Wireless on recumbents (RANS Stratus)

[Mark Leuck]

"Cletus D. Lee" <[email protected]> wrote in message
news:[email protected]...
> In article <[email protected]>,
> [email protected] says...
> >
> > "Cletus Lee" <[email protected]> wrote in message
> > >> Another choice, a Garmin Geko. uses GPS for speed and distance. No
wheel
> > magnet. It does under report speed by about 1.5% over a well calibrated wired computer.
> >
> > My Garmin Geko doesn't under report speed--it's right on.
>
> I am sorry, You are correct about the speed. Mine is accurate to the .10 mph. Except for the time
> it registered 888 mph in downtown Houston while I was stopped at a traffic light. It is distance
> that is under reported by ~1.5%. The reason the distance is off is because the Geko measures
> distance from point to point. A route with a lot of turns will be in error. Also I have found that
> false signals are causing route track errors in city canyons.
>
> --
> Cletus D. Lee Bacchetta Giro Lightning Voyager http://www.clee.org
> - Bellaire, TX USA -

Some of that tracking error may be avoidable, one problem is they are always off by a few feet
anyway and if I select power saver mode on my emap it checks position every 2 or 3 seconds instead
of 1, small difference but still a difference. Even then both gps and computer are pretty spot on
in the end.

Odd thing is mine always shows me at least 100 feet off in my home town, anywhere else is fine

 

[Mark Leuck]

"brian hughes" <[email protected]> wrote in message
news:[email protected]...
>
>
> Well, I'm not so sure about the trip distance being under-reported. When
I
> first bought it, just for an experimentation I brought it along and
watched
> it update while my wife was driving down the freeway. Assuming the survey mile markers were
> correct, I didn't see any noticeable error at all. I
also
> used it when I rode a century (actually a 108 mile ride), every rest stop was exact on my Garmin,
> as was the final distance when compared to the cheat-sheet. It was not 1.5 miles off, it wasn't
> even .15 miles off. But heck, I guess you could also argue that they probably used some kind of
GPS
> to calculate and make the cheat sheet to start with.
>
> Maybe you know for sure, but I theorize the Garmin does not simply triangulate between points
> (thus cutting corners as you indicate). I believe the Garmin somehow integrates velocity, and
> thus doesn't
necessarily
> cut off corners. The reason I believe this is I've seen the Garmin
continue
> to update trip distance for a several seconds (at least) after the GPS signal is lost.
>
> Brian

My only gripe about Garmin is they are wayyyyy too slow updating maps, I often ride on roads I know
have been around for at least 10 years that still don't show up on the gps.

 

[Mark Leuck]

"brian hughes" <[email protected]> wrote in message
news:[email protected]...
>
> "bentcruiser" <[email protected]> wrote in message
> news:[email protected]...
> > Rvc wrote:
> > > I've just purchased a Stratus and wonder if any improvements have
been
> > > made in the wireless computer world ?
> >
> >
> >
> > If you are talking distance trouble, try putting the computer on the derailleur tube.
>
> That's what I did on my RANS Tailwind, it works just fine on the derailler tube. Except last
> Tuesday when the weather was quite cold for these parts (about 15 deg F) on my morning commute. My
> wireless only reported about
> 1/2 the distance I really traveled--kept jumping between my actual speed
and
> zero. The rest of the week the temp was in the mid 20's and it worked fine--I guess I could try a
> new battery but this one isn't that old.
Seems
> like under 20 deg F it doesn't work dependably.
>
> I don't think poor cold weather performance is a wireless thing. On my V-Rex and my DFs I have
> wired computers and they don't work well either
when
> it gets well below freezing. Anyone know of a computer that works well in the bitter cold?

Some of that may be atmospheric conditions instead of the temperature

 

[Cletus D . Lee]

In article <[email protected]>,
[email protected] says...
>
> "Cletus D. Lee" <[email protected]> wrote in message
> news:[email protected]...
> > In article <[email protected]>,
> > [email protected] says...
> > >
> > > "Cletus Lee" <[email protected]> wrote in message
> > > >> Another choice, a Garmin Geko. uses GPS for speed and distance. No
> wheel
> > > magnet. It does under report speed by about 1.5% over a well calibrated wired computer.
> > >
> > > My Garmin Geko doesn't under report speed--it's right on.
> >
> > I am sorry, You are correct about the speed. Mine is accurate to the .10 mph. Except for the
> > time it registered 888 mph in downtown Houston while I was stopped at a traffic light. It is
> > distance that is under reported by ~1.5%. The reason the distance is off is because the Geko
> > measures distance from point to point. A route with a lot of turns will be in error. Also I have
> > found that false signals are causing route track errors in city canyons.
> >
> > --
> > Cletus D. Lee Bacchetta Giro Lightning Voyager http://www.clee.org
> > - Bellaire, TX USA -
>
> Well, I'm not so sure about the trip distance being under-reported. When I first bought it, just
> for an experimentation I brought it along and watched it update while my wife was driving down the
> freeway. Assuming the survey mile markers were correct, I didn't see any noticeable error at all.
> I also used it when I rode a century (actually a 108 mile ride), every rest stop was exact on my
> Garmin, as was the final distance when compared to the cheat-sheet. It was not 1.5 miles off, it
> wasn't even .15 miles off. But heck, I guess you could also argue that they probably used some
> kind of GPS to calculate and make the cheat sheet to start with.
>
> Maybe you know for sure, but I theorize the Garmin does not simply triangulate between points
> (thus cutting corners as you indicate). I believe the Garmin somehow integrates velocity, and thus
> doesn't necessarily cut off corners. The reason I believe this is I've seen the Garmin continue to
> update trip distance for a several seconds (at least) after the GPS signal is lost.

What I know is from observation. Take a look at the track produced by your Garmin. Compare on a map
against the roads actually taken. I think you will find the route is a series of chords. The
distance along those chords is the distance measured and reported by the GPS. The Chord is always
(by definition) shorted than the curve that it spans. This is the error that I refer to. It is not
much but it accumulates. More accurate on straight roads, less so on routes with lots of turns.

As for your 108 mile Century, if the sponsors did anything other than actually chain off the actual
distance, then it is an approximation.

If you want to really see how accurate your Garmin really is at distance, take it to a High School
track. At cycle speeds, do several miles of a known distance (caution the track may be metric). See
how that compares to the known distance.

I did this with my Magellan once and was lucky to get three points per lap. I would have sampled
more often if I had been walking.
>

--
Cletus D. Lee Bacchetta Giro Lightning Voyager http://www.clee.org
- Bellaire, TX USA -

 

[William Higley]

I agree with Cletus about chords being shorter than arcs. I would try his method of testing the
distance of a GPS unit at a High School track.

My Garmin GPS III+ takes readings when I change speed or direction. The following excerpt from
one of my journeys (this was in an uphill portion of the route) shows how frequently it will mark
a point on a track. When I change direction or speed it stores a "Trackpoint", provides the
latitude and longitude of the point, shows the time it created the point, then shows the
distance, elapsed time, speed, and heading to the next "Trackpoint" Trackpoint N47 48 39.4 W122
00 44.8 3/15/03 9:10:33 AM 36 ft 00:00:02 12.2 mph 278° true Trackpoint N47 48 39.4 W122 00 45.1
3/15/03 9:10:34 AM 17 ft 00:00:01 11.7 mph 270° true Trackpoint N47 48 39.0 W122 00 46.2 3/15/03
9:10:39 AM 88 ft 00:00:05 12.0 mph 242° true Trackpoint N47 48 38.8 W122 00 46.6 3/15/03 9:10:41
AM 39 ft 00:00:02 13.2 mph 227° true Trackpoint N47 48 38.6 W122 00 46.8 3/15/03 9:10:42 AM 20 ft
00:00:01 14.0 mph 220° true

The following excerpt shows the frequency when I am in flat and straight sections of a ride
Trackpoint N47 50 18.4 W122 04 25.5 3/15/03 9:24:04 AM 570 ft 00:00:18
21.6 mph 326° true Trackpoint N47 50 21.2 W122 04 30.4 3/15/03 9:24:17 AM 438 ft 00:00:13
22. mph 310° true Trackpoint N47 50 23.9 W122 04 36.1 3/15/03 9:24:32 AM 476 ft 00:00:15
23.7 mph 306° true Trackpoint N47 50 26.2 W122 04 39.6 3/15/03 9:24:41 AM 332 ft 00:00:09
24.1 mph 313° true Trackpoint N47 50 26.7 W122 04 41.0 3/15/03 9:24:44 AM 113 ft 00:00:03
25.6 mph 300° true Trackpoint N47 50 29.8 W122 04 47.1 3/15/03 9:24:58 AM 520 ft 00:00:14
26.3 mph 307° true Trackpoint N47 50 30.8 W122 04 49.4 3/15/03 9:25:04 AM 188 ft 00:00:06
27.3 mph 302° true

As you can see it does a fairly good job of reporting the changes in direction and speed. If my
direction and speed remain constant, the frequency of reporting get stretched out. Most roads have a
fairly large radius curves. If you have an 11.5 degree arc with a 500 foot radius the difference
between the arc length and the chord length is about 0.16 feet. This is in a nominal 100 section of
road. Most roads that I work with and ride on have a radius that is usually measured in the
thousands of feet. In cases like that the difference becomes even less.

Having said something about the theoretical side I would advise a person to use the approach
suggested by Cletus. That should give you a good reference for the accuracy of the unit you use.

William Higley, Sr. Vision R-50 RANS Rocket

"Cletus D. Lee" <[email protected]> wrote in message
news:[email protected]...
> In article <[email protected]>,
> [email protected] says...
> >
> > "Cletus D. Lee" <[email protected]> wrote in message
> > news:[email protected]...
> > > In article <[email protected]>,
> > > [email protected] says...
> > > >
> > > > "Cletus Lee" <[email protected]> wrote in message
> > > > >> Another choice, a Garmin Geko. uses GPS for speed and distance.
No
> > wheel
> > > > magnet. It does under report speed by about 1.5% over a well
calibrated
> > > > wired computer.
> > > >
> > > > My Garmin Geko doesn't under report speed--it's right on.
> > >
> > > I am sorry, You are correct about the speed. Mine is accurate to the .10 mph. Except for the
> > > time it registered 888 mph in downtown
Houston
> > > while I was stopped at a traffic light. It is distance that is under reported by ~1.5%. The
> > > reason the distance is off is because the Geko measures distance from point to point. A route
> > > with a lot of turns
will
> > > be in error. Also I have found that false signals are causing route track errors in city
> > > canyons.
> > >
> > > --
> > > Cletus D. Lee Bacchetta Giro Lightning Voyager http://www.clee.org
> > > - Bellaire, TX USA -
> >
> > Well, I'm not so sure about the trip distance being under-reported.
When I
> > first bought it, just for an experimentation I brought it along and
watched
> > it update while my wife was driving down the freeway. Assuming the
survey
> > mile markers were correct, I didn't see any noticeable error at all. I
also
> > used it when I rode a century (actually a 108 mile ride), every rest
stop
> > was exact on my Garmin, as was the final distance when compared to the cheat-sheet. It was not
> > 1.5 miles off, it wasn't even .15 miles off.
But
> > heck, I guess you could also argue that they probably used some kind of
GPS
> > to calculate and make the cheat sheet to start with.
> >
> > Maybe you know for sure, but I theorize the Garmin does not simply triangulate between points
> > (thus cutting corners as you indicate). I believe the Garmin somehow integrates velocity, and
> > thus doesn't
necessarily
> > cut off corners. The reason I believe this is I've seen the Garmin
continue
> > to update trip distance for a several seconds (at least) after the GPS signal is lost.
>
> What I know is from observation. Take a look at the track produced by your Garmin. Compare on a
> map against the roads actually taken. I think you will find the route is a series of chords. The
> distance along those chords is the distance measured and reported by the GPS. The Chord is always
> (by definition) shorted than the curve that it spans. This is the error that I refer to. It is not
> much but it accumulates. More accurate on straight roads, less so on routes with lots of turns.
>
> As for your 108 mile Century, if the sponsors did anything other than actually chain off the
> actual distance, then it is an approximation.
>
> If you want to really see how accurate your Garmin really is at distance, take it to a High School
> track. At cycle speeds, do several miles of a known distance (caution the track may be metric).
> See how that compares to the known distance.
>
> I did this with my Magellan once and was lucky to get three points per lap. I would have sampled
> more often if I had been walking.
> >
>
> --
> Cletus D. Lee Bacchetta Giro Lightning Voyager http://www.clee.org
> - Bellaire, TX USA -

 

[Howard .]

"William Higley, Sr." <[email protected]> wrote in
news:[email protected]:

> I agree with Cletus about chords being shorter than arcs. I would try his method of testing the
> distance of a GPS unit at a High School track.
>
> My Garmin GPS III+ takes readings when I change speed or direction. The following excerpt from one
> of my journeys (this was in an uphill portion of the route) shows how frequently it will mark a
> point on a track. When I change direction or speed it stores a "Trackpoint", provides the latitude
> and longitude of the point, shows the time it created the point, then shows the distance, elapsed
> time, speed, and heading to the next "Trackpoint" Trackpoint N47 48 39.4 W122 00 44.8
> 3/15/03 9:10:33 AM 36 ft 00:00:02 12.2 mph 278° true Trackpoint N47 48 39.4 W122 00 45.1 3/15/03
> 9:10:34 AM 17 ft
> 00:00:01 11.7 mph 270° true
<edit for brevity>
> 00:00:06 21.3 mph 302° true
>
> As you can see it does a fairly good job of reporting the changes in direction and speed. If my
> direction and speed remain constant, the frequency of reporting get stretched out. Most roads have
> a fairly large radius curves. If you have an 11.5 degree arc with a 500 foot radius the difference
> between the arc length and the chord length is about 0.16 feet. This is in a nominal 100 section
> of road. Most roads that I work with and ride on have a radius that is usually measured in the
> thousands of feet. In cases like that the difference becomes even less.
>
>
>
A couple of observations if I may.

Altitude is bizarre in GPS speak. For giggles google geoid. Beyond that, the earth model used by a
particular GPS can have an impact on accuracy, particularly given your surface route is in 3
dimensions.

Further be aware of the distinction between "taking a reading" and "storing a datum." While there is
an entire newsgroup dedicated to GPS, here are some basics.

Garmin supports two families of near real time reporting protocols, one proprietary and one per the
not-as-tight-as-it-could-be NMEA standard. Per NMEA, it will report each "sentence" every 2 seconds.
I don't remember the frequency of the proprietary messages, but it's something similar (not a useful
improvement for most applications. Its chief advantage seems to be in a more tightly written spec).
Given the RAM inside the device is finite, if you were saving vectors it would stand to reason to
skip any successive vectors where the data (excluding timestamp) were the same, and let the plotting
software figure it out. Think of it as a finely detailed cue sheet - the changes you don't make
aren't recorded. In any event, you'll have no saved updates closer together than some arbitrary
minimum time which is quite likely some multiple of the actual calculation frame. Anything less than
500ms would surprise me on a low-cost unit, with 1 sec being a more expected delta time. My hunch is
that they use the same "engine" on most of their GPS receivers, but dumb 'em down for non time
critical applications (read non aviation, non-NASA, and non-military). I'd suspect the device is
running through the position calculations every 20-1000ms; that's just a hunch though.

"guess where you are now!" Howard

bitshift blah blah.

 

[Mike Schwab]

In my property description, there is a description of a rail road right of way. It starts 1/2 miles
south of my house, 1/4 mile west, going east, and 2 degree turn to the north until some station
number, and somehow comes across my property. The only way that makes send is if it is about 1/4
radius turn.

"William Higley, Sr." wrote: <snip>
> Most roads have a fairly large radius curves. If you have an 11.5 degree arc with a 500 foot
> radius the difference between the arc length and the chord length is about 0.16 feet. This is in a
> nominal 100 section of road. Most roads that I work with and ride on have a radius that is usually
> measured in the thousands of feet. In cases like that the difference becomes even less.
<snip>
> William Higley, Sr. Vision R-50 RANS Rocket

 

[William Higley]

Railroad curves are treated a bit differently than road curves. A 2 degree railroad curve will have
a radius of approximately 2865'. The 2 degrees refers to the interior angle that well be subtended
by a 100' chord. Most modern highway curves use a 100' arc length definition. The older highways
used the railroad definition because at the time, most Engineers with route layout experience came
from the railroads.

I may have caused confusion in my original statement when I spoke of an 11.5 degree arc. More
properly stated it should have read "a curve with a 500' radius and a central angle of 11.5
degrees." This would provide an arc length of approximately 100'.

William Higley, Sr., P.L.S. Vision R-50 RANS Rocket "Mike Schwab" <[email protected]> wrote in
message news:[email protected]...
> In my property description, there is a description of a rail road right of way. It starts 1/2
> miles south of my house, 1/4 mile west, going east, and 2 degree turn to the north until some
> station number, and somehow comes across my property. The only way that makes send is if it is
> about 1/4 radius turn.
>
> "William Higley, Sr." wrote: <snip>
> > Most roads have a fairly large radius curves. If you have an 11.5 degree arc with a 500 foot
> > radius the difference between the arc length and the chord length is about 0.16
feet.
> > This is in a nominal 100 section of road. Most roads that I work with
and
> > ride on have a radius that is usually measured in the thousands of feet.
In
> > cases like that the difference becomes even less.
> <snip>
> > William Higley, Sr. Vision R-50 RANS Rocket

 

[Dean Arthur]

"William Higley, Sr." wrote:
>
> Railroad curves are treated a bit differently than road curves. A 2 degree railroad curve will
> have a radius of approximately 2865'. The 2 degrees refers to the interior angle that well be
> subtended by a 100' chord. Most modern highway curves use a 100' arc length definition. The older
> highways used the railroad definition because at the time, most Engineers with route layout
> experience came from the railroads.
>
> I may have caused confusion in my original statement when I spoke of an 11.5 degree arc. More
> properly stated it should have read "a curve with a 500' radius and a central angle of 11.5
> degrees." This would provide an arc length of approximately 100'.
>

Hmm, two radii of 2865' with outer ends 100' apart = two degrees of subtended arc, right?

Now, two radii of 500 foot length and included angle of 11.5 degrees has 100' arc length?

Scribble, scribble, erase, erase, scribble...

Tune in next year for response, if I can find my way back!

 

[Mike Schwab]

Dean Arthur wrote:
>
> "William Higley, Sr." wrote:
> >
> > Railroad curves are treated a bit differently than road curves. A 2 degree railroad curve will
> > have a radius of approximately 2865'. The 2 degrees refers to the interior angle that well be
> > subtended by a 100' chord. Most modern highway curves use a 100' arc length definition. The
> > older highways used the railroad definition because at the time, most Engineers with route
> > layout experience came from the railroads.
> >
> > I may have caused confusion in my original statement when I spoke of an 11.5 degree arc. More
> > properly stated it should have read "a curve with a 500' radius and a central angle of 11.5
> > degrees." This would provide an arc length of approximately 100'.
> >
>
> Hmm, two radii of 2865' with outer ends 100' apart = two degrees of subtended arc, right?
>
No. At 100 feet distance from starting point, the curve is 2 degrees to one side of the straight
ahead line. Repeart 45 times for 4500 feet along the circumference, you will have turned 90 degrees.
90 times for 9000 feet 180 degrees, 180 times 18000 feet 360 degrees. C = 2 PI r
18000 /2 = 9000, 9000 /3.1415926 = 2864.789 feet radius. That distance makes sense for this railroad
to touch my property.

> Now, two radii of 500 foot length and included angle of 11.5 degrees has 100' arc length?
At 100 feet, the curve is 11.5 degrees to one side of the straight ahead line. Reapeat for 15.65
times for 1565 feet for 180 degrees divide by PI 3.1415926 = 498.15 feet.

 

[John Tserkezis]

RVC wrote:

> I've just purchased a Stratus and wonder if any improvements have been made in the wireless
> computer world ? I've read the archives and it appears the main problem(s) is/are transmission
> distance and durability. I can mount by Garmin eTrex Vista on the bike as a solution, but would
> prefer a bike computer.

> Any recommendations, to include waiting for improvements!

I've mentioned this 'trick' on another cycling group some time back, but unless you're
specifically looking for it, you probably won't stumble across it with google.

I had serious problems with the Polar HRM and the wireless wheel sensor on my upright. I found the
Polar transmitter/receiver set is somewhat directional, and the wheel transmitter just happened to
be positioned in the worst possible orientation- which also was the only possible place it could
be mounted. It was unreliable at best even with the transmitter power turned right up.

I solved the problem by using a passive antenna retransmitter. A simple solution involving only
wire. A coil around the transmitter, another similar coil around the receiver, and the two joined
by a wire pair.

Use thin wire, not critical at all, but the thinner the better (less space). Wind about ten turns
around two fingers (large enough to fit over the transmitter) and a similar one near the receiver.
Join the two coils together with a pair of wires twisted together.

The idea being that the transmitter signal is received by the coil around it, the signal travels
along the wire pair to the other coil, and is effectively re-transmitted to the nearby receiver.
Sure this is not an efficient system, but plenty good enough for the purpose.

So much so, that I could turn the Polar transmitter powers to minimum, which saves on their
battery power.

The wheel and cadence sensors that used to be the least reliable, are now the most reliable. The
chest strap transmitter is now the relatively unreliable one, because it's still susceptible to
interference, the others now appear immune.

I did the same thing on the 'bent, works perfectly.

Disadvantages are cosmetic, as it's now wired. But basically, your
alternative is to have a beautiful bike with a computer that doesn't work.

--
Linux Registered User # 302622 <http://counter.li.org