FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Impedance vs Time in 3 Helium Runs

Written by Ryan Hunt on .

So far we have not seen any clear excess energy out of the cell during this Helium test.  Since our baseline was from a run with the exact wire being tested here, we did get a nice, close calibration curve, though.  More accurately, we did not see any significant deviation in temperatures from the base line run.  

One of the results from last night's run of the Loaded Celani Wire in Helium is that we got a glimpse into the impedance changes of the Celani wire pre-loading and post loading.  That's what I want to talk about now.

This is the results for the very first run with the Celani wire after installing it in the cell.  With every increase in power, the impedance went up in a nice step.  It looks like the first 6 steps maintain a positive slope, though.  The last 3 steps level out more at the top.

The second run was done at at 0.5 bar.  Notice a difference?

The step in the middle of the range displays a strong negative slope.  That occurred approx T_mica = 228C.

Then on this run, we saw the following curve.

A couple of interesting things here.  The downward slope appears again on the 4th and 5th step.  This was about 225 and 245C on T_mica, though the wire was probably higher temp.  On the highest step (T_mica=265) the slope continued positive, though.  Upon reducing the power back down to match the 225 T_mica, again, the impedance dropped well below the previous level at that temperature.

Questions these data raise:  

If impedance is a measure of loading, is the wire somehow gaining Hydrogen in this Helium environment?  It is possible there was some residual hydrogen available in the cell.  Perhaps it was dissolved in the NiChrome wire, or the copper wire, or the Mica wire supports.  

Or is it a migration of Hydrogen with the wire? Or is it some other metallurgical thing?  It's possible it has little to do with hydrogen and may be a function of cycling the wire temperatures.  Thoughts?

Why did we see some "loading" in the Run2He?  Again, maybe residual hydrogen from other things in the cell.  When we installed the Celani wire, we also replaced the Nichrome wire, so that probably was not a factor, but the other items may have been.

What is with the sudden drop as the power is turned on at the very beginning?  This I think we have an answer for.  At that time, the wire was powered by the power supply in automatic mode.  This new power supply (that replaced the one we damaged yesterday) has an improved control interface that actually goes all the way to zero when asked to.  That makes the impedance measurement very inaccurate.  Additionally, the sensor inputs that measure the current and voltage have an option for a zero offset calibration that we have never had to worry about.  Now calibrating that offset is on our to-do list.  All these little improvements add up after a while to some real good technique.

It'll be interesting to see how the impedance changes by morning.  Once again, I am up past my bedtime.  Good night.

Add comment


Security code
Refresh

Comments   

 
0 #21 HarleyJuicy 2018-07-31 22:16
Hello. I see that you don't update your page too often. I know that writing posts is boring and
time consuming. But did you know that there is a tool that allows you to create new posts using existing content (from article directories or other blogs from your
niche)? And it does it very well. The new articles are high quality
and pass the copyscape test. You should try miftolo's
tools
Quote
 
 
0 #20 Robert Greenyer 2012-11-16 21:13
@ Jfab

Interesting suggestion

B
Quote
 
 
0 #19 jfab 2012-11-16 19:15
Have you considered asking Celani to come to your lab with his own experiment setup and apparatus for a couple of days? You could test your measuring instruments on his cell?
I'm sure some of us would be keen to donate for its travel expenses. I would.
Quote
 
 
0 #18 Ryan Hunt 2012-11-16 16:02
Actually, I don't take off the shield. But I did notice that the IR gun was leaving a thin streak of film on the cell from the plastic melting where it touched. :(
Quote
 
 
0 #17 Ecco 2012-11-16 16:00
My mistake, then.

The process of taking off the lid and measuring temperatures does cause a slight drop in cell temperatures, though. Or at least, it did each time you said you were doing so.

Excess power calculation currently seems in line with that of the previous power step.
Quote
 
 
0 #16 Ryan Hunt 2012-11-16 15:39
@ Ecco: No, I did not get close to the cell at all this morning.
I did increase the power to see if the excess power increases with the temperature and the increased current.
Quote
 
 
0 #15 Ecco 2012-11-16 15:27
Let me guess: after increasing input power a bit, have you have measured wire temperatures with the IR probe? It appears that every time you do this (not the measuring itself, but removing the lid, staying in front of the reactor, etc), reactor temperatures drop a bit in a [now] typical manner.
Quote
 
 
0 #14 Robert Greenyer 2012-11-16 14:13
@ Ron

In
www.nist.gov/.../jpcrd221.pdf

Page 235 is interesting too. It is a long read!
Quote
 
 
0 #13 Robert Greenyer 2012-11-16 14:00
@ Ron

We have been discussing the potential for annealing to modify or disable the Celani wire. I guess as we start testing more wires in both the US and EU, running SEM on them afterwards - we will build up a pool of evidence we can dip into to understand what is happening. With the live and near live publishing of raw data and images and the help of people like those interested in looking at it on here we will together form a good understanding of what is going on.
Quote
 
 
0 #12 David Jones 2012-11-16 13:54
Interesting paper

lenr-canr.org/.../...


We speculate that the electrically self-heated wire method, is probably helpful in
obtaining higher loading values.
Quote
 
 
0 #11 Ron B 2012-11-16 13:40
Thanks to everyone for sharing ideas, this is great!
There's some research on alloy resistance across temperature that may be interesting to some.
www.nist.gov/.../jpcrd221.pdf
On page 53 there are graphs of electrical resistivity of Nickel/Copper alloys at different ratios across temperature.
There's also some interesting information about annealing of Ni in the temperature range that the wire has seen already. (just something to look at while waiting on the experiment! : )
Quote
 
 
0 #10 Ecco 2012-11-16 13:34
@Chauvin: not related to your hypothesis, but after reading recent Celani presentations, yesterday I found out that he typically loads Hydrogen at a starting pressure of around 7 bars, which is higher than the 3.5 bars typically used by the MFMP. This would increase convective losses toward the glass tube, but on the other hand wire temperature would also be more constant along its length, which coupled with higher hydrogen pressure, might significantly help loading and therefore showing anomalous behavior.

I'm also thinking that once an active wire has been well-loaded with Hydrogen at a high pressure, afterwards it might not be necessary anymore to keep it at 7 bar.

So I'm asking: are there plans to try this "high pressure path", which anyway would be more faithful to Celani's testing conditions? Can your reactors withstand them?

(There are also potential safety issues with this, btw)
Quote
 
 
+1 #9 Nicolas Chauvin 2012-11-16 13:19
It is possible that at this particular temperature, the hydrogen protons loaded in the wire get a higher mobility in the nickel lattice.
This would reduce the wire resistivity, even if no additional hydrogen/proton s are loaded in the wire.
Quote
 
 
0 #8 David Jones 2012-11-16 13:07
@ Robert - not a problem..

I keep forgetting to tell you how to simply add a large ballast vessel. We used to use an empty gas bottle (big nitrogen ones) and connect directly to the laser (in your case the cell). In this way you have a very large reservoir (>1000:1 ratio) filled to the gas pressure you require and you can simply use this to accurately refill and keep your cell at the required pressure.
Quote
 
 
0 #7 Robert Greenyer 2012-11-16 12:00
@ David

Good points as ever David - this is why we planned the internal replications first, to identify and iron out any doubts or bugs in the system in a public way to form a tight understanding of what affects certainty and how to deal with it.

Thanks for your diligent attention.
Quote
 
 
0 #6 David Jones 2012-11-16 11:26
My point is; until you address the calibration errors you can not positively identify any excess anomalous power output. Any possible anomalous heating point has to lie outside any calibration spread error to be identified with any certainty.
Quote
 
 
0 #5 David Jones 2012-11-16 11:20
Curve of “Total Power vs T_GlassOut Rise” is beautifully smooth. The standard deviation of points to the fitted curve is very small with minimal random error. This implies that your current, voltage and temperature measurements suffer minimal errors. But, a number of repeat calibration runs at the same pressure (first pump out and refill) would show a large spread in the points – as shown in previous calibration runs. This implies poor control and recording of gas pressure in the cell. One way to test this is to pick one power input point and do a standard deviation run by repeating a pump/fill cycle at the same pressure. The same test should be done in air for completeness – known pressure. This should show if the error lies with poor monitoring of the gas pressure. One way to address this issue is to connect a metal cylinder of volume very much greater than your test cell to smooth out the gas fill errors and also to calibrate the accuracy/sensit ivity of your pressure sensor.
Quote
 
 
0 #4 Robert Greenyer 2012-11-16 08:52
@ Alex

There appears to be a step down in ambient at that time - but pressure and impedance are both dropping.
Quote
 
 
0 #3 Ivone M. FitzGerald 2012-11-16 08:42
Guys, it's early days yet, but this null test is a strong indication yet that the regular beat of between 5.5 watt and 7.5 watt over a cycle of eight minutes that you had earlier is what you are looking for. Right now it's random noise. Next time aim for 10 watts average, and I think you can allow yourself some self congratulation. And then what? Manufacture your own New Fire wires, (rhymes) with a thick rough appearance, make say a thousand, pump 500 watts through them, get 10 KW, and you're away!
Quote
 
 
0 #2 Alex 2012-11-16 08:41
Hello;
Sorry for my poor english.

What happens to power-out on October 16 at 7:05 am? (HUGnetView™ time)
Sudden increase of 2.5 W ?

Thanks for your great job

Regards
Quote
 

Here is your generous contributions so far towards our $500,000 target, thanks everyone! : $45,020   Please Donate
See the current state of our booked costs here