bizarre behavior of cycling computer



On Tue, 12 Dec 2006 17:55:02 -0500, "Roger Zoul"
<[email protected]> may have said:

>"Jim Higson" <[email protected]> wrote in message
>news:[email protected]
>:: FWIW, My Specialized wireless cycle computer always
>:: shows crazy speeds when I take it on electric trains,
>:: even when I'm a few cars away from the bike.
>
>Electric trains - are those the ones with the huge electric currents running
>underneath? :)


Actually, the currents aren't as large as you might think, but the RFI
noise from the contact shoes can be strong enough to show up as
crackle and hiss on a hearing aid. When a radio signal is at a field
strength level that allows it to be picked up by the circuit paths
between components, the results can get interesting. That's how you
end up with CB radio chatter on the intercom system of a building near
a freeway even though the intercom has no radio receiver included.

--
My email address is antispammed; pull WEEDS if replying via e-mail.
Typoes are not a bug, they're a feature.
Words processed in a facility that contains nuts.
 
On Tue, 12 Dec 2006 16:57:58 -0600, richard wrote:

> Some LED lights will do this.
>
> [email protected] wrote:
>> I have a cheap cycling computer I got at Performance ("dB 4LW
>> Wireless"). Sometimes, when the bike is not moving, I see the computer
>> showing speeds over 70 mph. This does not happen when I am riding.
>> Obviosuly, this behavior causes over-estimation of my travel distance,
>> and it also screws up my max speed reading (I've never gone above ~55
>> mph, sorry). I've seen this behavior with 3 computers of this
>> particular type. Does anyone know why this happens?
>>


Mine have been affected by power lines. Of course, you can also get high
speeds if the computer picks up your wheel sensor and someone else's.


--

David L. Johnson

__o | If all economists were laid end to end, they would not reach a
_`\(,_ | conclusion. -- George Bernard Shaw
(_)/ (_) |
 
I could not care less about the max speed being messed up. But it also
messes up the distances travelled is left run alone for several
minutes.

So, OK, if I see this happenning I will try to change my x,y in space
until it stops. Or I will turn off the computer. Wait... mine does NOT
have an off button, it turns off when there is no movement... so when
it wrongly thinks there is movement, it will never turn off. #$%#$
great.
 
Just a thought on how the manufacturers can fix this.

The bizarre reading happens when I am not moving. So, why not
programming the computer in such a way that when it "sees" an
"acceleration" from 0 to 75 mph in less than 1 second (what is it?
cosmic speed?) to just ignore that reading and consider it zero? Once
the real movement begins (the magnet passes by the sensor) the noise
does not matter anymore, b/c the magnet signal dominates. THis can alos
be implemented in HRMs
 
[email protected] wrote:

> Just a thought on how the manufacturers can fix this.
>
> The bizarre reading happens when I am not moving. So, why not
> programming the computer in such a way that when it "sees" an
> "acceleration" from 0 to 75 mph in less than 1 second (what is it?
> cosmic speed?) to just ignore that reading and consider it zero? Once
> the real movement begins (the magnet passes by the sensor) the noise
> does not matter anymore, b/c the magnet signal dominates. THis can alos
> be implemented in HRMs


They could also use a more sophisticated protocol. When on the train, my
wifi (provided by the train company) is pretty fast. If the field is clean
enough to get the internet at high speed, I should be able to get the
*tiny* bandwidth from a cycle computer OK. Probably it is a trade off with
them not wanting to make the transmitter/receiver use up more energy on
more robust encoding.

The interference on trains could the cycle computer because I'm sat close to
several people, all of who have mobile phones.
 
Jim Higson <[email protected]> wrote in
news:[email protected]:

> Probably it is a trade off with them not wanting to make the
> transmitter/receiver use up more energy on
> more robust encoding.
>


That's part of it. Also, your Wi-Fi uses a "closed loop" protocol so any
bad packets can be rejected and the sender required to retransmit them.

In the case of a wireless bike computer, it's open loop: since there's no
way for the head unit to inform the sensor unit that it didn't
understand, it simply takes anything that it thinks is a signal and
counts it. That's because if it used a complex signal to weed out noise,
it would also drop a lot of the "real" counts due to the same RFI causing
the signal to look bad.

Then, as you indicated, they need to be as low power as possible. This
means that they most likely want to turn off the transmitter between
pulses. With the pulses coming in at about 10Hz or so, that doesn't leave
a lot of time to transmit a complex bit pattern and still "rest" between
pulses. If you compress the data into a smaller timeframe you need more
transmitter power.

Open loop is always a challenge to make reliable.

I will stick to wired computers. They are a pain to install - especially
if you want the installation to look clean, but they are FAR more
reliable and if the installation is done right it can be much less
intrusive looking than a big fat wireless pickup unit strapped to the
fork. Plus, once the installation is done, you don't need to do it again.
All the wired computers I've ever seen (including my ew ErgoBrain) use
simple reed switches as sensors. On my old bike, I changed the head unit
three times (each time to a different model) and was always able to
simply connect the old cadence and speed sensor wires to the new head
unit bracket.

Cheers,
David
 
"Pete Biggs" <[email protected]> wrote
in news:[email protected]:

> Wired computers giving double-speed readings is not uncommon when the
> sensor is too near the magnet and switches twice per wheel revolution.
> If it's just on brink of being OK, it may be fine for most
> revolutions but not for some others. I've experienced this several
> times with Cateye Mity 3s.
>
> I guess the same can happen with wireless as well (in addition to all
> the other problems).
>
>


Try rotating the magnet 90 degrees, or getting a different magnet. The
issue of double speed readings is that the reed switch is seeing both the
north and south poles of the magnet and opening between them thus
counting twice each time the magnet passes. (It should only be an issue
when riding slowly.) By rotating the magnet 90 degrees you can orient it
such that only one of the poles faces the sensor and the problem should
go away. The idea of a different magnet is to get one that has the poles
oriented perpendicular to the plane of the wheel.

Cheers,
David
 
[email protected] wrote in news:1165983960.292420.196720
@l12g2000cwl.googlegroups.com:

> I could not care less about the max speed being messed up. But it also
> messes up the distances travelled is left run alone for several
> minutes.
>
> So, OK, if I see this happenning I will try to change my x,y in space
> until it stops. Or I will turn off the computer. Wait... mine does NOT
> have an off button, it turns off when there is no movement... so when
> it wrongly thinks there is movement, it will never turn off. #$%#$
> great.
>
>


Do yourself a favor and replace the wireless computer with a wired one.
Yes, the installation is a much bigger pain, but once it's done, it will
be far more reliable and can even look MUCH better than a bulky wireless
unit. (Plus no sensor batteries to replace.)

A friend and I installed a computer on my new bike. It was a time
consuming task, but I'm extremely happy with the results. I uploaded some
photos to Flickr to illustrate the installation:

The computer wires are routed along with the front brake and front
derailleur cable housings using woven cable sheath (catalog photo):
http://www.flickr.com/photos/68241631@N00/321927262/ so I don't have an
unsightly wire wrapped around the brake cable as most installations use.

You can see this braded cable cover on the two cables in this photo:
http://www.flickr.com/photos/68241631@N00/321889543/) A trick here to
keep the ends looking clean is to fold the cut strands into the weave so
they don't fray. Note that you can also see the speed sensor and its wire
on the inside of the fork, but I had to take the photo at just the right
angle or you wouldn't know it's there.

The cadence sensor wire comes out where the derailleur cable housing
stops at the downtube. This is on a Klein, so the cables route internal
to the frame. The brighter red "dot" you see where the wire comes out is
just the camera flash reflecting off the end of the cablestop on the
downtube. (Sorry for the poor picture quality, but I just snapped them to
show what's going on and didn't set up proper lighting.)

The wire is routed to the bottom of the downtube where we used silicone
to "glue" it to the bottom of the tube. You can see both "fat" cables
compared to the rear brake and derailleur cables which don't have the
woven covering.

This photo: http://www.flickr.com/photos/68241631@N00/321889547/ shows
where the cadence wire routes on the bottom of the downtube and around
the derailleur cable guide. We mounted the cadence sensor on the bottom
of the left chainstay and the magnet was glued (more silicone) to the end
of the left pedal axle using a few small washers as spacers to make it
stick out from the hole in the crank arm. When we first installed it I
had put my old Speedplay X2 (steel axle) pedals on the bike and
everything worked fine. I replaced the pedals with X1 Speedplays and the
cadence sensor became intermittent. I added a couple of extra washers to
push the magnet a bit closer to the sensor which it seemed to make up for
the lack of a ferrous axle and it's worked reliably ever since.

You can see the head unit in this photo:
http://www.flickr.com/photos/68241631@N00/321899650/ Note that any extra
wire is tucked under the handlebar tape.

Finally, here's a photo of the whole bike:
http://www.flickr.com/photos/68241631@N00/316751635/ Although it's not
the best resolution, I think it still illustrates that the installation
is very clean.

I understand that this is not a job that everyone will want to tackle. I
didn't time it, but I think it took the two of us at least 5 hours.
However, to me, it was well worth the time (although I don't know if my
friend will agree since he also did a similar job on his girlfriend's
bike a few months before this one.)

Aside from just the esthetics, the lack of nylon cable ties means that
it's easier to clean the bike without getting caught on the ties. I also
think that the silicone keeps the wires from lifting away from the frame
and potentially getting snagged.

And while 10 man-hours might seem a bit "excessive", it's the last time I
will need to deal with the wiring on this bike. We did a similar
installation on my old bike and I replaced the computer a few times over
the course of more than 20 years - each time with a different brand or
model. But I was always able to use the existing harness and sensors.
(All wired computers that I know of use simple reed switches for the
sensors.)

I know this was a long winded explanation of this wiring job. But I just
had to post it here. After all, I'm quite proud of it and none of the
Rivendell fans at the local latte house would appreciate the elegance.
<grin>

Cheers,
David
 
Solvang Cyclist wrote:
> "Pete Biggs" <[email protected]> wrote
> in news:[email protected]:
>
> > Wired computers giving double-speed readings is not uncommon when the
> > sensor is too near the magnet and switches twice per wheel revolution.
> > If it's just on brink of being OK, it may be fine for most
> > revolutions but not for some others. I've experienced this several
> > times with Cateye Mity 3s.
> >
> > I guess the same can happen with wireless as well (in addition to all
> > the other problems).
> >
> >

>
> Try rotating the magnet 90 degrees, or getting a different magnet. The
> issue of double speed readings is that the reed switch is seeing both the
> north and south poles of the magnet and opening between them thus
> counting twice each time the magnet passes. (It should only be an issue
> when riding slowly.) By rotating the magnet 90 degrees you can orient it
> such that only one of the poles faces the sensor and the problem should
> go away. The idea of a different magnet is to get one that has the poles
> oriented perpendicular to the plane of the wheel.
>
> Cheers,
> David


Dear David,

Hmmm . . .

Threads like the one below suggest a fairly low practical limit to reed
switch activation times:

http://groups.google.com/group/rec....=gst&q="reed+switch"&rnum=14#22d22ef67f57e2ba

At 1 mph, a 2124 mm 700c tire is spinning at 12.63 rpm.

Can anyone re-produce the wildly high readings by just wiggling a
magnet past a pickup by hand?

I think that reed switches in cyclocomputers will suffer from the
equivalent of valve float before the leading and trailing edge of a
single tiny magnet can trigger them twice.

I hung my bike from a garage ceiling hook, added a second magnet, and
had some fun spinning the front wheel.

For a 36-spoke wheel with two magnets between 20 and 100 degrees apart,
the normal 15~16 mph speed never exceeded 27 mph as the wheel spun
down.

With the two magnets as close as possible, 20 degrees apart, the speed
would finally rise abruptly to around 7 mph from around 3 mph as the
wheel slowed down.

The magnets were out about 10 & 3/4 inches from the axle, while the
tire was about 14 inches.

The cyclocomputer is a $10 WalMart Schwinn that reads speeds in 0.1 mph
increments and distance in 0.001 mile increments (not that it's reallly
quite that accurate, a single wheelspin being just over a thousandth of
a mile).

Like every other cyclocomputer that I've ever had, it's wired and
ridiculously reliable. (Until they fail dramatically--possibly you're
not supposed to fill them with water or bang them on the ground.)

There are probably lots of odd details to consider, such as the effect
of asymmetrically placed magnets, how far out the magnets are placed,
various brands of cyclocomputer pickups, magnet strength, and so forth.

But if a lead-trailing magnet edge is somehow responsible on the
wireless cyclocomputers in question, then just wiggling the wheel with
the magnet lined up at the sensor should reproduce the problem.

Cheers,

Carl Fogel
 
[email protected] wrote in
news:[email protected]:

>
> Solvang Cyclist wrote:
>> "Pete Biggs" <[email protected]>
>> wrote in news:[email protected]:
>>
>> > Wired computers giving double-speed readings is not uncommon when
>> > the sensor is too near the magnet and switches twice per wheel
>> > revolution.
>> > If it's just on brink of being OK, it may be fine for most
>> > revolutions but not for some others. I've experienced this several
>> > times with Cateye Mity 3s.
>> >
>> > I guess the same can happen with wireless as well (in addition to
>> > all the other problems).
>> >
>> >

>>
>> Try rotating the magnet 90 degrees, or getting a different magnet.
>> The issue of double speed readings is that the reed switch is seeing
>> both the north and south poles of the magnet and opening between them
>> thus counting twice each time the magnet passes. (It should only be
>> an issue when riding slowly.) By rotating the magnet 90 degrees you
>> can orient it such that only one of the poles faces the sensor and
>> the problem should go away. The idea of a different magnet is to get
>> one that has the poles oriented perpendicular to the plane of the
>> wheel.
>>
>> Cheers,
>> David

>
> Dear David,
>
> Hmmm . . .
>
> Threads like the one below suggest a fairly low practical limit to
> reed switch activation times:
>
> http://groups.google.com/group/rec.bicycles.tech/browse_frm/thread/28c1
> 05c6b2356521/22d22ef67f57e2ba?lnk=gst&q=%22reed+switch%22&rnum=14#22d22
> ef67f57e2ba
>
> At 1 mph, a 2124 mm 700c tire is spinning at 12.63 rpm.
>
> Can anyone re-produce the wildly high readings by just wiggling a
> magnet past a pickup by hand?
>
> I think that reed switches in cyclocomputers will suffer from the
> equivalent of valve float before the leading and trailing edge of a
> single tiny magnet can trigger them twice.
>
> I hung my bike from a garage ceiling hook, added a second magnet, and
> had some fun spinning the front wheel.
>
> For a 36-spoke wheel with two magnets between 20 and 100 degrees
> apart, the normal 15~16 mph speed never exceeded 27 mph as the wheel
> spun down.
>
> With the two magnets as close as possible, 20 degrees apart, the speed
> would finally rise abruptly to around 7 mph from around 3 mph as the
> wheel slowed down.
>
> The magnets were out about 10 & 3/4 inches from the axle, while the
> tire was about 14 inches.
>
> The cyclocomputer is a $10 WalMart Schwinn that reads speeds in 0.1
> mph increments and distance in 0.001 mile increments (not that it's
> reallly quite that accurate, a single wheelspin being just over a
> thousandth of a mile).
>
> Like every other cyclocomputer that I've ever had, it's wired and
> ridiculously reliable. (Until they fail dramatically--possibly you're
> not supposed to fill them with water or bang them on the ground.)
>
> There are probably lots of odd details to consider, such as the effect
> of asymmetrically placed magnets, how far out the magnets are placed,
> various brands of cyclocomputer pickups, magnet strength, and so
> forth.
>
> But if a lead-trailing magnet edge is somehow responsible on the
> wireless cyclocomputers in question, then just wiggling the wheel with
> the magnet lined up at the sensor should reproduce the problem.
>
> Cheers,
>
> Carl Fogel
>
>


Carl,

I'm ashamed to admit that despite having worked with relay matrix test
equipment that used reed relays back in the 1970's (among other types), I
can't remember the specs. My recollection is that we ran the relays at
about 1200 "tests per minute". That cycle time had to account for ramp up
to 1000 volts and then back down at the end of each test. But of course,
that's only 20 cycles a second, so it's not very fast.

I agree that this is something that should be able to be tested. The only
issue is that it would need to be tested on a bike that is exhibiting the
problem or the test setup would require some experimentation to get the
distance between the magnet and the reed just right. If the sensor is in
line and close enough to the magnet, then it should stay closed between
the poles. But if it's far enough away to allow the reed to "relax" to
the point that the circuit opens and then closes again when the second
pole appears it would certainly get two "clicks". The question would then
be if the computer has a "debounce" feature to ignore these quick pulses.

If I have some time this weekend, I might try a few experiments.

Cheers,
David
 
In article <[email protected]>,
"Roger Zoul" <[email protected]> wrote:

> "Crescentius Vespasianus" <[email protected]> wrote in
> message
> news:[email protected]
> :: [email protected] wrote:
> ::: I have a cheap cycling computer I got at Performance
> ::: ("dB 4LW Wireless"). Sometimes, when the bike is not
> ::: moving, I see the computer showing speeds over 70 mph.
> ::: This does not happen when I am riding. Obviosuly, this
> ::: behavior causes over-estimation of my travel distance,
> ::: and it also screws up my max speed reading (I've never
> ::: gone above ~55 mph, sorry). I've seen this behavior
> ::: with 3 computers of this particular type. Does anyone
> ::: know why this happens?
> :: ---------------
> :: I have a new cateye micro wireless that gets it's max
> :: speed wacked almost always, when I'm riding in town.
> :: All the rest of the readings are ok, just the max speed
> :: shows some ridiculous high speed. Now when I get out of
> :: town and ride in the boondocks it never happens.
>
> Have you tried relocating the computer/transmitter?
>
> ::
> :: On my other bike I have an old specialized wireless, and
> :: it's high speed never gets wacked, ridden on the same
> :: routes, that wack the other one.
> ::
> :: I don't think there is a way to fix it, but the high
> :: speed reading isn't that important, it's just irritating
> :: that it isn't correct.
>
> It might be correct even if it seems believable! :)


Roger, you keep trying, but that dog won't hunt.
Wireless bicycle receivers will eat anything. I am
telling you, others are telling you. But do not believe
me. Buy yourself a wireless system and install it.
Better yet, send me your address and I will send you
one free, no shipping charge, original documentation
included.

--
Michael Press
 
[email protected] wrote:

> Can anyone re-produce the wildly high readings by just wiggling a
> magnet past a pickup by hand?


No, but I suspect I'm not wiggling well enough. The speed might have to be
within a certain range, and distance to the magnet would need to be
consistent. All this is difficult to manage by hand.

> I think that reed switches in cyclocomputers will suffer from the
> equivalent of valve float before the leading and trailing edge of a
> single tiny magnet can trigger them twice.


I've experienced double-speed readings several times on different bikes with
Cateye Mity 3 wired computers. Solved each time simply by moving the sensor
further out. (And solved for good by glueing magnet to spokes and using
soft rubber tape under the sensor to ensure no movement).

Perhaps Cateye's reed switches work faster than other brands? They make a
louder noise, if that has anything to do with anything?

~PB
 
On Tue, 12 Dec 2006 19:09:42 -0600, Werehatrack
<[email protected]> wrote:

>>Electric trains - are those the ones with the huge electric currents running
>>underneath? :)

>
>Actually, the currents aren't as large as you might think, but the RFI
>noise from the contact shoes can be strong enough to show up as
>crackle and hiss on a hearing aid.


On DC electrified railways, the currents are also strong enough to
distort the image and colour on CRT monitors used for platform
passenger information display systems.

The only way around that is to shield the monitors - and that's
expensive.
 
Andrew Price <[email protected]> wrote in
news:[email protected]:

> On Tue, 12 Dec 2006 19:09:42 -0600, Werehatrack
> <[email protected]> wrote:
>
>>>Electric trains - are those the ones with the huge electric currents
>>>running underneath? :)

>>
>>Actually, the currents aren't as large as you might think, but the RFI
>>noise from the contact shoes can be strong enough to show up as
>>crackle and hiss on a hearing aid.

>
> On DC electrified railways, the currents are also strong enough to
> distort the image and colour on CRT monitors used for platform
> passenger information display systems.
>
> The only way around that is to shield the monitors - and that's
> expensive.
>


I would imagine that the distortion would have more to do with the
magnetic fields than with the current. In fact, a steady DC current
shouldn't distort a CRT at all. In any case, perhaps the cheapest fix for
that problem would be to replace the CRT monitors with LCD panels. No
electron beam - no distortion from magnetic fields.

Cheers,
David
 
"Solvang Cyclist" <[email protected]> wrote in message
news:[email protected]
:: Andrew Price <[email protected]> wrote in
:: news:[email protected]:
::
::: On Tue, 12 Dec 2006 19:09:42 -0600, Werehatrack
::: <[email protected]> wrote:
:::
::::: Electric trains - are those the ones with the huge
::::: electric currents running underneath? :)
::::
:::: Actually, the currents aren't as large as you might
:::: think, but the RFI noise from the contact shoes can be
:::: strong enough to show up as crackle and hiss on a
:::: hearing aid.
:::
::: On DC electrified railways, the currents are also
::: strong enough to distort the image and colour on CRT
::: monitors used for platform passenger information
::: display systems.
:::
::: The only way around that is to shield the monitors -
::: and that's expensive.
:::
::
:: I would imagine that the distortion would have more to
:: do with the magnetic fields than with the current. In
:: fact, a steady DC current shouldn't distort a CRT at
:: all.

Why not? If the CRT is moving, then it won't matter if the current is DC or
not.

In any case, perhaps the cheapest fix for that
:: problem would be to replace the CRT monitors with LCD
:: panels. No electron beam - no distortion from magnetic
:: fields.
::
:: Cheers,
:: David
 
"Michael Press" <[email protected]> wrote in message
news:[email protected]
:: In article <[email protected]>,
:: "Roger Zoul" <[email protected]> wrote:
::
::: "Crescentius Vespasianus" <[email protected]> wrote
::: in message
::: news:[email protected]
::::: [email protected] wrote:
:::::: I have a cheap cycling computer I got at Performance
:::::: ("dB 4LW Wireless"). Sometimes, when the bike is not
:::::: moving, I see the computer showing speeds over 70
:::::: mph. This does not happen when I am riding.
:::::: Obviosuly, this behavior causes over-estimation of
:::::: my travel distance, and it also screws up my max
:::::: speed reading (I've never gone above ~55 mph,
:::::: sorry). I've seen this behavior with 3 computers of
:::::: this particular type. Does anyone know why this
:::::: happens?
::::: ---------------
::::: I have a new cateye micro wireless that gets it's max
::::: speed wacked almost always, when I'm riding in town.
::::: All the rest of the readings are ok, just the max
::::: speed shows some ridiculous high speed. Now when I
::::: get out of town and ride in the boondocks it never
::::: happens.
:::
::: Have you tried relocating the computer/transmitter?
:::
:::::
::::: On my other bike I have an old specialized wireless,
::::: and it's high speed never gets wacked, ridden on the
::::: same routes, that wack the other one.
:::::
::::: I don't think there is a way to fix it, but the high
::::: speed reading isn't that important, it's just
::::: irritating that it isn't correct.
:::
::: It might be correct even if it seems believable! :)
::
:: Roger, you keep trying, but that dog won't hunt.
:: Wireless bicycle receivers will eat anything. I am
:: telling you, others are telling you. But do not believe
:: me. Buy yourself a wireless system and install it.
:: Better yet, send me your address and I will send you
:: one free, no shipping charge, original documentation
:: included.
::

:)

That's okay. I have one...


:: --
:: Michael Press
 
"Roger Zoul" <[email protected]> wrote in
news:[email protected]:

> Why not? If the CRT is moving, then it won't matter if the current is
> DC or not.
>


Ah, yes, ypu are right. While in this case, the CRTs are not moving (on the
platform), the trains would be, so the net effect would be te same.

Still, my suggestion of using LCDs as a cheap fix should work.

Cheers,
David
 

Similar threads