next message in archive
next message in thread
previous message in archive
Index of Subjects
Steve Shaw wrote:
> Most likely with only 91 samples (not that many), it is going to look
> noisy, but
> when plotted, the result is actually quite striking and rather
> simple. The
> counts at all the even-numbered sample points (178, 180 etc) give a
> reasonable,
> normal-looking distribution of values centered on Dave = 180 lb, a bit
> skewed
> towards the low end. By contrast, the counts at all the odd-numbered
> samples
> (181, 183...) are relatively small and most are actually zero. So the
> scale appears
> to be counting in twos (almost).
Hi Steve, Peter, Patrick & All, Dec 5, 2006
The executive summary, as they say, is battery fatigue. And the
short answer to 'why bother ?' is (a1) that much of our experience now
passes through digital filters and (a2) much of nature is a puzzle.
This tendency for a low incidence of odd weights was what led me to
wonder about this in the first place and tabulate some of them.
The various comments, including 91 being too few weights, got me
past the energy barrier enough to tabulate all weights and look for
possible mechanisms. Fot the total of 254 weights, only 47 were odd; the
full array being--
192,1; 191,0; 190,1; 189,1; 188,3; 187,0; 186,4; 185,2; 184,11; 183,1;
182,21; 181,2; 180,29; 179,8; 178,24; 177,4; 176,43; 175,10; 174,16;
173,4; 172,21; 171,9; 170,8; 169,4; 168,16; 167,2; 166,8.
To proceed further, some shorthand is indicated. The question is
"V>x ?" represents "Does the output voltage of the transducer exceed the
voltage that would be generated by a weight of x pounds ?" [Note that
bathroom scales don't measure mass but weight. If you were in orbit then
your mass would remain unchanged but the bathroom scales would say
correctly that your weight was zero (if they worked correctly)].
For a thumbnail reminder of how numbers are represented by digital
switches and a key to additional jargon that I will use in this
schematic representation of possible mechanism-- The output register is
assumed to be 8 digital switches in which switches 1 to 8 (bits 1 to 8)
represent the numbers 1, 2, 4, 8, 16, 32, 64 & 128 respectively. Note
that all numbers between 1 & 255 can be represented by combinations of
these 8 values. If for example only bits 8 and 6 are set to 1, then the
binary number is 00000101 which represents 32+128= 160.
This unit runs on a 9 volt battery so circuit operations must be kept
to a minimum. If the circuit logic worked from low to high (Is V>1 ?, if
yes set bit 1; Is V>2 ?, If yes reset bit 1 and set bit 2...) it would
use a lot of power asking stupid questions, activating redundant
circuits and flicking gates on and off many times.
Therefore the circuit should work from high to low--
Is V>128 ? If yes then set bit 8 & goto B if no then is V>64 ?...; (B)
Is V>(128 + 64) ? If yes then set bit 7 & goto C..., if no then is
V>128+32 ?... The outcome of each test determines whether a bit will be
set or not and also determines which branch of the circuit will next be
used. So the last bit to be processed will be bit 1.
If output voltage decreases, due to battery fatigue, while these
necessary 8 questions are satisfied then even if there initially was
enough output voltage to set bit 1 with some left over, in addition to
all the other set bits, the voltage may have dropped too low to set bit
1 by the time question 8 is asked.
One could test this idea by wiring a second 9 volt battery in
parallel with the usual. This would take a fair amount of haywire, so as
a first step I am running this past Naturens and as a second action will
run it past a friend who has extensive electronics background.
Yours truly, Dave Webster, Kentville
next message in archive
next message in thread
previous message in archive
Index of Subjects