Spoke sensor for bicycle

On Tuesday, June 14, 2022 at 1:59:54 PM UTC-4, John Doe wrote:
> Ricky <gnuarm.del...@gmail.com> wrote: 
> 
> > John Doe wrote:
> >> Dimiter_Popoff wrote: 
> 
> >>> I started cycling again (used to as a kid...) some 4 years ago. And 
> >>> while I know there are plenty of super cheap "bicycle computers" I 
> >>> have 
> > 
> >>> been thinking of making my own. Thus so far 4 years no speed/mileage 
> >>> indicator for me, obviously. While this may likely remain the case 
> >>> forever I am still thinking about it - not the obvious things like 
> >>> the MCU, display etc., just the rotation sensor. They all use a magnet 
> >>> placed somewhere on one of the wheels and some sensor, I have used a 
> >>> Hall sensor for pretty precise positioning of the rotor of our TLD 
> >>> reader etc., but I only still keep on thinking of making the thing 
> >>> because I imagine it sensing just the spokes, i.e. making some 
> >>> inductive sensor. 
> >>> 
> >>> I know what I will try out etc., I may even get to designing one 
> >>> before the millennium is over but well, these thoughts seem to be on 
> >>> topic and I am sure there are people with more experience making 
> >>> inductive sensors than I have. 
> 
> >> This might be blasphemy, but... 
> >> 
> >> Speed and mileage indication is easily done with a smartphone these 
> >> days, at least here in America. 
> > 
> > People often talk about cell phones as if they had universal coverage. 
> > That is true for 99.9% of where people spend their time. But that 0.1% 
> > of the time, like biking in the woods, you are out of cell phone range. 
> > So any solution needs to not depend on a cell phone connection. If you 
> > can use the GPS in your phone without a cell connection, great! But I 
> > haven't found such a program as yet. Maybe I just haven't dug deep 
> > enough.
> On Android? Can anybody name one such Android app (with a decent rating) 
> that DOESN'T work off-line (at least after you have paid the 1 or $2 for 
> it)? There are plenty of hiking apps, obviously all of them work off-line. 

I can't get Alltrails to do much offline.  It does have a numeric display that shows your lat/long, but that's not much use.  To do anything else it's asking me to log in!  That's hard to do without a connection.  I guess people download stuff in advance of a hike or whatever... or they just use their data connection, when they have one. 

-- 

Rick C.

+ Get 1,000 miles of free Supercharging
+ Tesla referral code - https://ts.la/richard11209

On Tuesday, June 14, 2022 at 8:51:30 AM UTC-7, Dimiter Popoff wrote:
> I started cycling again (used to as a kid...) some 4 years ago. 
> And while I know there are plenty of super cheap "bicycle computers" 
> I have been thinking of making my own. 
> ...use a magnet placed somewhere on one 
> of the wheels and some sensor, ... I 
> imagine it sensing just the spokes, i.e. making some inductive sensor. 

A magnet secured to a spoke, and a reed switch,  makes a durable sensor, but
is somewhat clumsy.   A sensor of the metal spoke is hard, because spokes 
aren't all magnetic, there's stainless steels.   Have you considered a reluctance
sensor on a gear-cluster tooth?   There's some DC insensitivity, unlike with Hall
sensors, but maybe a  low-speed signal dropout isn't going to be a bother.

A reluctance sensor is a coil of wire between a long weak-ish magnet (you don't want to
attract metal bits) and a tooth's circular trajectory.   Into a high impedance, it generates a 
peak/valley two-lobed pulse at a tooth passage, where the volt-seconds of the peak 
and the volt-seconds of the valley are constants (so voltage goes low when speed is slow).
A set/reset flipflop makes a good debouncer for that.

You might consider putting a magnet not on a spoke, but inside
the tire; a button-size rare earth magnet won't show, but a Hall sensor of the non-hysteresis
type should pick it  up easily.   I'd put the magnet center-of-tread, and affix the sensor near the bottom bracket.

On 6/14/22 8:51 AM, Dimiter_Popoff wrote:
> I started cycling again (used to as a kid...) some 4 years ago.
> And while I know there are plenty of super cheap "bicycle computers"
> I have been thinking of making my own.
> Thus so far 4 years no speed/mileage indicator for me, obviously.
> While this may likely remain the case forever I am still thinking
> about it - not the obvious things like the MCU, display etc., just
> the rotation sensor. They all use a magnet placed somewhere on one
> of the wheels and some sensor, I have used a Hall sensor for
> pretty precise positioning of the rotor of our TLD reader etc.,
> but I only still keep on thinking of making the thing because I
> imagine it sensing just the spokes, i.e. making some inductive sensor.
> 

There are at least two reasons for the magnet. One is that it triggers a 
reed contact on the other side. This allows the whole speedometer to not 
have to be turned off by hand for battery conservation. It simply does 
that on its own and when the reed switch changes state again it turns 
itself back on. So you can just park your bike and walk away. Well, 
maybe not in a large city because then it gets stolen.

The second reason is that this system works down to very low speeds. 
There isn't actually a lower limit and it will still clock the miles 
correctly.


> I know what I will try out etc., I may even get to designing one before
> the millennium is over but well, these thoughts seem to be on topic and
> I am sure there are people with more experience making inductive
> sensors than I have.
> 

I can already smell an engineer's solution coming up. Huge sensor, 
multiple bus cables, a laptop mounted to the handlebar ... :-)

-- 
Regards, Joerg

http://www.analogconsultants.com/

On 6/14/22 10:16 AM, John Doe wrote:
> Dimiter_Popoff <dp@tgi-sci.com> wrote:
> 
>> I started cycling again (used to as a kid...) some 4 years ago. And
>> while I know there are plenty of super cheap "bicycle computers" I have
>> been thinking of making my own. Thus so far 4 years no speed/mileage
>> indicator for me, obviously. While this may likely remain the case
>> forever I am still thinking about it - not the obvious things like the
>> MCU, display etc., just the rotation sensor. They all use a magnet
>> placed somewhere on one of the wheels and some sensor, I have used a
>> Hall sensor for pretty precise positioning of the rotor of our TLD
>> reader etc., but I only still keep on thinking of making the thing
>> because I imagine it sensing just the spokes, i.e. making some inductive
>> sensor.
>>
>> I know what I will try out etc., I may even get to designing one before
>> the millennium is over but well, these thoughts seem to be on topic and
>> I am sure there are people with more experience making inductive sensors
>> than I have.
> 
> This might be blasphemy, but...
> 
> Speed and mileage indication is easily done with a smartphone these days,
> at least here in America.
> 

Then, later in the afternoon during an all-day ride, there will be a low 
battery warning, followed by a dark screen.

Why do we always have to make things so complicated? When I grew up 
speedometers didn't even need any electrical power. They just worked.

-- 
Regards, Joerg

http://www.analogconsultants.com/

On Tuesday, June 14, 2022 at 2:00:52 PM UTC-4, Dimiter Popoff wrote:
> On 6/14/2022 20:43, Jan Panteltje wrote: 
> > On a sunny day (Tue, 14 Jun 2022 18:51:21 +0300) it happened Dimiter_Popoff 
> > <d...@tgi-sci.com> wrote in <t8aapr$ab6$1...@dont-email.me>: 
> > 
> >> I started cycling again (used to as a kid...) some 4 years ago. 
> >> And while I know there are plenty of super cheap "bicycle computers" 
> >> I have been thinking of making my own. 
> >> Thus so far 4 years no speed/mileage indicator for me, obviously. 
> >> While this may likely remain the case forever I am still thinking 
> >> about it - not the obvious things like the MCU, display etc., just 
> >> the rotation sensor. They all use a magnet placed somewhere on one 
> >> of the wheels and some sensor, I have used a Hall sensor for 
> >> pretty precise positioning of the rotor of our TLD reader etc., 
> >> but I only still keep on thinking of making the thing because I 
> >> imagine it sensing just the spokes, i.e. making some inductive sensor. 
> >> 
> >> I know what I will try out etc., I may even get to designing one before 
> >> the millennium is over but well, these thoughts seem to be on topic and 
> >> I am sure there are people with more experience making inductive 
> >> sensors than I have. 
> > 
> > Or you could make something like my gm_pic2: 
> > http://panteltje.com/panteltje/pic/gm_pic2/ 
> > 
> > It uses GPS and writes to SDcard or EEPROM where you were at any time, logs radiation there too 
> > There is software to replay the trip on google maps. 
> > After all the GPS talk here .. 
> > Simple maaz using the space and time between data points will allow you to find speed and distances. 
> > A GPS module and a Microchip 18F14k22 + EEPROM is all you need. 
> > Oh and a battery... 
> > Now been working 24/7 for 8 years... As clock mostly. 
> > No tinkering with your bike needed. 
> > Also logs trips per bus or train. 
> > 
> >
> Is the GPS accuracy/latency of the speed measurement close to that of 
> a car's speed indicator?

"Close" is probably the best you can claim.  Each 1 second update is a second old by the time you get it from the GPS.  The accuracy can be there, but it depends on the constellation.  With some xx feet of accuracy, each calculation can have errors that are significant if you are not moving fast.  30 mph is 44 fps, so 14 foot accuracy (what I'm seeing now and a typical value) is significant.  GPS measurements are typically filtered.  That's one reason why your car navigation can prompt you rather late sometimes (or early).  It is hard for a GPS to know what lane you are in, but can tell if you are on the parallel access road and not the main highway.  

-- 

Rick C.

-- Get 1,000 miles of free Supercharging
-- Tesla referral code - https://ts.la/richard11209

On Tuesday, June 14, 2022 at 2:21:58 PM UTC-4, whit3rd wrote:
> On Tuesday, June 14, 2022 at 8:51:30 AM UTC-7, Dimiter Popoff wrote: 
> > I started cycling again (used to as a kid...) some 4 years ago. 
> > And while I know there are plenty of super cheap "bicycle computers" 
> > I have been thinking of making my own.
> > ...use a magnet placed somewhere on one 
> > of the wheels and some sensor, ... I
> > imagine it sensing just the spokes, i.e. making some inductive sensor.
> A magnet secured to a spoke, and a reed switch, makes a durable sensor, but 
> is somewhat clumsy. A sensor of the metal spoke is hard, because spokes 
> aren't all magnetic, there's stainless steels. Have you considered a reluctance 
> sensor on a gear-cluster tooth? There's some DC insensitivity, unlike with Hall 
> sensors, but maybe a low-speed signal dropout isn't going to be a bother. 

Does it have to be magnetic?  A conductor moving in a magnetic field creates a current which also creates a magnetic field which can be detected, no?  

-- 

Rick C.

-+ Get 1,000 miles of free Supercharging
-+ Tesla referral code - https://ts.la/richard11209 

tirsdag den 14. juni 2022 kl. 20.28.02 UTC+2 skrev Ricky:
> On Tuesday, June 14, 2022 at 2:00:52 PM UTC-4, Dimiter Popoff wrote: 
> > On 6/14/2022 20:43, Jan Panteltje wrote: 
> > > On a sunny day (Tue, 14 Jun 2022 18:51:21 +0300) it happened Dimiter_Popoff 
> > > <d...@tgi-sci.com> wrote in <t8aapr$ab6$1...@dont-email.me>: 
> > > 
> > >> I started cycling again (used to as a kid...) some 4 years ago. 
> > >> And while I know there are plenty of super cheap "bicycle computers" 
> > >> I have been thinking of making my own. 
> > >> Thus so far 4 years no speed/mileage indicator for me, obviously. 
> > >> While this may likely remain the case forever I am still thinking 
> > >> about it - not the obvious things like the MCU, display etc., just 
> > >> the rotation sensor. They all use a magnet placed somewhere on one 
> > >> of the wheels and some sensor, I have used a Hall sensor for 
> > >> pretty precise positioning of the rotor of our TLD reader etc., 
> > >> but I only still keep on thinking of making the thing because I 
> > >> imagine it sensing just the spokes, i.e. making some inductive sensor. 
> > >> 
> > >> I know what I will try out etc., I may even get to designing one before 
> > >> the millennium is over but well, these thoughts seem to be on topic and 
> > >> I am sure there are people with more experience making inductive 
> > >> sensors than I have. 
> > > 
> > > Or you could make something like my gm_pic2: 
> > > http://panteltje.com/panteltje/pic/gm_pic2/ 
> > > 
> > > It uses GPS and writes to SDcard or EEPROM where you were at any time, logs radiation there too 
> > > There is software to replay the trip on google maps. 
> > > After all the GPS talk here .. 
> > > Simple maaz using the space and time between data points will allow you to find speed and distances. 
> > > A GPS module and a Microchip 18F14k22 + EEPROM is all you need. 
> > > Oh and a battery... 
> > > Now been working 24/7 for 8 years... As clock mostly. 
> > > No tinkering with your bike needed. 
> > > Also logs trips per bus or train. 
> > > 
> > > 
> > Is the GPS accuracy/latency of the speed measurement close to that of 
> > a car's speed indicator?
> "Close" is probably the best you can claim. Each 1 second update is a second old by the time you get it from the GPS. The accuracy can be there, but it depends on the constellation. With some xx feet of accuracy, each calculation can have errors that are significant if you are not moving fast. 30 mph is 44 fps, so 14 foot accuracy (what I'm seeing now and a typical value) is significant. GPS measurements are typically filtered. That's one reason why your car navigation can prompt you rather late sometimes (or early). It is hard for a GPS to know what lane you are in, but can tell if you are on the parallel access road and not the main highway. 
> 

https://www.gps.gov/systems/gps/performance/accuracy/

tirsdag den 14. juni 2022 kl. 20.00.52 UTC+2 skrev Dimiter Popoff:
> On 6/14/2022 20:43, Jan Panteltje wrote: 
> > On a sunny day (Tue, 14 Jun 2022 18:51:21 +0300) it happened Dimiter_Popoff 
> > <d...@tgi-sci.com> wrote in <t8aapr$ab6$1...@dont-email.me>: 
> > 
> >> I started cycling again (used to as a kid...) some 4 years ago. 
> >> And while I know there are plenty of super cheap "bicycle computers" 
> >> I have been thinking of making my own. 
> >> Thus so far 4 years no speed/mileage indicator for me, obviously. 
> >> While this may likely remain the case forever I am still thinking 
> >> about it - not the obvious things like the MCU, display etc., just 
> >> the rotation sensor. They all use a magnet placed somewhere on one 
> >> of the wheels and some sensor, I have used a Hall sensor for 
> >> pretty precise positioning of the rotor of our TLD reader etc., 
> >> but I only still keep on thinking of making the thing because I 
> >> imagine it sensing just the spokes, i.e. making some inductive sensor. 
> >> 
> >> I know what I will try out etc., I may even get to designing one before 
> >> the millennium is over but well, these thoughts seem to be on topic and 
> >> I am sure there are people with more experience making inductive 
> >> sensors than I have. 
> > 
> > Or you could make something like my gm_pic2: 
> > http://panteltje.com/panteltje/pic/gm_pic2/ 
> > 
> > It uses GPS and writes to SDcard or EEPROM where you were at any time, logs radiation there too 
> > There is software to replay the trip on google maps. 
> > After all the GPS talk here .. 
> > Simple maaz using the space and time between data points will allow you to find speed and distances. 
> > A GPS module and a Microchip 18F14k22 + EEPROM is all you need. 
> > Oh and a battery... 
> > Now been working 24/7 for 8 years... As clock mostly. 
> > No tinkering with your bike needed. 
> > Also logs trips per bus or train. 
> > 
> >
> Is the GPS accuracy/latency of the speed measurement close to that of 
> a car's speed indicator?

a car speedometer is only required to be -0 to +10% 

On 6/14/2022 21:21, whit3rd wrote:
> On Tuesday, June 14, 2022 at 8:51:30 AM UTC-7, Dimiter Popoff wrote:
>> I started cycling again (used to as a kid...) some 4 years ago.
>> And while I know there are plenty of super cheap "bicycle computers"
>> I have been thinking of making my own.
>> ...use a magnet placed somewhere on one
>> of the wheels and some sensor, ... I
>> imagine it sensing just the spokes, i.e. making some inductive sensor.
> 
> A magnet secured to a spoke, and a reed switch,  makes a durable sensor, but
> is somewhat clumsy.   A sensor of the metal spoke is hard, because spokes
> aren't all magnetic, there's stainless steels. 

Aaaah, the stainless steel is a game changer of course. Had not thought
of that at all - though I know the spokes are not rusty (some at the end
where they are tightened to the wheel but not all and not much). And it
is not like they have not seen water, 1-2 months ago getting back
after waiting for a torrential rain to subside I was nearly in submarine
mode ...

>  Have you considered a reluctance
> sensor on a gear-cluster tooth?   There's some DC insensitivity, unlike with Hall
> sensors, but maybe a  low-speed signal dropout isn't going to be a bother. >
> A reluctance sensor is a coil of wire between a long weak-ish magnet (you don't want to
> attract metal bits) and a tooth's circular trajectory.   Into a high impedance, it generates a
> peak/valley two-lobed pulse at a tooth passage, where the volt-seconds of the peak
> and the volt-seconds of the valley are constants (so voltage goes low when speed is slow).
> A set/reset flipflop makes a good debouncer for that.
> 
> You might consider putting a magnet not on a spoke, but inside
> the tire; a button-size rare earth magnet won't show, but a Hall sensor of the non-hysteresis
> type should pick it  up easily.   I'd put the magnet center-of-tread, and affix the sensor near the bottom bracket.

Putting the magnet underneath the tire is an idea. Of course I might
lose it if I have a flat and repair it on the road (you won't believe
how often I was getting a flat until I got my latest tires, they claim
to have some Kevlar and I had just one flat in a year or two (can't
remember if I got them last year or the year before, almost replacement
due now). But then I have a decent supply of coin magnets, too (say 1mm
thick/10mm diameter)...

The Hall sensors I have "in the drawer" are some with hysteresis though,
they need such a magnet really close (within 1mm if not less) to work,
would need some other kind but that should be the easiest part.
But well, it looks I am not going to start the project any time soon.
And I did download some biking app to the phone, this makes the
project's prospects even more bleak...

On Tuesday, June 14, 2022 at 11:32:03 AM UTC-7, Ricky wrote:
> On Tuesday, June 14, 2022 at 2:21:58 PM UTC-4, whit3rd wrote: 

> >  ... a reluctance 
> > sensor on a gear-cluster tooth? There's some DC insensitivity, unlike with Hall 
> > sensors, but maybe a low-speed signal dropout isn't going to be a bother.

> Does it have to be magnetic? A conductor moving in a magnetic field creates a current which also creates a magnetic field which can be detected, no? 

A nonmagnetic gear tooth will (in motion) disturb a magnetic field, and make a pulse.
The eddy current response, though, is less than a ferromagnetic material's, and has a time decay
of its own,  added to the coil's DC insensitivity.

An IR (modulated?  polarized?) source, aimed at a retroreflector on a spoke, has good range and sensitivity, at a cost
of some power usage.  Rejecting ambient light and dirt are other issues there.