FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

Teasing things apart

Geschrieben von Ryan Hunt am .

The last 5 days, or so, have been a very interesting ride to follow.  Last Tuesday evening we started a run aiming at running in the same current range as Celani and in a 75%H/25%Ar gas.  The excess power calculation started out indicating approximately -2 watts and the pressure was 4.5 Bar.

Over the next few days we saw a slow temperature rise on all the temperatures as the pressure continually dropped.  We have confirmed that the leak is from the pass through where the power wires enter the cell. 

We made a few adjustments to the operating conditions along the way.

On Wednesday we raised the power in to 48.2 W to more closely match calibration points so we could make a better comparison to the calibration data we had.  We continued at that power and watched as the P_xs rose to over 1.1 on the minute averages and peaking just over 2 momentarily.  

On Saturday, as we were working around the cell, we noticed the instrumentation top logging.  While troubleshooting it, we caused the power into the cell to drop to zero for a couple minutes.  This provided an opportunity for a test that we were hoping to run, anyway.

As the cell rewarmed, it came back up to a approximately the same level of P_xs as before.  I was actually hoping it would come back to -2 for a while and then climb back up again slower to +1 watt, indicating a base line comparable to where the cell started

After we got it restarted, the zero offset calibration for the voltage and current inputs was off slightly.  When we noticed this a day later, and fixed it, we also noticed that the P_xs had been remarkably flat for that period and then got much noisier afterwards and gradually declined to near zero+/- 0.5 watts.

On Sunday, acting on a tip from Celani that he has found a sweet spot of 1.85 amps, we upped the power in to get to that level of current.  What we saw was the P_xs rise for a couple hours and then a slow decrease back to almost zero.  Didn't seem too sweet at that spot, so we turned the current back down to exactly where it had been before that (1.75 amps) and we saw the P_xs rise again up to the 0.75 range.

Since we are at a low enough power level where it is still ambiguous whether we are seeing real excess heat, the next steps I will take with the experiment are aimed at figuring out if we are seeing a verifiable rise above a base line where little or no LENR is occurring.  There have been many, many good suggestions about things to try.  I really appreciate the intelligent insights being provided as we collectively work on this problem.  

Options we could do:

1.  Turn the power off, the cell cool, and see if it comes all the way back up to where it was immediately before.  If the cell does not warm up that far, then we have some indication it is (or has been) working above a real baseline.  If it comes right back, we know that it either is at baseline, or returns quickly if it is active.

2.  Switch to NiCr wire and adjust it to the same power level.  Presumably, the temperature of the Celani wire will be decreased to a significantly lower level and we will see the base line temperatures for the various sensors for that power level.  If the P_xs is the same, then can assume we have not been generating any significant power.  A small reduction or rise in P_xs will be ambiguous because we are sensing glass temperature at only one point and could be subject to variation based on location of the sensor relative to the hot wire.  Plus, the NiCr wire is slightly thicker and will run slightly cooler on its surface, which could allow less heat to escape as IR radiation through the quartz.  

3.  Turn on both wires, which would heat the cell and make the Celani wire run hotter.  We have heard that the LENR effect improves with temperature.  This might boost production up to a more measurable level.

4.  Reset the gas blend to a higher pressure, again, and compare the performance to the -2 watts P_xs originally measured when the pressure was at 4.5 bar.  We could load enough gas in it to achieve 7 bar, allow it to coast down to 4.5 and below, and test whether the higher pressure appears to increase or decrease the calculated P_xs.  This approach is my favorite, since it will test and compare against the baseline we just set over the last 5 days.

5.  We could charge the cell with pure Hydrogen and see if that makes a more measurable effect.

6.  We could attempt a second method of regeneration of the wire that Celani suggested:  heat to 400C in 3.5 Bar Hydrogen for an hour or two.

And I am sure there are other options.  I will attempt #1 immediately and see what I can learn.  I would like suggestions on what to try after that and in what order.  

I am behind but working on getting some IR video of the cell and surrounding area edited, and getting some gamma spectra published.  I wish there was more time in a day to be able to engage all the suggestions and comments better, and do more experimenting.  Thanks for all the help and keep it coming.

 

 

Add comment


Security code
Refresh

Comments   

 
0 #24 Robert Greenyer 2012-11-28 04:46
@ David

Thanks again for this analysis. Could you possibly come up with an alternative Pxs estimation equation that takes into account these two components and their relative offset. This could be called Pxo maybe and obviously would not be representative until after 369s.

@others

Good ideas with the airflow sensors - the more accurate the better, buying off shelf is often cheaper than spending 100s hours reinventing the wheel - with calibration questions and difficulties for repeatability. I like the idea of something that can very accurately measure both flow and temp - that is a meaningful pair.
Quote
 
 
0 #23 Tony McConnell 2012-11-27 22:25
@Ged, the wireless sensor does look interesting!

There's a handheld ultrasonic anemometer here:

www.tsi.com/.../new_UA30.pdf

Looks considerably more expensive, but supposedly will measure to a resolution of 0.001 m/sec.
Quote
 
 
0 #22 Ged 2012-11-27 21:57
@Ryan

Here's a really promising looking sensor, used for home ventilation www.monnit.com/.../airflow.php

Seems it has a little ribbon that hangs beneath it and measures air flow. Pretty nifty looking.
Quote
 
 
0 #21 Ecco 2012-11-27 21:55
Power to the cell has been turned off; a thing to check might be how much time it takes for P_xs to reach 0 (if it reaches it at all, and not necessarily because of excess heat).
Quote
 
 
0 #20 Tony McConnell 2012-11-27 21:38
Ultrasonic air flow sensors?
Quote
 
 
0 #19 Ryan Hunt 2012-11-27 20:16
Maybe what we need for air flow sensor is a tiny strain gauge on a flag in a swivel because near the sensor it is all light and variable breezes. I wonder if anybody make a breeze detector.
Quote
 
 
+1 #18 David Roberson 2012-11-27 19:49
The time constants that I have measured most likely depend upon the gas pressure and type. I will be happy to continue to pursue this technique and apply it to any data that you suggest, but it is important that the systems be similar.

I have found that there appears to be two time constants mainly working upon the input heating power. One is at around 67 seconds, while the other is at 369 seconds. The transient magnitudes are 11.2 degrees and 108 degrees in that order;

The temperature fit is excellent with time exhibiting an error of typically less than .5 degrees at the maximums. Most of the time the error is less than .25 degrees.

The data being fit occurs when 48.7 watts is applied from 0.

I hope to apply my process to additional data soon.
Quote
 
 
0 #17 Ecco 2012-11-27 16:46
Is it only my impression or did active wire performance improve after the temporary input power reduction? If this is the case, then it might be interesting to test if after several quick cool-off->burn- in cycles, p_xs improves further.

Of course, then one would have to ignore P_Out calculations when power is temporarily removed/decreas ed, as they would erroneously indicate, due to thermal inertia, excess power.

Thinking about it, this is similar to my previous "ringing signal" idea I suggested some time back.
Quote
 
 
0 #16 Ecco 2012-11-27 15:31
I see that my suggestion of trying a lower input power is being tested right now. However I meant at the active wire rather than the heater one. No worry though: it might also be interesting to check out what difference there will be by applying the same 36W to the other wire only too.

Perhaps, as the active wire will be hotter by receiving power directly, an even lower input power could be tested instead.

Update: BTW, with a lower amount of indirect power, the active wire seemed to be doing well.
Quote
 
 
0 #15 Ged 2012-11-27 06:12
@Robert,

I don't have personal experience with a sensor, unfortunately, but I found at least one good airflow sensor. Actually used to measure things like mouse respiration. Kinda price at ~$400 dollars though kentscientific.com/.../...

There's some other industrial units I found, but those are probably extreme overkill for the application.
Quote
 
 
0 #14 Robert Greenyer 2012-11-27 05:25
@David

Thanks - this is really helpful, so what you are saying, quite correctly, is that there are two components that are differently out of phase. The IR lag is much less, and the convective element greater. The significance of the IR component also will be affected by thermalisation capability and conductive averaging of the type of glass chosen. The power in metric is instantaneous.

I think a better Pxs would come from incorporating the lag you have Identified especially in this Quartz cell given its relative transparency to IR. Could you run some of the recent data with this lag and calculate the resulting Pxs so we can see what that looks like, maybe we won't see such extreme highs and lows?

Until this is done, the 5 min average might be considered a more accurate representation of what is going on.
Quote
 
 
0 #13 David Roberson 2012-11-27 04:02
I have been playing with the time constant that exists from when power is added to the heated wire until it shows up on the glass out. I get a value of approximately 6.38 minutes. The value seems to vary somewhat, but this is a pretty good estimate.

I think that I can see the IR energy initially heating the glass outer surface but have not verified that my numbers make sense. When you dropped the power to 0 then heated only the inactive wire this last time I detect a 7.5 degree heating from the direct IR.

I hope these numbers are correct and help to answer your questions.

Keep up the good work.
Quote
 
 
0 #12 Robert Greenyer 2012-11-27 03:51
I think we need to think about the lag between the momentary fluctuations in power in and the time it take for this to be seen on the t-glass out. As the momentary Pxs is a calculation based on things that are out of phase, spikes or dips in apparent Pxs could be a phase addition or subtraction artefact. It might also reveal a more detailed understanding of transient air flow around the cell which could then be better assessed by an air flow meter as suggested below.

Can someone calculate an estimation of the time it takes for this thermal inertia from available data so that we might formulate a better live estimation of Pxs.

Also - once calculated, can someone run a bit of analysis on past data either side of this calculated value (slide the relative t-glassout & power in relationship) and see what conclusions could be drawn.
Quote
 
 
0 #11 Robert Greenyer 2012-11-27 03:16
@Ged

I do like the idea of having an airflow sensor next to the cell and would be great to have that data live also. Do you have a suggestion for a suitable component?
Quote
 
 
0 #10 Ged 2012-11-27 01:14
@Ecco,

Very true. So far the signal to noise ratio doesn't seem too bad, and hopefully will tighten. First time I've seen a 1.49 W spike for the P_xs, and so rapidly from the onset of the temperature boost. It's sitting in that "teasing" range at the moment.

I'll reiterate that it would be a very useful to have an airflow sensor near the cell.
Quote
 
 
0 #9 Ecco 2012-11-27 01:08
Yes, but as noted in other runs, the hotter the glass tube gets, the more its temperature randomly varies over time, perhaps due to the increase of convective air movement around it. This is affecting the stability of calculated P_xs.

I don't think there's much that can be done about this except insulating the reactor (or ideally increasing excess power - easier said than done - so that this random variation becomes less significant) and/or - maybe (wild guessing here) - mounting it vertically.
Quote
 
 
0 #8 Ged 2012-11-27 00:57
And we return to spikes 0.8 W of P_xs already. So far so good. The excess at the very least is staying stable at higher temperatures.
Quote
 
 
0 #7 Ged 2012-11-27 00:33
Oh hoh, right after I post, now it's time for the dual wires. That is the quickest thing we can do of the primary options, and a great one to run over night. Well, time to grab the popcorn and see what the data shows us. Will we get a bump in P_xs now that the wire seems functional?

Hopefully it'll be unambiguous one way or the other, no teasing this time.
Quote
 
 
0 #6 Ecco 2012-11-27 00:30
Power has just been applied to the heater wire as well (It's been mentioned before, but independent control of input power for each wire would be interesting, by the way).
Quote
 
 
0 #5 Ged 2012-11-27 00:23
Ok, it's definitely climbing right back up. I see the cell was powered off previously and that is what I missed. Nice we're still climbing right back up; getting a good handle on these current conditions.
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