From: John Tserkezis 
Reply-To: Fidonet AVtech Echo 

Bob Lawrence wrote:
>>(b) indicate, when the vehicle is travelling at a speed in excess
>>of 50 kilometres per hour, a speed that is not more than 10- less
>>than the actual speed, and

> JT> Thanks for that. At last, fact that isn't based on "so and so
> JT> told whoever that joe blow said...."

>  Did you actually read what it said? Firstly, Brenton misquoted, and
> second, it says that the speedo is not allowed to be correct. I tried
> to explain how it *actually* works in the real world in a message to
> Brenton...

  Yes, and I read the original via the link (I also had noted it was vague in 
that instance, so checked- he quoted correctly).

  I take it for what it's worth, written like a true Australian Standard, vague 
to the hilt.

>  Can you measure *road* speed using GPS? On a straight road, does the
> software do a speed calculation?

  Yes, it takes two consecutive point reads (and times), and calculates the 
horirzontal speed.  It does not cater for slopes, but on the road, you don't 
get enough of a steep slope to make a significant enough difference to note.

  Things to look out for, the fixes end up being about 1-2 seconds behind real 
life, so for fluctuating speeds it's useless, on the freeway, where you are 
steady for longer periods of time, it's quite right.  Some GPS recievers can 
average the speeds over a number of seconds, so it has even a slower update, 
but it's closer to the mark. (only small differences in this case though)

  Basically, I note what the speed differencial is at every 10Km/h, and use the 
car speedo because its update rate is good enough to use.

>  If it does, you have discovered the *only* way I know to do a simple
> speedo calibration.

  I can't claim discovery of this.  There are also provisos you have to account 
for while doing it too.

> Brenton's assertion that his speedo is "very
> accurate" is crap. I checked mine with a calibrated rev counter using
> gear ratios, but that doesn't allow for tyres and road surface. You
> either have to use a cop's laser or some way to actually measure road
> speed... not tailshaft speed. Timing a measured mile is usless.. I've
> tried it.

  Well, yes.  It only measures your speed _average_ over a mile.  Which is useless.

> JT> Yeah, they all use the standard frequency meter technique to
> JT> measure speed. 

>  Frequency meters measure *both* period and frequency.

  There are hardware implications if doing both, (easy to do, but still adds a 
little to hardare costs).  If you're using a microprocessor, there are no 
differences in hardware, but major differences in software.

> It's insane to
> measure frequency rather than time at that speed, especially when you
> only need three digits. I despair of enginerring in Australia
> today...

  It's cheaper. (in engineering costs, not hardware- hardware costs the same 
either way).  This is how scary it is: they measure that number of pulses over 
an experimentally determined time interval, to count nKm/h over whatever time 
period.  If there are not enough pulses (too long a time interval) they rig the 
sensors to output more pulses per revolution.
  This way they don't need to do any number crunching, it's just a direct 
display of a pulse count on the display.

  If you were to measure period, you would need to number crunch, as well as 
have a separate timer to handle the update rate. (so the update rate is a 
constant, separated from the current speed).

>  When we developed our own computer-controlled train, I was really
> impressed. I wondered where they got the engineers. Ten years ago, we
> were doing consumer-controls on washing machines and shit like that,
> and we could never get engineers or programmers who knew their arse
> from their elbow in the real world. How did they get engineers for a
> *really* complex problem like a train? 

  University graduates.  Sometimes you get lucky and find someone who knows the 
difference between the colour of their bodily fluids, but usually, you get 
someone who just did it because it sounded like a cool job.

  You cover your arse because you hired someone who was trained,and has a nice 
certificate in a handsome plastic frame.  Whether they can actually _do_ the 
work is another subject entirely.

  Can anyone say "millenium train"?

>  Now I know. The fuckers don't work...

  I'm not so sure it's as simple as that.  As far as I can tell the only ones 
left are the ones who are geek die-hards who wouldn't be caught doing anything 
else because they love it too much, or the wanna-bes, who are doing it because 
they are happy to take the shit pay for a job that requires extensive training 
and experience.
  I know plumbers who were quite cluey engineers.  Plumbing pays much better 
though.

-- 
       -o)
       /\\    Message void if penguin violated
      _\_V    Don't mess with the penguin

Linux Registered User # 302622                         http://counter.li.org>
Fido: 3:712/610  BBS/FAX: +61-2-9716-8310  Internet: jt{at}techniciansyndrome.org

--- ifmail v.2.15