SV: data collection and storage.

[Bernd Felsche]

espen hilde tapped away at the keyboard with:

> Thanks a lot for all suggestions.My problem is that I am not able
> to say this suggestion is better than this.....its a little above
> my head , please try to expaine like you where trying to tell a
> 10year old.....  if you got the time...

Does it help if you jump up and down? :-)

The values which would be of interest must be 'sampled' (measured)
as often as is characteristic for that type of value. As engine
speeds can change very rapidly, that's probably the one you want to
measure as often as possible. Around 10 times a second is not
unreasonable. At the other end of the scale, things like inlet
temperature change slowly, once every second is probably 'too often'
given the slow response of most temperature sensors.

Once you've figured out what you might want to measure and how
often, you can decide on the size and speed of the data capture
equipment. In general, make sure that the system can cope with
one-third more than what you plan, either 'out of the box' or by way
of expansion.

The crucial values are the total storage bandwidth, resolution and
capacity.

The bandwidth and resolution required is determined by the number
and type of values ... e.g. engine speed as an 8-bit value, 10 times
a second giving 80 bits per second. Similarly, an analogue value
such as inlet temperature could be sufficient as a 10-bit value,
once every second i.e. 10 bits per second. Add them all together.

A 10-bit value is more easily on word boundaries; usually
bytes-boundaries so such a number would require 2 bytes == 16 bits
of storage and hence increases the bandwidth.

Say you then get a total of 320 bits/second (bps) - you can then
calculate the capacity of storage required, determined by the time
you want to be logging data. e.g. for 10 minutes == 600 seconds ==>
192,000 bits or 24 kbytes. Don't forget to leave some space for
'headroom' in capacity. That headroom provides for management
overheads such as tagging samples with time-stamps, or the like.

Make sure that individual data sources can be measured as quickly as
you need... i.e. if the gadget does 3 samples per second on an
analogue input, you won't be able to use it to log data that needs
to be measured 100 times a second. 

It's all fairly logical once you crack the jargon and understand teh
fundamentals.

> The first races we got 5 and 6 places ,did a owerhaul on the
> engine ,a 15 hp Yamaha,worked on tilt angles ,this summer we
> straightened out the buttom and made sharp edges to the
> straiks.Did some 4places in the first day of a race in Sandefjord
> 2 weeks ago, next day we borrowed a propeller from a competitor,
> The boat got away like all other stod still......2place first heat
> 1 place 2 heat....WOW.  That was fun! Wery difficult to give that
> propeller back...how much..???do you want for it??

> The propeller did not give any more top speed just better
> accleration, Need to know more about propellers and need some good
> tools to measure the difference in real life....not forgetting to
> have full control on the engine combustion.

> Any who can say :this is the easiest ,best way of logging the datas?

You might also want to add some accelerometers - a minimum of three
is required to resolve all directions. It's just about the only way
you can quickly resolve the motion of a boat. GPS, even without
selective availability (SA) doesn't have the resolution especially
if you're changing direction quickly.

Sensor should be placed as far apart as possible for best accuracy,
avoiding local vibrations. e.g.  place one under the deck in the
bow, and one either side near the hull in the stern, as far as
possible from the prop, shaft and motor.

A bit of simple arithmetic with the figures from the sensors
will give you an idea of forward acceleration achieved for a
particular engine speed and propeller type. 'Integration' of the
acceleration figures will give you an idea of hull speed. That's not
as complex as it sounds ... it's a simple accumulation of numbers,
then multiplying by a constant and the 'time'.

An illustration of how simple that is, there's an automotive gadget
you can buy (IIRC it's a G-Tech) that gives you those sorts of
figures. You attach the device to the dashboard or windscreen and
take the car for a drive to obtain 'performance' figures. The
quality of the results depends on how level the road is, how
well-attached the unit, and how straight the road.

-- 
Bernd Felsche - Innovative Reckoning
Perth, Western Australia
----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org