiles for bike speed and cadence do not actually send RPM, so if you are getting 0 rpm that may be from your decoders handling of sensor information instead of the sensor itself. You may want to investigate what the sensor is actually transmitting.

Page 5 is one of the pages of data that the sensor can send to the display. See section 5.3.6 and 6.3.6 in the profile document. Not all sensors support it. When page 5 is not supported the displays usually detect 0 speed and cadence based on a proprietary algorithm.

Hopefully the above is the issue, and you don't need a different sensor. Unfortunately this behavior is not something we track for certified sensors. Your best bet is to look for a sensor with an on / off switch or similar. One of the reasons to identify a lack of speed or cadence is to put the device into a lower power mode. Simple sensors with an on / off switch won't need to do this. If the sensor doesn't detect low power state, then it should be the simple revolution counter you are looking for.      

Signature

Ian Haigh

Rank

Total Posts: 3

Joined 2020-06-11

PM

For communication we use a developed SDK that we do not manage ourselves (https://assetstore.unity.com/packages/tools/network/advanced-ant-71980)
I'll reach out to the developers, to see if there's anything on their end.

Furthermore, I've observed significant data loss, with the same sensors but in different locations. One of the bikes have a power adapter attached to the front, and suspect that it may be 2nd and 3rd harmonics from the AC circuit that interferes with the signal. I will test with a spectrum analyser next week.

Thank you so much for the insights. Also, I've also observed that some speed sensors measures more than one degree of freedom rotation (usually the cheaper ones) which permits attachments to the pedal arm rather than the wheel arm. By point is, it would maybe be possible to get someone to develop a custom ANT+ sensor that are optimally suited for this special case of use. I have not been able to locate sensors with an on/off switch. Alternatively, I'll probably buy and test a lot of different devices incl. susceptibility to external noise. We have a decent electronics lab in the university. I'll be happy to share the results with the community.