FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Some Comparison Results - Updated

Written by Ryan Hunt on .

We are trying to figure out if this previously damaged wire is giving us any real excess heat, or if we are seeing an artifact of pressure changes and other characteristics of the cell.  Towards that end, we have executed a few tests over the last few days to try to compare the effect of cell temperature, pressure, and cooling and reheating.

First, on the evening of the 26th of Nov, we turned on the Celani wire to 48 W to give us an immediate reference.  It settled to about 0.5 W

Then we turned on the NiCr wire to heat the cell while keeping the voltage and current through the Celani wire constant.  This resulted in a P_xs that started at 0.75 to 1 avg and ended up bouncing around 0.25 in the morning.  A slight increase over the roughly 0 to .25 we saw with just the Celani wire heating immediately previously.

Then we briefly turned off the Celani Wire and let it cool before turning it back on in order to try to estimate the working temperature of the wire at his higher cell temperature range.  Results on that to be shared later.  I am a little bit behind in my analysis.

Next, we turned off the Celani wire and adjusted the power on the NiCr wire to be exactly the same as what the Celani wire had been.  This passive heating of the active wire actually seemed to show a slightly higher P_xs in the 1 watt range.  This could be attributable to the variation in surface temperatures.

Finally, we decided to re-pressurize and replicate the exact conditions that we saw a week ago when we started the long rise run.  On the evening of Nov 20th, that run started out at 4.5 Bar and 43 Watts and a curiously low P_xs of -2.5 or so.  To explore the range just a bit more we pressurized the cell up to 6 bar when at temp and we are allowing it to slowly decline because of the known leak.

After we powered it up to 43 watts we saw the P_xs approach -6 at 6 bar.  As the pressure dropped, we saw the P_xs rise.  Once it hits 4.5 Bar we will see if it matches the previous data.  If it does, then we have pretty good evidence that we are not generating any more energy int he wire than we were a week ago, which is presumably nothing.

We do think we have a handle on why we are reading values that are so far below our calibration curve.

When we look at the T_mica and T_GlassOut vs Pressure at 110W and just today at 48 W, we see some very explanatory curves.

This first graph we have published before.  We took points from each calibration right around 110 watts and plotted the temperature achieved versus the cell pressure at that time.  It is interesting to note that the H/Ar line crosses the He line at about 4.7 Bar.  This lead us to do the next 3 graphs.

First we asked what it would look like at points near 48 Watts.  Same basic shape, but the line for H/Ar crosses the He line at 3.7 Bar. That is close to the range of pressure where our P_xs crossed zero last week. 

But we are measuring based on the T_GlassOut, so we plotted those, too.  First at 110W range and next at the 48W range.

In the 48 W range that we were at during the slow rise last week, we see some enlightening relationships.  First, we see the cross between the Helium line and the H/Ar line crosses, again, at 3.7 bar - where our long rise last week reached zero.  That makes sense because we are using data from the Helium run with this wire as our calibration data.  If we imagine our pressure starting out at 4.5 bar for this run and moving slowly to the left on the red line, we can see that we would measure a temperature below the blue line, then approach unity, and keep rising up to a max and then hold steady above the blue line from that point on.  That is almost exactly what we saw during our run.  (Time to go plot that T_glass_out vs Pressure over this graph, I guess, since most of it was at 43 or 48 watts.)

What this makes me wonder is how this curve looks up to 8 Bar where Celani was working?  And then, how would that pressure related curve reflect in the Stephen Boltzman calculations that Celani was using?  Should be easy enough to estimate.  Anybody up for it?  

Update:

Malachi and I made a couple more graphs, but I though they fit best in this blog article rather than in an entirely new one. 

Here is one of the graphs we made above with the data from various runs we have done in the last week.  I gotta admit, it was not as easy to understand as I thought it would be.  There are a number of variables, still, since these were just sections of data that we happened to have near these conditions.  I think this calls for a controlled run aimed at comparing to these curves.  

Notice, though, the blue squiggly line from the long, slow rise we saw last week ends with a couple jumps up?  That is quite intriguing.  I didn't expect that.  All the data was in the 75%H2/24%Ar blend.

And this is the above graph normalized somewhat by dividing the T_GlassOut rise by the input power.  This does favor the lower power levels somewhat, but does help make things more comparable, overall.

If you can't really come to any insights fro the above 2 graphs, I don't blame you.  If you can, please share.

Below is a graph Celani made of one of his successful runs including a loading phase and a running phase.

 

Can the phenomenon we are seeing with the pressure effect explain some of the excess power in the above graph?  If so, how much?  It is interesting that his leak is roughly the same rate as ours, too.  

Wouldn't this whole thing be easier with a calorimeter?

Add comment


Security code
Refresh

Comments   

 
0 #39 Eric Walker 2012-12-01 06:25
Thanks for the interesting graphs.

From Celani's chart above, two things stand out:

1. His experiment is operating in somewhat of a different pressure regime -- between 3 and 9 bar. I think the highest I've seen in the present experiment is 4.5 bar.

2. In Celani's chart, I see that the pressure is going down, and there is a similar inverse relationship between pressure and P_xs. You can also see a response in P_xs to pressure transients.

Other points:

3. Like Dave says, there are a large number of variables. They are greatly complicating the analysis. I wonder if there is a way to systematically control for each variable.

4. In the charts above, the real runs are a lot choppier than the calibration runs. I have looked at the calibration data, and it is also choppy -- am I mistaken, or are unfiltered live data being juxtaposed in the charts with smoothed calibration data?
Quote
 
 
0 #38 Robert Greenyer 2012-12-01 05:35
@Ecco

We will look to arrange SEM analysis of the first US wires as was always part of our plan
Quote
 
 
0 #37 Robert Greenyer 2012-12-01 05:34
Hi Ecco,

Mathieu will clarify this in a post soon, but the calibrations in the EU cell are currently being done with the same constantan source wire of same diameter but instead of being treated by the Celani process it is merely oxidised.

This is going to be a one of a series of meticulous optimised calibrations that kill variables and learn from the US experience to give us better reference for the active wire runs.
Quote
 
 
0 #36 Ecco 2012-11-30 21:18
@Robert Greenyer: off-topic here, but looking at the real-time data for the EU cell, are you sure that the high power (~100 watts directly applied) calibration procedure isn't "frying" the active wire? That's something that was also done for the US cell.

I think SEM pictures of the active wire currently installed in the US cell will be interesting on this regard.
Quote
 
 
0 #35 Ged 2012-11-30 20:27
@Ascoli,

Completely agreed.
Quote
 
 
+1 #34 Ascoli65 2012-11-30 19:28
@ Sanjeev and Ged,
Up to now, the HUG team performed an accurate calibration and, if they think it is worthwile, they are able to improve it, including the effect of gas pressure and composition. This will allow them to reduce the magnitude of their P_Xs, which more properly is the error in estimating the P_in. But, in my opinion, the first step is just to reproduce the data showed at ICCF17, by using the same calibration procedure.

In his slideshow Tyler wrote that “The Celani demo is considered to be one of the best, simplest, and most conclusive demos to date”. This is the phrase with which I agree more and I am confident that the this worthy MFMP initiative will come to something meaningful and really conclusive.

We should recognize that, letting his cell and his wire be tested and providing his open and loyal assistance, Celani is giving one of the best and commendable examples of application of scientific method to the CF field.
Quote
 
 
0 #33 Robert Greenyer 2012-11-30 19:13
@Ecco

We are feeding him findings and he is assisting greatly
Quote
 
 
0 #32 Ecco 2012-11-30 19:03
@Robert Greenyer: just wondering, is Dr. Celani too receiving feedback from these experiments which he might find helpful or is it still pretty much a one-sided learning process on your part?
Quote
 
 
0 #31 Robert Greenyer 2012-11-30 18:36
Even if we have found something that might appear to account for some of the affect, we are aware that our cell has important differences to Celanis and this wire could be toast also. Let's see what the last few experiments may show us before getting to excited - and remember, we could be wrong.

With the EU cell running calibrations now and the US getting closer to running more fresh wire there is going to be much more data to work with and to build on our experiments to date.

Celani has spent best part of 2 years on this line on the back of over 20 years of experience. We are coming to the end of week three. But with your help and a scaling of the number of experiments our learning will accelerate.

Isn't science FUN!
Quote
 
 
+1 #30 Ged 2012-11-30 17:00
@123star and Ascoli65,

You forget Celani did controls; you also both I think are overlooking the magnitude of Celani's results versus the ones here; an order of magnitude different. That cannot be explained simply by minor math or baseline differences, especially when that does not explain all of the results here either. Don't jump the gun now.
Quote
 
 
0 #29 123star 2012-11-30 16:52
@Ascoli65
Excellent!
The leak (both ours and Celani's) + the type of calibration used by Celani could explain everything!

@Sanjeev
We already did a "same gas - same pressure - different wire" test. Compare Cal1 and RunHe1. They gave different results. Why? Probably this is related to the wire emissivity (and hence the radiated power spectum). Or, in other words, our calorimeter is not a calorimeter :P
Quote
 
 
+1 #28 Sanjeev 2012-11-30 14:39
@Ascoli65

By your equation I meant the SB eqn which you used to plot the curve. I wanted to see the effect of compensating for pressure variable, but now I think its not very fruitful. At most you will see a bit less P_Xs, but nothing can be concluded.

I'm not sure if Celani's excess energy is also a mathematical artifact because, if it is so, all kinds of wires will show the same excess heat. I'm sure that if Celani repeats his experiment with a steel wire of same size, we won't see any P_Xs even with same simple calibration and calculations.
Quote
 
 
0 #27 Robert Greenyer 2012-11-30 09:17
@Ascoli65

By being thorough and with the help of you guys the truth will distill out.

Expect a data flood to come.
Quote
 
 
+1 #26 Ascoli65 2012-11-30 07:49
Hi, Sunjeev, the equations on the graphs I have shown are not mine. They are the equations that Celani used for deriving his “excess power” (see my previous comment quantumheat.org/.../.... The graphs show, unfortunately, that this “excess heat” his only a mathematical artifact, which depends solely from a too simplistic (and favourable) calibration procedure.

I fear that using that “self-calbratio n” you get some “excess heat” whatever wire and gas is used, provided that the cell has a gas leak and the test lasts for a few days. Sorry.
Quote
 
 
0 #25 Ged 2012-11-30 07:12
hmm, 6.5 bar of pressure, but only -2.9 W? Is this the pure hydrogen?
Quote
 
 
0 #24 David Roberson 2012-11-30 03:33
I obtained another curve fit for the latest power on and the time constant is very similar to the earlier measurements at 333 seconds. This is from power input to outside glass temperature. The curve fit is very reasonable.

It is also evident that the outer glass temperature is slowly rising after approximately 3 time constants. I also notice that the pressure is dropping during this same period.

The rate of outside glass temperature is smooth and in the vicinity of .0015 degrees C per second. I suspect the pressure change is somehow related.

There are plenty of variables to pin down. How are you going to separate the individual pieces? Perhaps another calibration sequence should be obtained using our new tools to verify.
Quote
 
 
0 #23 123star 2012-11-30 02:39
@Rats
Of course you need a calibration run first.
Only then you can evauate P_xs by comparing the data from the experimental run with the data from the calibration run.
Quote
 
 
0 #22 Rats 2012-11-30 02:15
@Sanjeev
There is no P_Xs for the Euro cell. How do we determine excess power?
Quote
 
 
0 #21 Sanjeev 2012-11-30 01:09
Looks like EU cell is now on. Data is coming.
Quote
 
 
0 #20 Sanjeev 2012-11-30 00:40
The rate is actually 1.7 if you see the last test, but 1 is on the conservative side, some excess will be seen.
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