FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Is it, or isn't it?

Geschrieben von Ryan Hunt am .

What everyone wants to know is:  Are we REALLY making excess heat?

Current Conditions:  H2@3.5 Bar (starting pressure) and heating with the Nichrome (close to 22 ohms), and T_ambient of 23C

Comparisons for Current Conditions:
We have a number of comparable results, but none is a perfect control.  The post warming loading, the closest to current conditions could be demonstrating excess energy.

Run Name Description Gas Heating Wire Ambient
Early Calibration A calibration test done before beginning with Isotan 44 H2 @ 3.5 Bar NiChrome @ ~25 ohm 24 =>21
Cal 7 With Isotan H2 @ 3.5 Bar Isotan @
~16 Ohms
22.2 =>23.2
Run  He 1 Just after installing Celani Wire (1 wrap short on the end) He @ 3.5 Bar Celani Wire @ ~18 Ohms 25 =>24
Post Loading Warming Stepped increase in power like a calibration, but after loading the wire H2 @ 3.5 Bar Nichrome @ ~21 ohms 19.5 =>19.7

Compared on Graphs

The major uncertainty comes from not having the wire installed in exactly the same manner.   Otherwise, it looks promising that we are above the other baseline calibrations and right at the same point or higher than as the test with the Helium, which has a lower thermal conductivity and usually makes the T_mica run hotter.  

If we saw the temperatures and excess power slowly climbing, that would be another good clue.

Time to change something.  Higher power to try to trigger?  Lower power to attempt more loading?  Or switch to heating from Celani Wire?  Or attempt to open it up and re-wrap the wire more like the control wire was?

And again, thank you for all the wonderful suggestions and insights from the comments.  I regret not having the time to engage them all in discussion, but I do read them and will re-read them as we design the next experiment.

Add comment


Security code
Refresh

Comments   

 
0 #17 Al Potenza 2012-11-16 18:29
IIRC, Celani did say that the wires are hard to make. I wonder if there is an effort on his part or yours to devise ways to ease that process or perhaps automate it somehow and maybe generate more, lots more, of those wires?
Quote
 
 
+2 #16 Ryan Hunt 2012-11-16 02:30
Two main reasons. First, we have been focused on doing a replication of the experiment done with one wire. Second, we have only been gifted a precious few wires. We have not sought to try to make any ourselves.

But in the future there are literally hundreds of variations like that to try.
Quote
 
 
0 #15 Al Potenza 2012-11-16 02:27
I'm really curious: why do you not use MULTIPLE Celani wires? Maybe 3? 5? 10? Then the signal to noise ratio would be much better would it not? One might expect for a given input power, you'd get much more output power. After all, the input power should be doing nothing except heating up the wires.

Are the wires extremely rare or expensive to make? Is there some other reason not to increase their number? Certainly there is plenty of room on your mica form.
Quote
 
 
0 #14 Jack Cole 2012-11-15 18:57
I'd be interested to see what happens with getting it back up around 270 C, run for an hour, drop the pressure of the H2 down below 1 ATM for an hour, then bring it back up again.

I think static conditions are not helpful, and that maybe even Celani's hydrogen leak may have helped to trigger the reaction.

Here is a paper discussing problems with static equilibrium and triggering.

lenr-canr.org/.../...
Quote
 
 
0 #13 Ecco 2012-11-15 18:45
By the way, to not complicate things too much at this stage and help the system raise temperatures (as mentioned in other blog posts here) without applying too much power (either directly or not), you might want to try limiting hydrogen pressure to 1 bar or less (absolute). During NIWeek2012 and ICCF17 (I think) Celani had to do that because he wasn't allowed to bring containers of pressurized flammable gases to the venues. Yet, excess heat apparently was still triggered (although the active wire was already well-loaded), with his wire reportedly achieving the best excess heat results up to that time. Again, maybe asking Celani for data about those tests might be helpful (did the wire impedance anomaly show, etc).

I don't know if you have performed calibration runs at low gas pressures, though.
Quote
 
 
0 #12 David Jones 2012-11-15 18:35
Beaten to it by Rob L...
Quote
 
 
+3 #11 David Jones 2012-11-15 18:32
At some stage I would suggest you put insulation around the tube and endeavor to reduce the heating power input requirement to the 20W level – i.e. comparable to the expected anomalous heat power output.

You have nothing to lose doing this.
Quote
 
 
+2 #10 Rob L 2012-11-15 18:30
Can't tell - But it is probably a negative indicator that you don't see the Post Loading Warming tracking along the Cal 7 line and then jumping up when the temperatures gets up above 125°C (where if I recall correctly Celani saw onset of LENR heat).

This setup (and I understand the desire to mimic Celani as closely as possible) is inimical to doing clear demonstrations - it requires far too much power to maintain temperatures in the desired range, introducing large error bars, and requiring too many compromises in terms of limiting max temp, and introducing leaky seals. The sooner it is ditched in favour of a low power input high temp calorimetry method the better. But in the absence of anything else encapsulating most of the reactor in insulation leaving only small viewing windows would help to reduce the error bars by reducing necessary power input. (Or you could put the whole unit in a constant temp enclosure at 1-200°C - a deep fryer would do the job)
Quote
 
 
0 #9 Ecco 2012-11-15 18:20
It looks like live data isn't available anymore. Was it turned off on purpose?

Anyway, I agree with the choice of trying to find a temperature range which makes the sudden and anomalous decline of wire impedance to show. That is supposed to be a preamble for excess heat with Celani's wire, at least according to what I know. Maybe it might be a good idea to consult with Celani to know if with his latest wires excess heat has ever appeared without that happening first. If not, then, there might be problems with your wire.
Quote
 
 
0 #8 JOB001 2012-11-15 17:52
Sometime when you're not busy, provide a toggle to also show inverse kelvin temperature vs log watts which will have the neat advantage of showing self calibration for a wire at low temperature = negligible excess heat. The linearized slope change will precisely show where the reaction starts. google "Arrhenius equation graph" for many examples.
This graph method will allow calculation of the coulomb energy barrier also from the slope change.
Quote
 
 
+1 #7 Mitch 2012-11-15 17:46
From the peanut gallery, a suggestion. Unfortunately, it requires resources I cannot contribute towards.

You need your perfect control before going much farther. You should set it up as a simultaneous cell as close to the test cell as possible, and you should do it DOUBLE BLIND, so that only someone unconnected with your experiments knows which has the Celani wire and which does not. (If there are visual tells on the Celani wire, the experimenters should not have visual access to the wire.)

It's very easy for even sincere and honest investigators to subconsciously start interpreting ambiguity in their favor, so going double blind at the start will really help you build confidence that what you are seeing (if anything) is not wishful thinking.
Quote
 
 
0 #6 Ryan Hunt 2012-11-15 17:41
We are turning off the heat. We will top the gas off and then sweep through the expected loading temperature over the next few hours and see if we can find any drop in the impedance of the Celani wire, again. That will be roughly 120C to 300C, If we see any impedance decline we will let it go till it stops declining.

If we don't see any change in impedance, we will switch to the Celani wire and see if we see any rise in excess power.
Quote
 
 
0 #5 Ecco 2012-11-15 17:41
Personally, I would keep the more risky choices as secondary alternatives.

Rewrapping the Celani wire around the mica support without a spare one in case of damage (since it's brittle, etc) would delay experimentation and progress. Also, it might need to be planned in advance due to the time needed (I guess, unless you can virtually stop the reactor now and fitting the wire again in a couple hours).

However, if you choose to refit the Celani wire but you can't do it right now you can still experiment with higher temperatures in the meanwhile.

In either cases, if you decide to increase temperatures (doing this by increasing power to the heating wire seems the safest choice currently. There's quite some headroom left, I think) I would suggest monitoring manually from time to time with the IR probe the center portion of the Celani wire which as seen yesterday gets hotter than on its ends.
Quote
 
 
0 #4 charlie tapp 2012-11-15 17:34
hi just wondering not to throw a stick in the spokes i realy want this to work also even trying some things myself. is it possible on the very small amount of exess heat you saw that it could be caused by hydrogen embritalment ? some factories and industrial plants have to be carefull of this because it ruins metal the hydrogen finds its way into metal and tears it apart, looking at the celani wire under your electron microscopes kind of looks that way and when you distort a piece of metal it puts off heat is there any kind of calculation for that , also with the supposed ecat or even celani , if with the exess heat has anyone calculated the cost of making the hydrogen to cause the reaction into the cop just some quick questions good luck will be sending money on payday
Quote
 
 
0 #3 Andre Blum 2012-11-15 17:33
...either that or re-wrap the celani wire soon. You want that wrap tightness hypothesis behind you ASAP.
Quote
 
 
0 #2 Andre Blum 2012-11-15 17:29
One thing you might want to think about is whether you can further explore your theory about the tighter wire wrap without touching setup 1.

I do not know the current status of setup 2, but you could consider running some calibration runs with normal isotan in a tight wrap just like you have in setup 1.

Or maybe a somewhat less meticulous test can be performed in a quick and dirty third setup with just the components essential to find out how the temperature would vary with wrap.
Quote
 
 
0 #1 Ged 2012-11-15 17:17
I think you should try heating from the Celani wire first. I remember from Celani's paper, that using the Celani "active" wire doubled the excess heat effect over heating indirectly with an inactive wire. So, if we are replicating Celani's LENR effect here, switching over to the active wire for direct heating would be the fastest way to verify, by watching for a jump in excess heat production.

That's my thought. This is very interesting, quite the twist.
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