Setback with Stainless Cell
Part of the process of commissioning the stainless cell with the insulation jacket that we blogged about a few posts ago is to heat the cell and burn off anything stinky in the insulation jacket. Once it was up to temperature of 400 some and slowly rising, we left it to bake overnight. That was a bit of a mistake.
When we came back the next morning, we noticed the cell was cool and no current was flowing through it. It looked fine, but I figured it burned out a wire. When we pulled it out of the insulation, we could see it had been quite hot. And then it got stuck and wouldn't slip out of the aluminum sleeve even with a great deal of force. That meant we had to cut the insulation jacked open so we didn't damage the thermocouples. Once we opened it up we saw why it got stuck.
The aluminum sleeve had been melted! That means we had over 630 C there. Looking at the core of the cell, we learned more of the story.
The wire was burned off in at least 2 places. What happened was that the copper wire bowed out as it got hot and shorted against the heating wire. The shirt made a low resistance path and it drew way more power. As it heated up faster, now, it shorted out again! This time it got hot enough to melt the aluminum and, eventually, to burn out what was left of the heating wire. The Macor was discolored near the bottom where it had been the hottest, especially around the corresponding wire notches. Unfortunately, the epoxy in the pass throughs had melted out, too, so the cell was no longer tight.
All together, it means a lot of re-work.
Notice the vacuum hose to remove the dust in the air from the buffing wheel.
And below, is the re-furbished product with clean looking macor and new pass-throughs being glued in.
Time to make a new aluminum sheath.
Hopefully we can reuse the insulation shell since it is already baked out, but almost everything else will have to be redone before we can get this together and set into the air flow calorimeter for calibration. And then we'll keep a much better eye on it.
The good news is that the cell was showing about 8 C/Watt before I left it alone, so it should be nice and sensitive to excess heat generation.
Comments
We have been informed that sulphur in some steels can kill the effect, so lets hope that is not a problem for this cell. We are looking to specifically address this issue in further work.
We are likely to take this steel cell into calibration in the next few days and then into live runs. I think that the high temperatures for low input - say 20W in achieving 160+ degrees means that we may see far higher signal to noise in the configuration as is - this is good.
But fear not, yet other cells are being conceptualised that will learn from and extend this work. With the communities help and contributions we feel we are working towards a practical experiment. Every step makes the path ahead clearer.
Thank you for your kind words, It is certainly exciting to be involved as you say - we need everyones eyes on the data over the coming weeks because soon the EU cell and this steel cell may be running and this will start producing a LOT of data.
We will also need help spreading the word when the Kickstarter is launched. We just need a few 10,000s people to make a small donation and then we believe if it can be done, we can deliver a device that will enlighten the world in order to start this revolution.
Good idea, and if you could get the heating wire a bit closer to the active wire then its resistance variations would clearly show the temperature of the Celani wire along its entire length.
If you take care to have the heating wire set into the Macor a bit more to assure that if it does sag it won't connect with the Celani wire.
So exciting to be involved in this phase. A quick glance at the data just now looks very encouraging! Great work team Fire.
EDIT: I mean having the heater wire located where the green lines lay in this edited photo:
i.imgur.com/Shnjz.jpg
I have one question: I notice that there is one graph shown that shows the curve of temperature versus power expressed as a quadratic. The label for the quadratic shows 4 decimal digits for the coefficients. Are these the numbers you use for your curve generation? I have found that 6 digits works better for me.
Thanks again, and good luck with that stainless cell.
It's here:
docs.google.com/.../edit
That shows the power out calculated from T_Glassout to be low, too, on both cells. -13 on cell 1.0 and -2 on 1.1.
Here is the data for the calibration including times.
docs.google.com/.../edit
recently I have been using calibration data for the cells that apparently is different than yours. Is it possible for someone to give me the time stamps that point to the actual set of measurements that you are using for your calibration?
I am following twp experiments (cells 1.0 and 1.1) and would love to be able to verify the excess power that you seem to be obtaining. Thus far I detect no excess power with the calibration I am using. I am actually calculating less power output than is inputted by your supply when I monitor the outside glass temperature. For this reason I suspect my calibration is off.
Thanks for any assistance you may offer.
You can also purchase replaceable fusible links for heavy current loads.
All the suggestions here are good - current limiting, temp based control or power control. The newest software in our instrumentation package includes those features, but they aren't well tested yet.
This incident does point out that care must be given to safety in the design if these units will be shipped out.
RSS feed for comments to this post