From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.tech
Subject: Re: cycle computer
Date: 21 Feb 2000 23:41:33 GMT

Jon Isaacs writes:

>> How do cycle computers calculate altitude?

> I believe that the two common Cyclometers the discontinued Avocet 50
> and the Cateye both measure the air pressure and use that to
> calculate the altitude.

These instruments use barometric pressure and to measure altitude
during changing weather requires adjusting the current altitude that
can vary by 100m from fair to stormy weather.  Because most users are
interested in accumulated climb rather than current altitude, this
setting is less often used, the change in altitude being unaffected on
first order by the zero setting.  That is because most important
places, like mountain passes or peaks, are marked by a sign.  The
AVO50 makes use of a patented accumulator that is essential for
reliable climb accumulation in that it does not record bumps in the
road that are less than 10 meters (RR overpasses) while recording
continuous climbs entirely.  Cateye does not infringe and also gets
different results.  Avocet is bringing its Vertech wrist watch back to
market in the near future and after that hopes to get a new Cyclometer
out that has altitude and a string of new features.

> The Ciclomaster IIA altitude measurement was different and rather
> hokey IMHO.  It used the angle of the bike and the velocity and used
> that to measure calculate the altitude.  The one I had (have) was
> not reliable or consistent.

A pendulum on a moving vehicle does not know where "down" is and
therefore, the Ciclomaster wad doomed to fail in recording altitude.
I find interesting that such a company could go all the way to market
with such a bizarre idea, and beyond that that no one in the bicycle
press questioned it.

> Perhaps Jobst can discuss the actually sensor in the Avocet 50, I
> would be interested in know the what sort of sensor techniques are
> used.

New and better sensors are now available, ones that are moisture
proof, the bane of all AVO50's.  The pressure sensor is a silicon chip
with an evacuated void spanned by a semiconductor strain gauge.  Newer
chips have the strain bridge immersed in oil under a membrane.  That
way, even under water, moisture cannot can reach the strain sensors or
chip.

I have no dates for introduction of any Avocet products.  We'll have
to wait and see.

Jobst Brandt      <jbrandt@hpl.hp.com>

From: jbrandt@hpl.hp.com (Jobst Brandt)
Newsgroups: rec.bicycles.tech
Subject: Re: cycle computer
Date: 22 Feb 2000 15:55:40 GMT

Jon Isaacs writes:

>> New and better sensors are now available, ones that are moisture
>> proof, the bane of all AVO50's.  The pressure sensor is a silicon chip
>> with an evacuated void spanned by a semiconductor strain gauge.  Newer
>> chips have the strain bridge immersed in oil under a membrane.  That
>> way, even under water, moisture cannot can reach the strain sensors or
>> chip.

> Thank you for answering my questions.  Are these chips also used in other
> applications such as aircraft altimeters??

Yes, these are generic pressure sensors that come in various pressure
ranges.  The ones used in the altimeter are designed for zero to
somewhat over one atmosphere pressure.

Jobst Brandt      <jbrandt@hpl.hp.com>

From: jobst.brandt@stanfordalumni.org
Subject: Re: Computer w/altimeter
Newsgroups: rec.bicycles.tech
Message-ID: <W7d5a.67478$Ik.2910409@typhoon.sonic.net>
Date: Thu, 20 Feb 2003 23:02:14 GMT

Peter Headland <PHeadland@actuate.com> writes:

>> Others such as the Cateye are optimistic in this regard

> Uh? If I climb 3 feet, I climb 3 feet and I expend energy so to do.
> Measuring that accurately isn't "optimistic", it's a statement of
> fact.

Oh you don't say.  So if I ride from my house to San Jose, both at
elevation 60ft and no hills in between, you want to register the more
than 100ft of 0.25% grades along the route as elevation climbed.
That's only the more apparent problem but atmospheric variations
usually cause more than that in pressure fluctuations even though the
barometer at the start and finish of such a ride is also unchanged.

> OK, I grant you there has to be some lower threshold, otherwise we
> would be measuring every ripple in the pavement, but 10m is excessive
> in my view. The resolution on most of these instruments is generally a
> small (single digit) number of feet (so they don't detect rollers
> smaller than that), and that seems to work tolerably well.

The 10 meter threshold was chosen because it eliminates RR underpasses
or overpasses, something most athletically inclined people take as not
a climb, all of it being an optionally anaerobic effort.  For folks who
don't ride mountains, this may seem a Draconian theft of
accomplishment but these folks don't need an altimeter anyway.  The
instrument is designed to give an accurate representation of climbing
over a mountainous route and it does that with greater accuracy than
others.

> The real problem is jitter (where the reading is near the boundary
> between one increment and the next and very small undulations or
> perturbations in air pressure can cause measured elevation changes
> of several feet).  Certainly a good algorithm would ignore changes
> of less than a single increment (but perhaps Avocet's dubious patent
> prevents other makers from so doing - if so, thank you Avocet).
> More annoying still are those instruments that have an over-long
> sampling interval (20 seconds) and perform cumulative calculations
> based upon those snapshots - you can go down and back up quite a
> long way in 20 seconds.

Are you basically opposed to Intellectual Property (IP) protected by
copyright and patents?  That is a discussion subject that is widely
discussed in appropriate forums.  When I proposed the 10m threshold in
the design, others did not understood why that might be valuable,
indicating to me that the feature should be patented since it was not
obvious to others involved in designing the barometric accumulator.
This feature also allows the Vertech wristwatch to count ski runs, an
equally patentable application.

http://www.avocet.com/vertechpages/vertechski.html

> If your complaint is that the accumulated climb is not equal to the
> difference between the elevations at the top and bottom of a climb,
> I think you are being unrealistic - a climb such as our beloved Page
> Mill road here in the SF peninsula undulates quite a bit, and you
> certainly have to work to regain the losses.  FWIW, my Suunto has a
> "stopwatch" mode that can measure the difference between the start
> elevation and the current elevation, in addition to the accumulated
> elevation change.

I should turn your comment around because insignificant barometric and
elevation dithers are not what most healthy bicyclist wants to record
when characterizing a route.  In fact, without rejection software,
other instruments record significantly greater elevation gains even on
monotonic climbs with level sections.

> The bottom line is that these instruments are plenty accurate enough
> for their purpose; expecting perfection (even if you can define what
> that is in this instance) is foolish.  The fact that one maker's
> device may give a slightly different result than another's is not
> significant.

So why are you making a fuss over it?

Jobst Brandt
jobst.brandt@stanfordalumni.org
Palo Alto CA