At 01:44 PM 6/1/2010, Jed Rothwell wrote:
http://lenr-canr.org/acrobat/SartoFelectrodes.pdf
The excess heat value in Table I for Sample #64 was listed as 80 -
100%. Apparently that was a typo. It has been changed to >1000% (10
times input). That is a big difference! I am glad to hear they got
such pronounced heat, but I wonder how they managed to make such a
larger error in the manuscript.
This is an important paper, by the way.
Looks like it. But this is what is frustrating about a lot of
publications in the field, particularly with excess heat, but it
afflicts many other phenomena.
There are some experiments where it's only practical to do one or two
runs. Because, however, CF tends to be chaotically reproducible, for
a report to be truly meaningful, there should be a run of experiments
that are (hopefully) identical, and all the results should be reported.
1000% sounds great, but what if this was one or two cells in a series
of, say 64 cells? After all, it was sample #64! How impressive it is
would depend on all those other results! Now if someone runs 64
cells, and gets 500% excess heat *average*, I'd start to get excited!
But what seems to be reported is the highest result obtained, which
then leaves the field entirely open to charges of cherry-picking results.
And then, of course, other findings in the experiment series would be
reported as correlated or not correlated with the excess heat. Helium
is an obvious one, if collection and measurement of helium is
practical. But also radiation, palladium surface morphology after the
experiment (presumably the cathodes start out the same),
transmutation byproducts, etc.
It's easy to ascribe transmutation byproducts to contamination, and
contamination can be quite subtle. But if the byproducts turn out to
be correlated with excess heat.... As a chaotic reaction, it might be
all over the map, and transmutation isn't the primary reaction,
normally (except for the formation of helium!), but if no heat -> no
excess trace elements, bingo.
unfortunately, again, what is often done is to vary the conditions
all over the place in the hope of finding the magic combination. From
an engineering perspective, that's an error. The first goal should be
to get something reliable even if small. Then one starts varying the
conditions, in baby steps. At that point, one might set up a series
of cells, each with one variation. Then, when something looks
promising, that variation is repeated in a substantial series to
confirm that it wasn't just statistical variation, the chaotic effect.
This is why I'm so keen on designing a cell that is cheap, because
cheap means that one can run a lot of them if using them for serious
research, as distinct from my other market, people simply interested
in a demonstration. I'm thinking that, after the work with standard
SPAWAR cell types, I'll make the cell even smaller. I'd rather have
two small cells with an active 0.5 cm of cathode wire, compared to
one cell with 1 cm of cathode.
I guess I'm just oppositional....
