At 09:50 AM 12/16/2012, Jones Beene wrote:
Very silly error by Naudin (discovered up by George Holz on the MAGH forum)
in that JLN was using a 5% duty cycle and "calculated" the gain by reducing
the input readings - instead of measuring them directly.
Naudin's error was a multiple of 20 (reciprocal of the duty). He deflated
both the voltage AND current readings instead of just the voltage; so there
was a systemic underestimate of P-in by a factor of twenty. The calorimetry
was good. Actual P-in was near 85 watts, had he done the correct
calculation.
In the end, there was probably slight gain - not 21:1 but more like 1.1:1
(COP of 1.1)
Or no gain.
In any case, I wanted to verify what Jones wrote. In short, yes. Very
silly error, but compounded, seriously, by simply ignoring the
problem once it was raised.
That mailing list was interesting.
http://tech.groups.yahoo.com/group/aH-gen/messages/11?l=1
Started by Sterling Allen. Early participants included familiar
names. Dennis Cravens, cold fusion researcher. Jones Beene himself.
Dennis Cravens raised some question about input current values.
http://tech.groups.yahoo.com/group/aH-gen/message/27
Robin van Spaandonk raised the issue more extensively.
http://tech.groups.yahoo.com/group/aH-gen/message/32
Jones points out a problem related to input power.
http://tech.groups.yahoo.com/group/aH-gen/message/40
George Holz nails the error.
http://tech.groups.yahoo.com/group/aH-gen/message/41
Robin confirms this. http://tech.groups.yahoo.com/group/aH-gen/message/44
Jones explained the input problem more clearly:
http://tech.groups.yahoo.com/group/aH-gen/message/112
Several people commented on the error made by averaging voltage and
averaging power and multiplying the average figures together. If
there is a five percent duty factor, and you have direct current,
when it's on, you have V and I, for voltage and current, but then
switched for the duty factor, the average voltage is V * 0.05 and
average current is I * 0.05, and the product of voltage and current
is V * I * 0.05 * 0.05, whereas the actual average power is V * I *
0.05. If Naudin did in fact calculate the input power this way,
multiplying average voltage by average power, he understated input
power by 0.05, which happens, then, to be close to what he measured
as dissipated power.
Naudin reported, in fact, RMS voltage and RMS power. Those are
average measurements. For DC, they *are* average voltage and power.
Naudin used, it appears, RMS meters.
Yet no response was obtained from Naudin on this crucial question,
which essentially, if true as to what he did, blows the whole result
of major excess power out of the water. Completely. End of story. And
at least one page on the experiments shows that this is exactly what
Naudin did, see http://jlnlabs.online.fr/mahg/tests/index.htm. He
reports, in his chart, RMS voltage and RMS current. Those will be,
for pulsed DC, the same as average voltage and average current, and
Naudin does multiply them. That page, though, shows an apparent input
power waveform of a half-vave rectifier.
That isn't the same as the pulsed square-wave DC being considered by
the people on the list. However, the same error is being made. It is
effectively switched AC with a 50% duty cycle. If there is no phase
shift, the measurements will be off by a factor of two. That is
*very roughly* the XP being shown on the page.
It's an appalling error to be still standing without correction, as
Jones has noted.
Knowing this field, it is not suprising that discussion continued on
the mailing list about what might be the cause of the very possibly
non-existent anomaly.
Okay, here is the test that the people on the list were talking about:
http://jlnlabs.online.fr/mahg/tests/pultests.htm -- yes. 5% duty
cycle, with DC, input power calculated by multiplying measured RMS
voltage and power. Error, then, by a factor of twenty.
In the long chart, one can see that the "Efficiency" is always less
-- substantially less -- than the input power measurement error,
which will be 1/DTC (duty cycle).
There is a nifty plot of "efficiency" vs. "duty cycle," which shows
the trend, toward infinite efficiency at 0% duty cycle. That should
have been a clue, but sometimes you need to hit people with a shoe to
wake them up.
Jones, responding to a person who had posted possible libel of Naudin
commented: http://tech.groups.yahoo.com/group/aH-gen/message/177
This potential problem (with the duty cycle) is NOT a cunning
trick, and was pointed out weeks ago here as a likely problem. Why
don't you take the time to read the past messages?
That observation was communicated to Naudin a month ago by way of
Moller. They responded that they trust the measurements but there
could be problems in the semantics (translation from French), But
they do not believe that the issue of a "reciprocal" of the duty
cycle, i.e. 5% = 20 is the problem. Others have mentioned this
here previously. In testing the device in Dec 2003, when the duty
cycle was 50% they achieved OU of claimed 250%, which even if
halved is still overunity. Are we to believe that in the
intervening 18 months, that nothing was accomplished?
Jones, that is embarrassing. If the duty cycle was 50%, the error
introduced was 50%, leading to an expected "efficiency" of 200%.
Above, I happened to first look at that study, it seems. The actual
results were not "250%." That was the high end of a series of
results, which ranged from 144% to 282%. That averages only a bit
over 200%. If there is an anomaly here it is small. On the page
showing much higher "efficiency" the results are dismal, after the
error is corrected, there is no overunity, instead they are missing
lots of power. Maybe it is being stored up in the ZPE field?
Look, smart people can make mistakes like Naudin. But, above, they
were informed, and "how did you measure your input power" is kind of
hard to lose in translation. "We trust the measurements" would
rationally mean that they trust what their instruments were showing
them, but that was *not* the question. It was how the measurements
were converted into an estimate of input power.
Yes, with my spectacular hindsight, I can affirm that "nothing was
accomplished in the intervening 18 months," except that the most
obvious bonehead you can make in running experiments like this, it's
well-known, was maintained without correct, and as far as I can tell
so far, is still being maintained. Making a mistake, anyone can do
that. Even spectacular ones. But allowing the error to continue? That
demolishes any respect that the researcher might have enjoyed. Hey,
nice clean lab, pretty instruments ... but, sorry, no good.
It looks like Naudin never responded. Some people questioned the
alleged error. It's really amazing how people can fool themselves.
The error is a basic one, obvious to any electrical engineer. It's
easiest to understand with the switched DC. Naudin reported RMS
voltage and RMS current. For pulsed DC, the RMS voltage will be the
peak voltage times the duty cycle. The RMS *power* will be the peak
power times the duty cycle. If you multiply the RMS voltage times the
RMS current, you will be multiplying the duty cycle twice, not once.
That introduces an error of 1/DTC, the inverse of the duty cycle.
It's truly basic, and this is not complicated. But some made it so,
for themselves, using resistance, etc.
The mailing list became largely a repository of spam. One of the last
substantial messages was from Terry Holmes, August 30, 2006.
http://tech.groups.yahoo.com/group/aH-gen/message/309
At the Tesla Tech convention in Salt Lake City on July 28, Thorsten
Ludwig of the German Association for Space Energy gave a
presentation on Naudin's MAHG device, which he (Ludwig) said had
been replicated by his group. His presentation showed various
photos of Naudin, Moller and himself with Naudin. He also claimed
that Moller was in the process of producing 300 of the tubes like
the one Naudin had used in his publicized results. Ludwig hoped to
acquire one or more of these tubes for his group to use. Moller
told him that they were expected to be ready in the next few weeks.
His presentation pretty much consisted of the 2005 Naudin results
with photos from the web site from last summer--complete with the
Van Nostrand quotes implying that the energy of combination of
hydrogen was 90,000 times greater than the energy of disassociation
(?!).
That's the error just mentioned earlier today. Looks like people
repeat "information" without checking. Hey, guy gives a presentation
at a conference, must be an expert, eh?
(That was probably misreported. The dissociation energy for H2 - 2 H
is 104 KCal/nole. The reported figure from Langmuir (actually only
one of his figures) was 90,000 cal/mole. As publicized by Moller, the
encyclopedia purportedly show 104 cal/mole. What's three decimal
places among friends? This is Moller's trope, so Ludwig was repeating
it, probably.)
After the lecture I told him about the concerns that some
calorimetry people in the U.S. had tried to communicate to Naudin
and Moller which were met by only silence--an inappropriate response
in my view. He was surprised to hear the Naudin did not
communicate, but himself did not evince any interest in what the
calorimetry questions might have been.
Who cares? "We trust the measurements," what, do you think Naudin was lying?
No. He made a bonehead mistake in calculating input power, that's
all. Obvious from the experimental report, once you know where to look.
He also confirmed that
Frolov and Naudin were in basic disagreement about the technical
issues involved (feud was the word I suggested). One thing we did
agree on was that the 300 devices being produced were likely to
settle the matter of whether there was anything here or not.
Well, did they?
I'm going to close this post here, but I intend to see what else is
out there on this mess. Did Allen write about this debacle?
