FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Write up of Calibration Variation

Written by Ryan Hunt on .

There is a video here to support this blog post.


Over the past few days we have come up with two slightly different baselines against which we estimate the heat flow out of the test cell.  Both are based on the long series of calibration runs, but they are based on different temperature sense locations.  There are many subtleties to it.  I will try to explain as clearly as I can here.
 
The first graph has the calibration data that matches with current operating conditions of 75%H2/25%Ar starting at 3.5 bar pressure.  We did 2 calibration runs at this setting in which we heated with the Isotan 44 control wire.  Then we averaged the data for all 3 runs to get a more meaningful curve and got an indication of variability between runs.  We started out with the
T_Rise =T_Mica - T_Ambient
 Then, later, we changed to
T_Rise= T_GlassOut - T_Ambient
 
Why did we do  this?  Because the variation in the T_GlassOut line seemed to be much less affected by gas composition and pressure.  This seemed like a really good idea given that we have a known but slow leak that would continually be changing the pressure, decreasing the thermal conductivity of the cell, and perhaps artificially raising the temperature of the Mica.  See blog article Towards a new calibration formula for more details.
 
This graph includes the measured average data points of T_Mica Rise and T_GlassOut Rise versus the total input Power.  The dashed lines are the best fit curves from zunzun.com along with the equations and parameters for both.  The best fit lines, indeed, fit very nicely.  During the experimental runs, then, these curves are used to calculate the amount of heat leaving the cell based on the temperature of the measured points on the mica and/or on the exterior of the glass.
 
 
The second Graph is a zoomed in view of  the above.  It is zoomed to operation conditions and highlights the current (as in an approximate) value of the operating conditions as of approx 2012-11-13 18:00.  From the points, if we draw a line straigh upwards, we see that the the T_GlassOut calibration line indicates approx 3 watts excess heat.  The T_Mica line indicates 8 or 9 watts.  Is it excess energy?  If so, why don't they agree closer?  What else might explain it?
 
 
The next two graphs depict Cal1 compared to RunHe1 and Cal3 compared to RunHe2.   We are trying to find a hint as to why we see the interesting behavior above.  In each of these graphs, T_Mica Rise and T_GlassOut Rise are compared between a calibration run with the bare ISOTAN44 and an identically conditioned run with the treated ISOTAN44 (Celani Wire).  We can see that the Celani wire runs hotter than the bare ISOTAN44 in all cases.  We also notice that the difference in temperature between the same sensor in each set of comparisons is larger for T_Mica Rise than for T_GlassOut Rise.  And, the difference is very close to what we are seeing right now.  Interesting...Why?
 
 

There are two possibilities we have come up with so far that may explain the difference in achieved temperatures:

Compressed wrapping of the Celani Wire - When we installed the Celani Wire in the cell we were extra careful with it because we had heard that it was brittle and we knew that the coating was somewhat prone to flaking off.  As we carefully wrapped it around the mica supports, we did not wrap it as tight as the Isotan wire, and as a result, we ended up about 1 wrap short at the far end of the cell.  One possible result of this is that if we put 100 watts into the cell, that heat is being dissipated less than all of the tube, effectively concentrating the heat and making it appear slightly warmer than in the control runs.  

Differences in thermal radiation emissivity of the Celani wire vs the Isotan control wire.  The Celani wire has a rough, black surface, which has a very high emissivity constant.  The Isotan wire has a smooth surface.  That should make the Isotan wire itself run hotter than the Celani wire for a given amount of input power because it can't radiate off as much energy.  That difference along with the relative transparency of the quartz tube at infrared wavelengths may have something to do with it.   I know there is somebody else out there who can shed more light on this and how it would affect the cell temperature.  Another complication is that the 0.020 inch diameter thermocouples inside the glass and outside the glass and the kapton tape affect the temperature read at that spot on the glass because they all absorb some infrared heat.

Other observations:

While the calibration based on T_GlassOut may be more independent of gas conditions in the cell, it is strongly and inversely affected by changes in the room temperature.   As T_Ambient rises, the T_Rise gets smaller, giving the inverse relationship. It takes a little while for the cell to come to equilibrium and give a valid measurement of P_xs after that. I am hoping to see everything relatively steady for 45 minutes before I have much of confidence in the output.

We have not had a chance to analyze the data for using the difference between T_GlassIn and T_GlassOut as another measure to correlate against.  It has the potential to be nicely physically based because it models the heat flow through the glass like a thermal shunt.  We should get to this soon.  

Fitting the Stefan-Boltzman calculations to this is another possibility.  Anybody want to run that on the data and see if it can be made to correlate?

We'll keep working to thoroughly understand the dynamics of this apparatus and keep you posted.

Add comment


Security code
Refresh

Comments   

 
0 #23 Parking Sensors 2018-06-05 06:36
Hi there Dear, are you genuinely visiting this website regularly,
if so afterward you will absolutely get nice knowledge.
Quote
 
 
0 #22 Parking Sensors 2018-06-05 06:36
Hi there Dear, are you genuinely visiting this website regularly,
if so afterward you will absolutely get nice knowledge.
Quote
 
 
0 #21 Ryan Hunt 2012-11-14 18:13
You guys are quick. New blog entry explains it.
Quote
 
 
0 #20 Al Potenza 2012-11-14 18:04
Your results suggest that you could improve the experiment by using true heat flow transducers to measure the actual heat passing through the glass envelope.

In the alternative, you could fluid cool the glass envelope with temperature-reg ulated coolant using a jacket or container for the existing device. That would enable you to perform flow calorimetry on the coolant by measuring its delta-T and flow rate.

I realize that flow calorimetry raises the difficulty and cost. If you think that the effect you are looking for is so large that it will be very obvious, then you won't need it. If the effect is more subtle, maybe the complexity is worth it. You seem to be demonstrating very nicely the problems with using spot temperature measurements instead of true calorimetry.
Quote
 
 
0 #19 123star 2012-11-14 17:52
@Ryan (I think)
Ok, maybe before I misread, are you saying that the Celani wire (RunHe2, right?) is black, so it should have a high emissivity, and should be colder (with the same power input), and yet the thermocouples give a higher temperature reading?
Quote
 
 
0 #18 Ecco 2012-11-14 17:34
As of now it appears that after vacuuming the cell the the MFMP team started injecting hydrogen again at a reduced pressure. No power applied yet.
Quote
 
 
0 #17 Lu 2012-11-14 17:27
It looks to me that pressure was steadily declining throughout the night. Something also happened about 1:30CST that changed the trend line for P_xs. A quick update on the status of the experiment would be appreciated.
Quote
 
 
0 #16 Ecco 2012-11-14 16:55
@GreenWin: input power has been switched on/off a couple times. It is currently off and as far as I can see, there's no thermal anomaly ongoing. I think they have done something with the power supply fan since before the latest switchoff the usual voltage dip (due to periodic fan activation) did not appear anymore. Maybe there there have been problems with it?
Quote
 
 
0 #15 GreenWin 2012-11-14 16:51
Am I correct to assume the experiment is not running - or not feeding the HUGnet database?
Quote
 
 
0 #14 Ron B 2012-11-14 15:22
Do you have a voltage sense line in the power supply setup? If so, what is it connected to If it's connected at the load, what is the length/gauge of the wires for both the load and the sense lines?
Quote
 
 
+2 #13 Achi 2012-11-14 15:18
Please ignore me if I'm wrong, since I'm just an observer making a comment on something I know very little about. From what I've understood from the many LENR theories I've read a smoother surface on the wire would lend itself to being less reactive in both the loading stages and the burning stages. From what I understand the rougher surface allows the H more surface area to penetrate the metal lattice, and also more surface area to be active in the reaction. That might explain the temp differences.
But like I said, I'm just a lay observer who only knows what he's read.
Quote
 
 
0 #12 David Jones 2012-11-14 15:11
@ Ecco

You are almost correct. This would help - but contamination of the vessel would occur. But, surprisingly some loss of hydrogen would also still occur - but at a very very reduced rate and probably so small as to be Irrelevant.
Quote
 
 
0 #11 David Jones 2012-11-14 15:06
Since almost no one is one is posting at the moment I’m going to labour the point.

Imagine the Celani wire is operating in a vacuum. With a thermocouple on a quartz glass outer tube that are both totally transparent at all wavelengths i.e. can only absorb/measure conducted heat. Then no temperature rise will be recorded at that thermocouple (assuming no electromagnetic absorption in the surrounding air). But if some mica is in contact/close with/to the wire and you also record the temperature of the mica you will notice the temperature rise in the mica.
Quote
 
 
0 #10 Ron B 2012-11-14 14:59
Oh, missed the ambient temp drop. it seems to drop off at about the same time.
Quote
 
 
0 #9 Ecco 2012-11-14 14:58
@David Jones: please correct me if I'm wrong, but shouldn't running the experiment at less than 1 bar of internal pressure (= slight vacuum) help with hydrogen leaking issues?

@Ron B: as much as I would like the excess heat to clearly trigger, I think this might have something to do with a sudden change in ambient/room temperature conditions occurred a few minutes ago.
Quote
 
 
0 #8 Ron B 2012-11-14 14:54
At 6:35PST it looks like it might be taking off. Excess power seems to be on an upwards trend. I didn't detect any other aspect of the measurements that's changing in relation to the upward trend in excess power!
Quote
 
 
0 #7 David Jones 2012-11-14 14:47
Another useful item is to have a gas reservoir (metal tube) of volume very much greater than your active chamber. This obviously reduces the variation of gas pressure over time in the event of a leak - as experimental setups like yours using hydrogen are almost impossible to get gas tight.
Quote
 
 
0 #6 David Jones 2012-11-14 14:27
A suggestion I have is to run a separate control tube besides the experimental tube. The control tube should be connected to the experimental tube via a gas line. Then a leak in one will affect the other. The control tube should have a normal wire in it – or possibly an unloaded Celani wire? This should enable better quantification of calibration and running errors.

This all costs money and time though…
Quote
 
 
0 #5 David Jones 2012-11-14 14:12
Possible answer for differences between glassout and mica ('possible') excess energy readings - they are observing different things.

The mica reading will include conduction, convection and radiated heat from the wire. The glassout will record far less radiated heat as you have shown (much to my surprise at these low temperatures for quartz glass!). Thus if a large proportion of heat loss is via radiation, glassout will not record it - whereas the mica temp will. Mica is a very good infrared emitter (absorber) – radiant fires are made from mica.

This is why I suggested monitoring the mica temp, as, it might more readily show up an effect – even though it is subject to more noise and calibration spread.
Quote
 
 
0 #4 123star 2012-11-14 11:50
I just read the bottom of the article and I find that a change of emissivity is a very good hypothesis that could also explain why the wire/mica/well is getting hotter when hydrogen loading occurs. In other words, what if hydrogen loading suppressed the emissivity of the wire (and no excess power)? In this case, what would happen to T_glassout, would it increase anyway?
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