*GlowStick* [UPDATE#11 - Images from GS3 run and re-heat data]

Down in sunny Santa Cruz, California, Alan Goldwater has up-specced his *GlowStick*. It has a thermocouple in the core and one on the outside. It also has a pressure sensor.

It is coupled to HUGNetLab for data acquisition and a combination of PID, SSR and Variac for power and control.

Alan says "For the pressure sensor, I built a dedicated 5 v regulated analog supply. Noise on the output is below the threshold of my scope, <50 uv. I lowered the input impedance of the HUG board to 10k, which reduced the sensor noise to ± 0.2 psi. I got the a/d calibration to track within that limit up to 100 psi, should be good enough."

Alan adds "I have the reactor mounted and ready for initial temp/pressure testing. It's loaded with 0.1 ml of water. I expect the thermocouple wire plastic insulation will melt."

Sat 28th 2015, Alan received some Carbonyl Nickel from Bob Higgins to go with the LiAlH4 that was kindly donated by Robert Ellefson earlier in the week (thanks Robert!). This can be run first if the Parkhomov powder is still in transit.

The first calibration/pressure test at 720C

Apparatus construction details with locations of the sensors

This image shows the 30 second averages for the entire run. As BobH predicted, after the initial peak, most of the water seems to have condensed in the cool ends. No evidence of leakage is seen - positive pressure was maintained for the entire run.

This image shows the thermal delay from a power step to be 12 seconds. Little or no time offset between internal and external core temperature rise is seen.

In this image you can see the power, control and data acquisition systems.

This is a stocky *GlowStick*, it ha a pressure gauge at one end and a thermocouple inserted at the other end.

There is another thermocouple on the outside in the centre of the reaction zone.

The first calibration/pressure test at 720C

Apparatus construction details with locations of the sensors

This image shows the 30 second averages for the entire run. As BobH predicted, after the initial peak, most of the water seems to have condensed in the cool ends. No evidence of leakage is seen - positive pressure was maintained for the entire run.

This image shows the thermal delay from a power step to be 12 seconds. Little or no time offset between internal and external core temperature rise is seen.

In this image you can see the power, control and data acquisition systems.

This is a stocky *GlowStick*, it ha a pressure gauge at one end and a thermocouple inserted at the other end.

There is another thermocouple on the outside in the centre of the reaction zone.

Meanwhile, up in Minnesota, Ryan has been preparing the automated power controller for the *GlowStick*

He notes:

"I just got the controller started on a 12 hour test. Just with an empty tube and scrap of coil. It does data log. I still need to get the PCE830 to data log"

This is the controller we have there

This is the software that can be used to monitor it

The plan is to start with the Kanthal A1 heated mini-*GlowStick* and see if that can be slowly heated up to 1200C and then see if the power needs to be reduced to maintain the temperature as in the report of the recent Parkhomov experiment.

The real test will be with the pair of longer *GlowStick*s, these are heated by Inconel heating wires and will be of the same starting resistance. One tube will have Nickel powder only. The other will have Parkhomov powder (or initially Vale 255 if the powder has not arrived in time) and LiAlH4.

They will be wired in series so that the current is the same. The voltage drop across each cells will be monitored as it may vary slightly with temperature. The control temperature will be taken from the 'active' cell - Otherwise they will be equivalent.

In this way, if the 'active' cell gets hotter than it should, it will dial back the input power to both and there should be an immediate visual difference between the two cells as the control cell will not have the 'additional' heat from the core. There will be the additional evidence of the temperature difference between the cells.

These things will be the same

- current
- frequency
- structure of reactors
- Nickel powder
- all times of changes to full voltage applied across the two reactors
- all ambient conditions (within reason, they'll be a little distance apart)

This will not be the same

- one will contain LiAlH4
- slight variation in power between cells as the resistance will change (will be logged)

UPDATE#1 - Prep tests

If our data is to be believed, we feel it is important to share our calibrations, even when a team member isn't quite hooked into doing things fully live. That way, when you do see live data, then everyone knows how a null system behaved.

Alan Goldwater did a pressure / heating test yesterday with a little bit of water in the cell to see how things went and you can see the results in the attached images, there are descriptions attached.

Meanwhile Ryan took a dummy coil on a 12 hour heat-up to test using the old controller we bought. With 40V and 3 Ohms, it maxed out at 700ºC which means we may have to run through a 1:2 transformer so that we can get to the right temperatures. This may also mean we have to run control and active cells separately after the mini *GlowStick* run... if we see something interesting however, we can invest further in the experiment we want to do.

UPDATE#2 - Bill Of Materials

Basic BOM for the "Glow Stick" reactors built during our February experiments, and currently under test at HUG in Minnesota. Alan's GS2 design uses a larger (0.375 x 0.125") core tube.

BOM for Glow Stick reactor design  30 January 2015   Prices in US$.

McMaster-Carr - quantities to taste:
Core tube   12"   0.250" x 0.125"          8746K15      15.20
Core plug    12"    0.125"    ±0.004        87065K42    20.94
Outer tube  12"    0.500" x 0.375"        8746K21      36.27

Swagelok fittings:
Al ferrules   A-400-SET          1.76
SS cap         SS-400-C             6.90

Kanthal A1 or Inconel wire 18 gauge -  on hand

Misc:
Structural supports
Alumina felt for ends of outer tube
Thermocouples
Copper split nuts for heater wire
Diamond burr for grinding core plug to fit (lathe toolpost grinder needed)

Dimensions:
The core tube will be 12" long with 8" or less active region
The core plugs will be 2" or more long each end
The outer tube will be 10" long to allow clearance for the core fittings

UPDATE#3 - Preparing for first live run in California

1st April 2015

Alan did a series of settled power steps with the PID controller bypassed. Then he entered corrections for the increase of resistance with temperature, and used 5-minute average power from the Cal1 voltage data. The result seems good enough show the measurement technique will work, so he'll take the cell down for fuel loading tomorrow morning.

This morning he opened the cell by removing the 1.8" inside thermocouple and found it survived, though the tube cracked mid-way from being a tight fit. The was then put on vacuum at 200 C, initially 1 torr and dropping as the water vapour is extracted.

This afternoon he will make up the fuel mixture. As Dr. Parkhomov's Nickel has not yet arrived, he will be using the Hunter Chemical AH50 carbonyl nickel that Bob Higgins sent him along with some LiAlH4.

He says "I need to devise a system to fill the cell without removing it from the test fixture, maybe a syringe with a 1/8" tube attached. I hope to have all that done by end of day.

If all goes well, I'll start the live run around 6 AM tomorrow. Interesting things should start happening by Noon and excess heat by 6 PM [we wish!]. Not sure what to do if it's stable and still running at Midnight. Can I safely get some sleep with a 'nuclear' reactor running in my garage?"

Fuelling

Alan describes how he is fuelling the reactor

"Here's what I've come up with for fuelling. The larger brass tube is 0.125 OD, 0.094 ID. The ram rod is 0.093 OD. I'll weigh and mix the fuel in a small cup, then fill the tube, all in a dry bag. The loaded tube is then placed in the reactor slowly, allowing displaced air to escape without carrying the fuel with it. Then the ramrod inserted about 3 inches (length of the filler rod), and by withdrawing the tube while holding the ramrod in place, the fuel is ejected along the length of the cell bore."

UPDATE#4 - Fuelling done, planning for 6am California start

Alan says, "I got the fuel in without spilling any. The glove bag didn't work (much too clumsy) so I just used a mask and lab coat, no problem. I did have to apply some vibration to the brass tube, because the powder packed pretty tight initially. Once it loosened up a bit, the ramrod worked OK."

UPDATE#5 - All raw data made public

3 April 2015

All data for *GlowStick* GS2 made public... Thank Alan!

http://bit.ly/1IbuTsQ

UPDATE#6 - teardown of cell

Skip and Alan disassembled the GS2 reactor and extracted the fuel by scoring the alumina tube lengthwise and splitting it with a small chisel. We found the fuel residue to be a typical sintered rod, all at one end of the reaction chamber. We tested the tube section holding the fuel under the gamma spectrometer, and found nothing above background noise.

The fuel was separated into three parts of about 0.18 g each and stored in vials for third party analysis. We kept some small chips for a look under a microscope.

Earlier today Alan finished testing the rebuilt control/DAQ system and also completed grinding a new reactor tube and wound a fresh coil, identical to the GS2 heater. Skip fabricated a new core thermocouple and the Al ferrules arrived from BobH. So we're ready to assemble and calibrate on Monday, and could start a run with fuel Wednesday.

UPDATE#7 - Preparing the *GlowStick* GS3

Even spacing for even heating...

Alan and Skip are building the best *GlowStick* yet - attention to detail is key and, given the burn outs and uneven heating of previous experiments by several parties including the previous Santa Cruz GS2, they are taking no chances.

This is what Alan had to say

"Yesterday during burn-in I noticed that the heat output of the fresh coil was not well balanced. The thermal camera image shows this, though not as clearly as could be seen from the visible glow. So this morning I built a wire guiding jig, set the lathe for 22 threads per inch and made a fresh coil. The result is very evenly spaced, with about .003 wire separation, so no stretching is needed to prevent shorted turns. The last image shows the nicely uniform and symmetrical glow at 550 watts.

Later today we'll assemble the cell and cement the coil and covers in place. After drying overnight we'll do a final cement cure at 200 C, then a full calibration of both sides of the split cell against a fresh type K TC inside the core."

The result is a thing of beauty.

This design brings with it the previous GS generations ability to go to high temperatures and pressures, but will be calibrated on both sides internal TC to external TC. Like the last *GlowStick* designs, this will mean we will have a very good idea of the internal temperature, when we are measuring a lower external temperature that is well within the range of a K-type. The big change here though is that the reference TC is outside the external Alumina sheath, and so it will not be at risk of failing in the same way that the TC did on the previous Santa Cruz reactor re-heat.

Additionally, the design allows for an in-line control, that is to say, one side will have fuel and the other will be filled with alumina - same pressure, same materials, same wire, same current etc.

---

UPDATE#8 - Explanation of the heater coil design features

The calibration to 800ºC passivates the wire also - that is to say that the aluminium in the Kanthal A1 comes to the wire surface and is oxidised to Al2O3 forming a highly insulating layer that also protects the coil from further degradation in air.

In addition, we have established that the coils can elongate with temperature - this is dealt with in two ways, the centre connector is held in place with high temperature refractory cement - and the rest is free to move outwards from the centre. Also, there is extra space between the reactor core and the ID of the ceramic sheath, this means that the coil has room to expand out a little radially rather than be forced to elongate.

It does however elongate a little still, something you can see clearly in the Padua recording - but this has the effect of separating the coils a little more as the temperature rises - just what is needed to decrease the likelihood of shorting.

One side effect of the elongation is that the local power density is not what it should be as power rises, and so that is why a calibration is done, ideally with a pre-heat to bake in permanent deformation and preliminary passivation.

---

UPDATE#9 - First calibration done on *GlowStick* GS3 done, second underway

[]=Project Dog Bone=[]

In Alan Goldwater's own words...

"First calibration data for the *GlowStick* GS3 is graphed in the attached image. We reached the target temperature of 1250 C inside the core with 800 C outside the heater covers. The divergence of the active vs. null temperature is probably due to small differences in contact between the coil and the covers.

We'll install an extra thermocouple for the active side data, slightly closer to the center, while leaving the existing one controlling the PID. That way, the calibration of set point against core temperature will be valid, and the comparison of the active vs null temperatures in the streaming data can be used to look for excess heat in real time."

---

UPDATE#10 - Second calibration done on *GlowStick* GS3 done... ready to run...

In Alan Goldwater's own words.....

"Here's the temperature profile for today's GS3 test, and the calibration data used to generate it.

We'll be starting soon, but the target temperatures and dwell times can be changed on the fly."

We might need to adjust the timings based on what pressures we see.

---

UPDATE#11 - Images from GS3 run and re-heat data

Here is the photo set from various points in the *GlowStick* GS3 run.

Data from the re-heat is published here

https://goo.gl/OPlJDq