FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Help Needed: Wire Temp Hypothesis

Written by Ryan Hunt on .

We started a run heated with both wire in 3.5 Bar of 75%H/25%Ar.

When we turned up the temp with both wires we saw the impedance drop first, then start to rise again. When we review Run 1 on Nov 12, we see that we were absorbing quite a bit at T_mica=156 and Current in the Celani wire = 1.7 Amps. So we decided to run both wires so that we were running 1.7 amps through the Celani wire, but heating with both wires. We will see what this does to the loading (if anything). Then we will turn off the NiCr wire and see if it continues to absorb at that current and a lower temperature. We also saw that the impedance during that run was way lower than we currently have. Maybe it has to continue to absorb more. Or maybe those parts of the wire that didn't turn black again are not changing impedance as much.

Anyway, when we stepped the power down, I noticed something in the shape of the impedance graph.  It might lead us to be able to estimate the wire's operating temp.  I propose that the wire drops in a second or two to the temperature inside the cell.  The size of that drop might be used to calculate what temperature the wire was operating at.  

The power drop:

The cell temperature drop:

So, I need help from all the wonderful commentators to figure out how to correlate this back.  Any suggestions?

Add comment


Security code
Refresh

Comments   

 
0 #22 Arnold Isenberg 2012-11-22 03:49
You must think of the "wire" as a very non-homogenous metal mixture at the surface, due to the selective oxidation of the elements, i.e. the copper diffuses out of the constantane alloy first and forms an oxide, which, when reduced to pure copper does not diffuse back to the physical place where it was before oxidation. At this place the constantane alloy is irreversibly destoyed, this mechanism forms the "layers". The more layers, the more destroyed constantane. In turn, this may form the thin layers of active structures. I'm looking at this subject only from a well understood metallurgical process point of view. So, when you talk about the Celani-wire do not think of it as a uniform resistor, it may vary from experiment to experiment, depending of is "thermo-chemica l-history". Therefore, oxidation time, gas composition, etc. must be very much controlled. "Reworking" the wire may be successful in reaching your set goal, however repeating it will be a bear. Good luck.
Quote
 
 
0 #21 Ecco 2012-11-21 22:47
@Rats: have a look here: lenr-canr.org/.../...
Quote
 
 
0 #20 Rats 2012-11-21 22:40
This upwards trend for P_Xs is promising but is still too small to determine if it is indeed excess energy. Does anyone know if Celani's cell also demonstrated a similar trend i.e. a very slow gradual increase?
Quote
 
 
0 #19 Ecco 2012-11-21 22:20
@Ged: a change in room conditions? Ambient temperature has increased over the previous two hours. That might have been accompanied by air movement removing heat from the reactor. Maybe there are people working/doing something in its proximity.

This however would show that if we are to measure excess heat in the order of a few watts (although as I mentioned previously, this isn't what was originally expected), then testing conditions must be kept as stable as possible, and the reactor has to be isolated as much as possible from external influences. The kind of calorimetry currently used is ok for large anomalies, but it's too fragile for small ones.

*By the way, I'm curious to know too if that camera flash thing has actually been attempted!
Quote
 
 
0 #18 Ged 2012-11-21 21:46
Hm, seems the outer glass is suddenly cooling. I wonder what is up, can't really see a reason. Either wait, or maybe blast that thing with a camera flash/other perturbation just to see what happens, since it seems so borderline at the moment?
Quote
 
 
0 #17 Ascoli65 2012-11-21 19:16
Hi, very good job! My suggestion.
Could Celani's wire resistance mainly depend from chemical/morpho logical transformation?
The so called repairing treatment oxidizes the wire. The black color is typical of CuO (1), whose resistivity is much higher than metal. In Cu-Ni-Fe alloy, coppers can be found at higher concentration on the intergranular matrix and CuO formation has the lowest heat of formation among other oxides (2). Oxidation of the wire increases its resistivity.
During the so called loading phases, the H2 gas progressively reduces the CuO to metallic Cu and resistivity drops to the original value. The reaction rate increases with temperature, starting from about 200 °C (3).
1 - en.wikipedia.org/.../...
2 - sciencedirect.com/.../...
3 - pubweb.bnl.gov/.../...
Quote
 
 
0 #16 Ron B 2012-11-21 19:01
Ryan,
Celani reported gamma bursts that lasted about an hour when the wire temp was about 100C. Have you had the chance to monitor it on this run?
Quote
 
 
0 #15 Ged 2012-11-21 18:52
The T_mica is still well below (by ~100 C) the temps we were thinking we needed to get to, and yet we are starting to see these events. This is quite interesting.
Quote
 
 
+1 #14 Ryan Hunt 2012-11-21 18:38
It is still tough to tell if we are seeing real energy excess because of hte change of pressure. We just upped the power to 48 W, so we can make a better comparison against both Celani's original report and against our calibration runs. If the reaction goes faster with higher temps, as Celani stated, then this should improve it.
Quote
 
 
0 #13 Robert Greenyer 2012-11-21 15:57
@ All

Too early to call - but this might be an experiment to watch
Quote
 
 
0 #12 Sanjeev 2012-11-21 15:52
@Ryan Hunt

Since your last voltage change, pressure dropped by 1 bar and T_Mica increased by 3`C. I know its not much to conclude anything, but the vessel can be filled up to 4.5 bars after say 24 hrs and if the T_Mica still keeps rising, we have something.

Anyway, I would say let it run till P_Xs touches +2W approx. Then you can play with pressure if needed at all.

P_in has actually gone down a bit due to rising impedance, still the temperature is rising, thats amazing.
Quote
 
 
0 #11 Ecco 2012-11-21 15:26
[portion about previous run edited out due to my mistake]

What I think is happening currently is that both sensors are increasing in reading and trying to reach each other as internal gas pressure decreases and convection is decreasing. The rest (external glass very slightly heating as well) might be due to the system slowly trying to reach thermal equilibrium.

ATM there's nothing yet suggesting me that LENR is occurring; current cell behavior doesn't look that interesting either. Keep in mind that unambiguous signals were originally expected, so calling a potential couple of watts of apparent (calculated) excess heat a success it's too premature at this stage.

@Ryan Hunt: this might be the right time to try "stimulating" the cell with the camera flash suggestion I proposed last time. That increase in P_xs was quite anomalous. Have you already tried? I know it sounds like a dumb idea, but...
Quote
 
 
0 #10 Ryan Hunt 2012-11-21 14:22
It looks very interesting, doesn't it? We are going to compare the current conditions against the Calibration runs 4, 5, and 6 which were in this gas blend at various pressures in order to have a comparison of how to expect the temperatures to change as the pressure drops in the cell.
Quote
 
 
+1 #9 Sanjeev 2012-11-21 13:15
In last 12 hours T_Mica has gone up by 2.5°C.
From the trendline of P_Xs, it should go above 0 in next 12 hours.

T_GlassOut is also clearly increasing.
Quote
 
 
0 #8 Andy 2012-11-21 09:47
I love this project. It looks like you have a great team. Best of luck!

I had a side-thought about an easy treatment for the wire - rub the wire down with ferric chloride? I've seen some papers about the synthesis of nickel nanoparticles using FeCl3 - so maybe this could cause nanofracturing of the wire?
Quote
 
 
0 #7 Sanjeev 2012-11-21 08:03
If you plot the data from 20/11/2012 23:00, something good is happening it seems.

T_Mica has risen by 1.5`C, while T_Amb and P_in are constant and for the first time P_xs is showing a continuous upward trend, not random.

Impedance is not dropping, so its not the usual loading phase. Note that in Celani's graph the impedance was also rising during excess energy phase.

Only thing that dropped is pressure, so the temperature increase can be due to pressure drop (less conduction). But I think you are near the sweet spot at this time. (hopefully)

Is there any calibration curve for T_Mica/pressure , so we can see how much pressure drop causes how much increase in T_Mica ? If the increase is greater than calibrated value, then you have excess heat.
Quote
 
 
+1 #6 Carta 2012-11-21 07:46
Thank you Robert for the aswer.

Why not ask Celani for an already tested (i.e. excess heat in Celani lab) wire ?

In turn, you may send your probably damaged wire to Celani and he could verify if it works or not.

This should tell if the problem is in the wire or in the measurement set up (of your reactor or the one of Celani)

Is it feasible?

Thanks and keep on!!

Ric
Quote
 
 
0 #5 Pekka Janhunen 2012-11-21 06:08
(Copied from previous orphan page:)
Two questions:
- Does daylight have access to the room, I guess not?
- How quickly was vacuuming/press ure reduction done, in comparison to Celani? If pressure is reduced too fast, maybe hydrogen outgasses violently, perhaps it kills the wire.
Quote
 
 
0 #4 Robert Greenyer 2012-11-21 05:05
@ Carta

Not that we are aware of - straight from manufacture
Quote
 
 
0 #3 Carta 2012-11-20 22:39
Just a question:
Did Celani tested the wire before send it to you?

Thanks

Ric
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