FacebookTwitterDiggStumbleuponGoogle BookmarksRedditTechnoratiLinkedin

US V1.3 Cells -Curious Results and Next Steps

Written by Ryan Hunt on .

While the EU Cells were switched to a pure differential mode so they could run at higher pressure and more autonomously while Mathieu travels, the US cells are sticking to the V2.0 protocol so far.  We are on Step 8 and seeing some results this morning that I just have to run by the crowd and get opinions on.

 

Below is the data set from the beginning of step 8 of the V2.0 protocol, in which we loaded the active wire, again, and then heated directly with 25.0 W.  The noteworthy features of this graph are:

  •  The Resistance of the wire rises in a few different slopes.
  • The vacuum level of the cell drops in a pretty close mirror of the resistance of the wire
  • The T_Ext and the T_mica both rise way above where they settle when the vacuum level and the resistance have settled.  The T_mica peaks where the resistance rise becomes more gradual.
  • For about 8 hours the T_Ext temperature was elevated.
  • The indicated power out maxed out at 1.8W, but settled at ~1.0W 

 

 

So, for energy output, it was very much in line with the performance during the previous passive heated part of the test.  The story I am getting is that the hydrogen leaves the wire and the vacuum pump tries to keep pumping it out.  That is fascinating, itself.

The Potential explanations for the higher temperatures are:

  • It is either the presence of Hydrogen, or the few extra millibar, or both that is causing different thermal conductivity.
  • Or the Hydrogen leaving the wire monotomically is recombining and releasing that chemical energy.
  • Or possibly, the flux of hydrogen out of the wire along with the current and heat in the wire is causing a LENR reaction till the hydrogen runs out.

 

The trick, now, is to try to separate the physical and chemical effect from any LENR effect. One potential test I can imagine is to remove the Celani wire and run a control with the NiChrome and some hydrogen.  What would come after that?  After removing the active wire in cell A, we could later run in differential mode like the EU cells and try a range of pressures and power levels and see if we can find more obvious results.

What other tests would help us sort this out?

 

 

 

Add comment


Security code
Refresh

Comments   

 
0 #19 Ecco 2013-08-07 14:11
@MFMP: before you start the next live run try checking if you can fix the live data feed at the moment - it's currently silent.
Quote
 
 
0 #18 Ecco 2013-07-19 19:03
@MFMP: under such a vacuum, seeing how resistance is settling at a still relatively low value [actually it appears it's increasing more or less linearly at the moment, although at a rather slow rate], the active wire in US Cell A must be receiving a very limited amount of heat through small patches in contact with the mica support and through some heat radiation by the Ni-Cr wire.
Quote
 
 
0 #17 Ecco 2013-07-16 01:02
AH, yes, sorry. I was looking at Cell A and didn't realize B was powered too since I assumed that A was in loading phase and when that happens, B is usually left unpowered.
Quote
 
 
0 #16 Ron B 2013-07-16 00:51
@Ecco,
07/15/2013 12:00:00 07/15/2013 23:23:48

US 1.3B -- Pb(bar) shows .01 ... do you see something different?
Quote
 
 
0 #15 Ecco 2013-07-15 23:45
@Ron B: the cell has been filled with 2 bar of H2, the Vac (mbar) field isn't reporting meaningful values anymore.
Quote
 
 
0 #14 Ron B 2013-07-15 23:40
07/15/2013 12:00:00 07/15/2013 23:23:48
It seems that pulling the stronger vac on US1.3B is increasing the temps, is that expected? I guess I could see the mica temp going up, but the ext temps as well?
Quote
 
 
0 #13 Robert Greenyer 2013-07-14 12:13
@Edwin Pell

Well, part of the MiniProject is to put across some of our ideas and get you guys to critique it / suggest ways to optimise it.
Quote
 
 
0 #12 Edwin Pell 2013-07-12 21:32
Quoting Robert Greenyer:
@All

We are readying our latest designs for the sequel to the Steel and Glass, for a mini project to let you guys let rip. It will be very flexible, allowing for single and dual chambers, flow through, fluid calorimetry, arc/glow and RF triggering.

Bob

Excellent, I am excited. Do you have an order of magnitude feel for the cost of making one? Thanks.
Quote
 
 
0 #11 Ecco 2013-07-12 21:23
US Cell 1.3 A is showing interesting longer term data:


i.imgur.com/ILnJZNt.png


i.imgur.com/ZC5z6MU.png


i.imgur.com/nNvULTt.png

Mica temperature appears to be slowly increasing during steady state operation. Unfortunately increasing pressure while the cell is hot breaks/resets this behavior.

Overall there's an oscillatory pattern in cell temperatures. However ambient temperatures don't seem to show such a marked pattern:


i.imgur.com/Va5Ivfp.png

Since the pattern appears to follow diurnal variations overall, I wonder if the T_Amb sensor is reporting correct temperatures or if somehow changes in ambient temperatures get amplified, affecting the cell(s) more than they would normally.

Also, why did active wire resistance start decreasing during active operation under vacuum? Incidentally, that's also about when mica temperatures started increasing weirdly over time.

EDIT

This is ambient temperature data from the CTC cell (which is in the same lab):



It seems to correlate pretty well with temperature data for US V1.3 cells.
Quote
 
 
0 #10 Robert Greenyer 2013-07-11 16:48
@All

We are readying our latest designs for the sequel to the Steel and Glass, for a mini project to let you guys let rip. It will be very flexible, allowing for single and dual chambers, flow through, fluid calorimetry, arc/glow and RF triggering.

Bob
Quote
 
 
0 #9 David Jones 2013-07-11 09:34
I think the control cell should be pumped through the active cell. This way the control cell is effectively in the ‘back-water’ and should follow the pressure rise of the active cell. This setup should enable de-convolution of any heating effect from LENR and simple pressure rise.

Using, different lengths of active heating wire in the two cells (say ratio of 1:3) should enable other spurious excess energy results not linked to gas pressure rise to be de-convolved from the LENR effect.

@Ecco
The cells could also be coated with black ceramic paint.
Quote
 
 
0 #8 Ecco 2013-07-10 17:48
@David Jones: it still won't prevent all possible issues/artifact s that could arise by having a transparent glass tube.
Quote
 
 
0 #7 David Jones 2013-07-10 17:40
Perhaps a better experiment.

Control cell and active cell connected together and thus kept under the same high/low static pressure of hydrogen.

Control cell has one length of active wire while the active cell has at least twice the length of active wire or if possible significantly more. Pump the active cell via the control cell for the first test then pump the control via the active cell for the second test. Arrange the pumping lines so that this can be achieved on the fly.

Run both cells in passive heating mode and initially calibrate with helium gas. Load with hydrogen, now, both should give excess energy with the 'active' cell giving significantly more in both pumping arrangements provided pressure is equalised - if pressure is not then this will be evident in switching the pumping direction . If the excess energy is identical in both cells then...


Does this set up not get around all the current experimental problems?
Quote
 
 
+1 #6 Robert Greenyer 2013-07-10 12:09
@Ron

That is an excellent idea.
Quote
 
 
+1 #5 Ron B 2013-07-10 06:12
Ecco,
You know, if we could demonstrate that the hydrogen remains in the wire under high pressure inert gas, it might be possible to do side by side tests with inert gas in both cells.
Prepare the two cells, one with a hydrogen loaded wire and the other without
Quote
 
 
0 #4 Ecco 2013-07-09 19:26
@Ron B: that could be an interesting test to try out.

An option to prevent this from happening could be to not start the test I previously proposed as soon as the wire is loaded, but to apply power under dynamic vacuum for a while first. I believe most of the hydrogen gets outgassed during this phase, when active wire resistance increases (EDIT: corrected) the most.
Quote
 
 
0 #3 Ron B 2013-07-09 19:15
Ecco,
Do you think that the out-gassing would occur even if the inert gas pressure was very high?

If you did that we could check to see if the pressure drops as it did with hydrogen. I understand that the atoms of H are much smaller but still, if there's a drop in pressure with the inert gas then we can make some assumptions about hydrogen.
Quote
 
 
0 #2 Ecco 2013-07-09 19:11
By the way, I wonder if before removing the active wire from Cell A you could do as follows:

- Quickly reload Cell A in H2
- Once the wire has loaded, power the cell off
- Flush Cell A and Cell B from their current gases, then fill them with Helium at low pressure, while power is still off. Do this a couple times.
- Apply vacuum again, connect both cells' vacuum lines together (as in EU Cells)
- Apply power again

This test should clear at least in part issues about the difference in thermal conductivity of the residual gases.
However, Cell A will keep outgassing hydrogen under load and this might interfere with this test.
Quote
 
 
+1 #1 Ecco 2013-07-09 17:40
I find curious that active wire resistance is now decreasing again even though there is still a vacuum in the cell:


i.imgur.com/Kg39WMB.png

The changes in wire resistance apparently also caused changes in mica temperatures:


i.imgur.com/cp9rkpw.png

On the other hand, output power appears to be overall strongly correlated with the vacuum level:


i.imgur.com/IGACqFS.png

I think that unless vacuum can be precisely and tightly controlled, no conclusion can be drawn whether there actually is excess heat or not. Plus, since the cell has a transparent tube there still is the very possible artifact of wire emissivity changing with H2 loading, affecting IR absorption by the glass tube and thus temperature readings by all thermocouples.

Bonus chart: output/input power ratio for the dataset I used in the screenshots above:


i.imgur.com/hZbWlRp.png
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