Re: GPS accuracy question

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

[Sorry for cross-posting to u.r.c partway through a discussion. I had
intended to cross-post from the start, but obviously forgot. I've left in
more context than I would normally as a result. ]

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Alan Adrian" <[email hidden]> wrote in message

news:<[email hidden]>...

Quoted message said:

Quoted message said:

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...
> I've recently acquired a Geko 201 and started using it while

cycling.

Quoted message said:

Quoted message said:

Quoted message said:

>
> The Geko routinely shows a current speed 0.3-0.5 mph slower than the

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster

than

Quoted message said:

that on the Cateye. Sorry for the confusion.

Quoted message said:

Quoted message said:

Cateye,
> but at the end of a journey the trip computer shows an total trip

distance

Quoted message said:

Quoted message said:

> of a few hundreths of a mile longer than the Cateye.
[...]
> Is there an obvious explanation?

I would guess that your cateye is misscalibrated.
[...]

And if you had a few turns with some cover, you would see some

distance

Quoted message said:

Quoted message said:

Quoted message said:

gained on the Geko due to overshoot at the corners.

I must check the tracklogs against a map and see how bad that effect
might be. And try some long straight line journeys, where this
shouldn't happen. There's a dead straight 2 mile stretch nearby.
Presumably I shoudln't get this effect there.

Unfortunately, I did. Geko and Cateye agree within 1% on overall distance
and on average speed over the journey, but the geko instantaneous speed is
3-5% different from the Cateye's. This on a dead straight road with a

clear

Quoted message said:

view of the sky all the way.

My inclination is to assume that the calibration of the cateye is correct,
given that the overall distance recorded is pretty much the same, and that
there's an explanation for why the speed is different. I just don't know
what the explanation is!

Is there anybody out there with a GPSR and a cycle computer that agree

about

Quoted message said:

everything? Or are at least consistently disagree?

Thanks,
Steve.

Over the 2 miles was your speed constant? If not is it possible the cateye
or GPS has some form of smoothing - I know on my GPS you can set some form
of smoothing and the length of period that is done over.

Eric
--

Steve Rumsby said:

[Sorry for cross-posting to u.r.c partway through a discussion. I had
intended to cross-post from the start, but obviously forgot. I've left in
more context than I would normally as a result. ]

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Alan Adrian" <[email hidden]> wrote in message

news:<[email hidden]>...

Quoted message said:

Quoted message said:

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

>I've recently acquired a Geko 201 and started using it while cycling.
>
>The Geko routinely shows a current speed 0.3-0.5 mph slower than the

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster than
that on the Cateye. Sorry for the confusion.

Quoted message said:

Quoted message said:

Cateye,

>but at the end of a journey the trip computer shows an total trip

distance

Quoted message said:

Quoted message said:

>of a few hundreths of a mile longer than the Cateye.

[...]

>Is there an obvious explanation?

I would guess that your cateye is misscalibrated.
[...]

And if you had a few turns with some cover, you would see some distance
gained on the Geko due to overshoot at the corners.

I must check the tracklogs against a map and see how bad that effect
might be. And try some long straight line journeys, where this
shouldn't happen. There's a dead straight 2 mile stretch nearby.
Presumably I shoudln't get this effect there.

Unfortunately, I did. Geko and Cateye agree within 1% on overall distance
and on average speed over the journey, but the geko instantaneous speed is
3-5% different from the Cateye's. This on a dead straight road with a clear
view of the sky all the way.

My inclination is to assume that the calibration of the cateye is correct,
given that the overall distance recorded is pretty much the same, and that
there's an explanation for why the speed is different. I just don't know
what the explanation is!

Is there anybody out there with a GPSR and a cycle computer that agree about
everything? Or are at least consistently disagree?

Two things that might affect your results:

- I don't know how your GPS works, but if it calculates speeds/times/etc
in fixed short intervals, it's effectively turning a smooth track into a
series of short straight sections.

- ionospheric irregularities will change the position from GPS by up to
a few meters during periods of solar activity.

Quoted message said:

Quoted message said:

Quoted message said:

> I've recently acquired a Geko 201 and started using it while

cycling.

Quoted message said:

Quoted message said:

Quoted message said:

>
> The Geko routinely shows a current speed 0.3-0.5 mph slower than the

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster

than

Quoted message said:

that on the Cateye. Sorry for the confusion.

Quoted message said:

Quoted message said:

Cateye,
> but at the end of a journey the trip computer shows an total trip

distance

Quoted message said:

Quoted message said:

> of a few hundreths of a mile longer than the Cateye.
[...]
> Is there an obvious explanation?

I would guess that your cateye is misscalibrated.
[...]

And if you had a few turns with some cover, you would see some

distance

Quoted message said:

Quoted message said:

Quoted message said:

gained on the Geko due to overshoot at the corners.

I must check the tracklogs against a map and see how bad that effect
might be. And try some long straight line journeys, where this
shouldn't happen. There's a dead straight 2 mile stretch nearby.
Presumably I shoudln't get this effect there.

Unfortunately, I did. Geko and Cateye agree within 1% on overall distance
and on average speed over the journey, but the geko instantaneous speed is
3-5% different from the Cateye's. This on a dead straight road with a

clear

Quoted message said:

view of the sky all the way.

My inclination is to assume that the calibration of the cateye is correct,
given that the overall distance recorded is pretty much the same, and that
there's an explanation for why the speed is different. I just don't know
what the explanation is!

Is there anybody out there with a GPSR and a cycle computer that agree

about

Quoted message said:

everything? Or are at least consistently disagree?

Thanks,
Steve.

Steve, this is a coincidence. Just finished a cycle with a milometer and
gps. Milometer says I cycled 8.25 miles, but the gps says I cycled 10.3
miles. Sat down to compose a post about this and saw yours. I bought my bike
second hand and the milometer was already attached to it. I did not get any
instructions with it and I am not 100% certain it is set right. Can anyone
tell me the correct positioning of the contacts on the wheel and spar for
say a 27" wheel, or does it differ?

Rory

Electronic "milometer" (computer)? The positioning of the magnet and pickup
don't matter. They generate a pulse every time your wheel goes around.
It's up the electronic computer to know (you tell it during setup) how far
each pulse is in distance.

I'd try searching for your computer on the net and see if you can find an
instruction book for it online. They work much like a digital watch, you
hold down a certain button in a certain screen and then enter the correct
numbers (from a table) to setup the computer to your wheel diameter.

If it's one of those old mechanical ones with multiple screws hitting a
wheel, I completely forget how they work, sorry.

Al...

"Rodders" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

Steve, this is a coincidence. Just finished a cycle with a milometer and
gps. Milometer says I cycled 8.25 miles, but the gps says I cycled 10.3
miles. Sat down to compose a post about this and saw yours. I bought my

bike

Quoted message said:

second hand and the milometer was already attached to it. I did not get

any

Quoted message said:

instructions with it and I am not 100% certain it is set right. Can anyone
tell me the correct positioning of the contacts on the wheel and spar for
say a 27" wheel, or does it differ?

Rory

Rodders said:

Steve, this is a coincidence. Just finished a cycle with a milometer and
gps. Milometer says I cycled 8.25 miles, but the gps says I cycled 10.3
miles. Sat down to compose a post about this and saw yours. I bought my bike
second hand and the milometer was already attached to it. I did not get any
instructions with it and I am not 100% certain it is set right. Can anyone
tell me the correct positioning of the contacts on the wheel and spar for
say a 27" wheel, or does it differ?

Think about it - the computer counts wheel rotations, so the positioning
does not matter, so long as you have just the one magnet, and that it
accurately detects each revolution.

You need to calibrate the computer by feeding it either your wheel
diameter or circumference, depending on model. Here's a good way to do it:

1) Inflate your tyres to normal riding pressure, and load the bike up
with a normal amount of gear.

2) Select a smooth dry pavement somewhere.

3) Draw a thin line of something like washing-up liquid across the pavement.

4) Ride over the line as straight as you can, and keep going for at
least one wheel revolution, preferably two or three.

5) Measure the distance between the original line and the others put
down by your tyre, dividing by the number of additional lines to get an
average value for greater accuracy. This gives you a value for your
wheel circumference - to get diameter, divide by pi (3.1416 is good
enough!).

6) Program the required value into your computer.

--
Mark.

Quoted message said:

If it's one of those old mechanical ones with multiple screws hitting a
wheel, I completely forget how they work, sorry.

From what I remember they work the same as a computerised milometer - count
the wheel revolutions, and in the case of the mechanical version the
internal cogs will be specific for a particular wheel size, and no
calibration is possible. Do they still sell these?

Quoted message said:

Quoted message said:

Quoted message said:

"Steve Rumsby" <[email hidden]> wrote in message
>I've recently acquired a Geko 201 and started using it while cycling.
>
>The Geko routinely shows a current speed 0.3-0.5 mph slower than the

Steve Rumsby said:

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster than
that on the Cateye. Sorry for the confusion.

That does put rather a different complexion on the problem.

Quoted message said:

Quoted message said:

Quoted message said:

>but at the end of a journey the trip computer shows an total trip distance
>of a few hundreths of a mile longer than the Cateye.

Quoted message said:

Quoted message said:

Quoted message said:

>Is there an obvious explanation?

I would guess that your cateye is misscalibrated.

eg: reading short and hence slow.

Quoted message said:

Unfortunately, I did. Geko and Cateye agree within 1% on overall distance
and on average speed over the journey, but the geko instantaneous speed is
3-5% different from the Cateye's. This on a dead straight road with a clear
view of the sky all the way.

Recorded distance on the GPS may be under the true distance if you have
any dropouts; that is if it loses satellite reception at any time during
the trip.

If the cateye was calibrated 5% low and the GPS was losing 3% to 4% of
true distance, the cateye would read 5% slower, but only 1% to 2% less
distance.

Rodders said:

Steve, this is a coincidence. Just finished a cycle with a milometer and
gps. Milometer says I cycled 8.25 miles, but the gps says I cycled 10.3
miles. Sat down to compose a post about this and saw yours. I bought my bike
second hand and the milometer was already attached to it. I did not get any
instructions with it and I am not 100% certain it is set right. Can anyone
tell me the correct positioning of the contacts on the wheel and spar for
say a 27" wheel, or does it differ?

If it is counting successfully then moving the parts should not change
the calibration. It counts whole turns of the wheel, not "distance" of
wheel going past.

"Graham W" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

Steve Rumsby said:

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster

than

Quoted message said:

Quoted message said:

that on the Cateye. Sorry for the confusion.

That does put rather a different complexion on the problem.

Yeah, sorry. I knew what I meant, but the connection from brain to fingers
was obviously under strain when I typed the original message.

Quoted message said:

If the cateye was calibrated 5% low and the GPS was losing 3% to 4% of
true distance, the cateye would read 5% slower, but only 1% to 2% less
distance.

My problem with this explanation is that the conditions on the road the few
times I've done a comparison have been pretty much perfect for GPS
reception. Clear sky, no tree cover, etc. The Geko clearly thinks is has
good reception because it is claiming accuracy of <20ft. I know that's not
an indication of real accuracy, but its estimate does get worse in bad
conditions (I've seen it 100ft+ when under tree cover) so it does mean
something.

If the Geko is really losing that much true distance in what seem like ideal
conditions, I'm a little disappionted. Of course, that could be might due to
my unreasonable expectations of a GPSR from knowing little about them. I'd
much rather believe I'd calibrated the Cateye incorrectly:-)

If the Geko was losing distance in this way, how would it show up in the
tracklog? Would that have a distance longer than the distance on the trip
counter (because the tracklog filled in the gaps with straight lines that
weren't counted in the trip distance)? That would be easy to check. I
normally get breaks in the tracklog if reception is lost completely (when
walking in dense woods) and that isn't happening here.

Anyway, I'll repeat the experiment a few more times on the same stretch of
road as the opportunity presents itself, and see what happens.

Sorry if this seems like making a mountain out of a molehill. I don't like
it when experiment reveals major flaws in my understanding of something,
even in something that isn't really that important!

Thanks,
Steve.

In sci.geo.satellite-nav Steve Rumsby said:

If the Geko is really losing that much true distance in what seem like ideal
conditions, I'm a little disappionted. Of course, that could be might due to
my unreasonable expectations of a GPSR from knowing little about them. I'd
much rather believe I'd calibrated the Cateye incorrectly:-)

In general, the GPS will have very good velocity measurements at steady
state. I would expect that the slower it goes, the worse the distance
(odometer) measurements are.

Is the course you took one that you could drive a car on? If it's not
too twisty, I would run the same course with the car at higher speeds
and see how much the distance measurement changes. I would expect the
higher speed reading to be more accurate.

--
Darren Dunham [email hidden]
Senior Technical Consultant TAOS http://www.taos.com/
Got some Dr Pepper? San Francisco, CA bay area
< This line left intentionally blank to confuse you. >

"Steve Rumsby" <[email hidden]> wrote

<snip>

Quoted message said:

Is there anybody out there with a GPSR and a cycle computer that agree

about

Quoted message said:

everything? Or are at least consistently disagree?

I've got a Garmin GPS12 (at least 5 years old) and a cordless CatEye.

I find that the GPS distance over a 10+ mile ride is always a few percent
lower than the computer distance. This isn't due to the GPS not getting a
lock until I've been cycling for a while because I let it get a lock while
I'm getting ready to go out.

The GPS speed and computer speed are usually the same +/- 0.1 m.p.h.

I calibrated my computer wheel circumference setting carefully but I've not
checked the accuracy over a known distance so it could be a bit out. I've
had other computers on other bikes with the same GPS and the GPS distance
reading has always been consistently lower than the computer value.

At one time I assumed this distance difference was due to the GPS estimating
the distance travelled by a series of straight lines between the points in
the track log. The real path between two points wouldn't be a perfect
straight line. I have my doubts about whether this would introduce much of
an error though.

--
Best Regards
Dave

e-mail [email hidden]
web www.dlowther.demon.co.uk/local.htm

Top posting - sorry

Used to know the answer to this but it's a bit (lot) hazy now. One thing I'm
sure of is that there is a proper technical answer for it.

.... but it goes something along the lines of it's down to the way speeds are
calculated by the two devices - the cateye counts revolutions - times these
by wheel circumference and thus calculates distances which are accurate -
the distance is then used along with a given time to calculate an average
speed for the instantaneous speed readout.

The GPS calculates speed slightly differently and gives the same correct
distance reading but averages the speed a slightly differently way - can't
remember quite how and cant remember which is the more accurate but the
distance for both is accurate. I think the cateye might be the more accurate
but couldn't be sure

Or something like that - I'm sure I picked this up on s.g.s-n years ago -
have you tried the info on pete mafferty's site ? Does that still exist?
been a long time since I lurked in s.g.s-n

HTH
Russ

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

[Sorry for cross-posting to u.r.c partway through a discussion. I had
intended to cross-post from the start, but obviously forgot. I've left in
more context than I would normally as a result. ]

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...

Quoted message said:

"Alan Adrian" <[email hidden]> wrote in message

news:<[email hidden]>...

Quoted message said:

Quoted message said:

"Steve Rumsby" <[email hidden]> wrote in message
news:[email hidden]...
> I've recently acquired a Geko 201 and started using it while

cycling.

Quoted message said:

Quoted message said:

Quoted message said:

>
> The Geko routinely shows a current speed 0.3-0.5 mph slower than the

Oops. I started off with the wrong information - where I said "slower" I
actually meant "faster". The indicated speed on the geko is 3-5% faster

than

Quoted message said:

that on the Cateye. Sorry for the confusion.

Quoted message said:

Quoted message said:

Cateye,
> but at the end of a journey the trip computer shows an total trip

distance

Quoted message said:

Quoted message said:

> of a few hundreths of a mile longer than the Cateye.
[...]
> Is there an obvious explanation?

I would guess that your cateye is misscalibrated.
[...]

And if you had a few turns with some cover, you would see some

distance

Quoted message said:

Quoted message said:

Quoted message said:

gained on the Geko due to overshoot at the corners.

I must check the tracklogs against a map and see how bad that effect
might be. And try some long straight line journeys, where this
shouldn't happen. There's a dead straight 2 mile stretch nearby.
Presumably I shoudln't get this effect there.

Unfortunately, I did. Geko and Cateye agree within 1% on overall distance
and on average speed over the journey, but the geko instantaneous speed is
3-5% different from the Cateye's. This on a dead straight road with a

clear

Quoted message said:

view of the sky all the way.

My inclination is to assume that the calibration of the cateye is correct,
given that the overall distance recorded is pretty much the same, and that
there's an explanation for why the speed is different. I just don't know
what the explanation is!

Is there anybody out there with a GPSR and a cycle computer that agree

about

Quoted message said:

everything? Or are at least consistently disagree?

Thanks,
Steve.

Quoted message said:

Sorry if this seems like making a mountain out of a molehill. I don't
like it when experiment reveals major flaws in my understanding of
something, even in something that isn't really that important!

GPS tells you where you are. Everything else your receiver tells you is a
builds on that. It will give you a very accurate distance if you record
your starting point and let it calculate the distance to your finishing
point.

As you ride your bike the receiver is taking positions at given time
intervals. At each one of these points in time your are going to get some
error due to the accuracy of the receiver. On your trip your receiver is
accumulating these small distances and also accumulating small errors. I
would think that if you measured the same course several times using gps,
you would get several different distances. I would think measuring the same
course with your bike computer several times would yield more consistent
results. I would think that a calibrated bike computer is more accurate for
measuring accumulated distance, and average speed. Hills will have a slight
affect also as gps measures horizontal distances and not along a slope as
your bike computer does.

A poor analogy would be to measure something fairly accurate like 440 yard
running track with a yard stick and see how close you are at the end.

Mark