Before I get to the main topic.
Topeak, famous for saddle bags and portable pumps, once had a brand called "Panobike" that offered
cycle computers and
an app service for managing their logs.
The image above is
a speed/cadence sensor from that Panobike.
What's shown on top is an Elite roller trainer mat and
a belt for a three-roller set.
↑This kind of thing
The C on the lower left is the magnet-sensing part for cadence
The part marked S on the flipped side is
the magnet-sensing part for speed.
Panobike
adopted a sales policy of selling
the Pano cycle computer and
speed/cadence sensor separately,
but since this sensor was cheap enough to sell for around 2,000 yen (at the time),
those unscrupulous folks who caught on to this
resorted to the shady practice of buying only the sensor,
which resulted in
cycle computers alone sitting unsold for a long time at the distributors.
If you search for "panobike," you'll find
images of just this sensor and at best the heart rate sensor,
but almost no images of the cycle computer itself.
What, me? I only bought the sensor.
So I bought an iGP SPORT (hereafter IGP, too much hassle typing the lowercase letter up front)
accelerometer-type cadence sensor, but
it won't pick up maximum cadence (around 200 rpm) when I put out a momentary high cadence on the three-roller set.
To investigate whether this is a general tendency of accelerometer-type sensors
or just an IGP issue,
I also bought a Cateye accelerometer-type
cadence sensor,
and paired each with different cycle computers.
Though they're both Polar wrist-mounted heart rate monitors.
For what it's worth,
what's laid underneath is a place mat that I use as a dust cover for
a laptop for the roller trainer,
and what looks like a wheel is
a picture of a cross-section of a citrus fruit.
I've paired the Vantage V2 on the left with Cateye,
and the Pacer on the right with IGP.
These can display 1 to 4 items on one screen,
but I've set it to display only maximum cadence.
↑This is cadence I produced by hand-cranking the crank during pairing confirmation.
Thinking about it afterwards,
I probably should have woken them up at a slightly lower value
to see the behavior better.
For just a moment, I did some pedaling that would have put out
at least 170 rpm by any conservative estimate,
but IGP only detected the cadence that dropped from there,
and Cateye showed no change at all.
I did the same thing two more times.
The first time both showed no change,
but on the second attempt, IGP updated to 142-156
from the estimated 170 cadence drop.
Cateye ignored all three efforts.
I unpaired the Cateye from the Vantage V2 on the left
and re-paired it with the Panobike sensor.
The right continues with IGP.
I didn't expect Cateye to drop out like that.
I hand-cranked the crank at a lower cadence than before.
From there, I stabilized my pedaling at just over 130 rpm.
Both showed roughly similar values,
but while Panobike displayed 100→115→128→134 as it changed,
IGP jumped straight to 100→133.
IGP's slower update frequency is not related to
the difference in grade of the Polar heart rate monitor (the left is higher-end).
I pedaled at a momentary 170+ rpm.
Again, IGP didn't detect it.
While updating the 173 that appeared on the Panobike side
and doing a good bit of sustained hard effort
(getting pretty tired by now)
IGP also picked it up and showed an even higher value.
I think this is just because the update frequency timing
happened to coincide
with the peak cadence.
This particular test is a pretty unusual use case, so
I don't want to take the stance that accelerometer-type
cadence sensors are no good based on this.
Even in my case, if I mounted it on a road bike and
rode it on roads,
the update frequency and accuracy would be no problem.
Also, I've only experimented with two, IGP and Cateye, so
I can't conclude that
"accelerometer-type sensors can't pick up momentary high cadence,"
but I'd say it's fair to say
"there's a tendency toward that."
In my view, the accelerometer-type
cadence sensors I'm talking about here
include the cadence measurement function
contained in power meters of the type that are
stuck to the inside of the left crank arm.
Topeak, famous for saddle bags and portable pumps, once had a brand called "Panobike" that offered
cycle computers and
an app service for managing their logs.
The image above is
a speed/cadence sensor from that Panobike.
What's shown on top is an Elite roller trainer mat and
a belt for a three-roller set.
↑This kind of thing
The C on the lower left is the magnet-sensing part for cadence
The part marked S on the flipped side is
the magnet-sensing part for speed.
Panobike
adopted a sales policy of selling
the Pano cycle computer and
speed/cadence sensor separately,
but since this sensor was cheap enough to sell for around 2,000 yen (at the time),
those unscrupulous folks who caught on to this
resorted to the shady practice of buying only the sensor,
which resulted in
cycle computers alone sitting unsold for a long time at the distributors.
If you search for "panobike," you'll find
images of just this sensor and at best the heart rate sensor,
but almost no images of the cycle computer itself.
What, me? I only bought the sensor.
So I bought an iGP SPORT (hereafter IGP, too much hassle typing the lowercase letter up front)
accelerometer-type cadence sensor, but
it won't pick up maximum cadence (around 200 rpm) when I put out a momentary high cadence on the three-roller set.
To investigate whether this is a general tendency of accelerometer-type sensors
or just an IGP issue,
I also bought a Cateye accelerometer-type
cadence sensor,
and paired each with different cycle computers.
Though they're both Polar wrist-mounted heart rate monitors.
For what it's worth,
what's laid underneath is a place mat that I use as a dust cover for
a laptop for the roller trainer,
and what looks like a wheel is
a picture of a cross-section of a citrus fruit.
I've paired the Vantage V2 on the left with Cateye,
and the Pacer on the right with IGP.
These can display 1 to 4 items on one screen,
but I've set it to display only maximum cadence.
↑This is cadence I produced by hand-cranking the crank during pairing confirmation.
Thinking about it afterwards,
I probably should have woken them up at a slightly lower value
to see the behavior better.
For just a moment, I did some pedaling that would have put out
at least 170 rpm by any conservative estimate,
but IGP only detected the cadence that dropped from there,
and Cateye showed no change at all.
I did the same thing two more times.
The first time both showed no change,
but on the second attempt, IGP updated to 142-156
from the estimated 170 cadence drop.
Cateye ignored all three efforts.
I unpaired the Cateye from the Vantage V2 on the left
and re-paired it with the Panobike sensor.
The right continues with IGP.
I hand-cranked the crank at a lower cadence than before.
From there, I stabilized my pedaling at just over 130 rpm.
Both showed roughly similar values,
but while Panobike displayed 100→115→128→134 as it changed,
IGP jumped straight to 100→133.
IGP's slower update frequency is not related to
the difference in grade of the Polar heart rate monitor (the left is higher-end).
I pedaled at a momentary 170+ rpm.
Again, IGP didn't detect it.
While updating the 173 that appeared on the Panobike side
and doing a good bit of sustained hard effort
(getting pretty tired by now)
IGP also picked it up and showed an even higher value.
I think this is just because the update frequency timing
happened to coincide
with the peak cadence.
This particular test is a pretty unusual use case, so
I don't want to take the stance that accelerometer-type
cadence sensors are no good based on this.
Even in my case, if I mounted it on a road bike and
rode it on roads,
the update frequency and accuracy would be no problem.
Also, I've only experimented with two, IGP and Cateye, so
I can't conclude that
"accelerometer-type sensors can't pick up momentary high cadence,"
but I'd say it's fair to say
"there's a tendency toward that."
In my view, the accelerometer-type
cadence sensors I'm talking about here
include the cadence measurement function
contained in power meters of the type that are
stuck to the inside of the left crank arm.