Write up of Calibration Variation
Over the past few days we have come up with two slightly different baselines against which we estimate the heat flow out of the test cell. Both are based on the long series of calibration runs, but they are based on different temperature sense locations. There are many subtleties to it. I will try to explain as clearly as I can here.
There are two possibilities we have come up with so far that may explain the difference in achieved temperatures:
Compressed wrapping of the Celani Wire - When we installed the Celani Wire in the cell we were extra careful with it because we had heard that it was brittle and we knew that the coating was somewhat prone to flaking off. As we carefully wrapped it around the mica supports, we did not wrap it as tight as the Isotan wire, and as a result, we ended up about 1 wrap short at the far end of the cell. One possible result of this is that if we put 100 watts into the cell, that heat is being dissipated less than all of the tube, effectively concentrating the heat and making it appear slightly warmer than in the control runs.
Differences in thermal radiation emissivity of the Celani wire vs the Isotan control wire. The Celani wire has a rough, black surface, which has a very high emissivity constant. The Isotan wire has a smooth surface. That should make the Isotan wire itself run hotter than the Celani wire for a given amount of input power because it can't radiate off as much energy. That difference along with the relative transparency of the quartz tube at infrared wavelengths may have something to do with it. I know there is somebody else out there who can shed more light on this and how it would affect the cell temperature. Another complication is that the 0.020 inch diameter thermocouples inside the glass and outside the glass and the kapton tape affect the temperature read at that spot on the glass because they all absorb some infrared heat.
Other observations:
While the calibration based on T_GlassOut may be more independent of gas conditions in the cell, it is strongly and inversely affected by changes in the room temperature. As T_Ambient rises, the T_Rise gets smaller, giving the inverse relationship. It takes a little while for the cell to come to equilibrium and give a valid measurement of P_xs after that. I am hoping to see everything relatively steady for 45 minutes before I have much of confidence in the output.
We have not had a chance to analyze the data for using the difference between T_GlassIn and T_GlassOut as another measure to correlate against. It has the potential to be nicely physically based because it models the heat flow through the glass like a thermal shunt. We should get to this soon.
Fitting the Stefan-Boltzman calculations to this is another possibility. Anybody want to run that on the data and see if it can be made to correlate?
We'll keep working to thoroughly understand the dynamics of this apparatus and keep you posted.
Comments
if so afterward you will absolutely get nice knowledge.
if so afterward you will absolutely get nice knowledge.
In the alternative, you could fluid cool the glass envelope with temperature-reg ulated coolant using a jacket or container for the existing device. That would enable you to perform flow calorimetry on the coolant by measuring its delta-T and flow rate.
I realize that flow calorimetry raises the difficulty and cost. If you think that the effect you are looking for is so large that it will be very obvious, then you won't need it. If the effect is more subtle, maybe the complexity is worth it. You seem to be demonstrating very nicely the problems with using spot temperature measurements instead of true calorimetry.
Ok, maybe before I misread, are you saying that the Celani wire (RunHe2, right?) is black, so it should have a high emissivity, and should be colder (with the same power input), and yet the thermocouples give a higher temperature reading?
But like I said, I'm just a lay observer who only knows what he's read.
You are almost correct. This would help - but contamination of the vessel would occur. But, surprisingly some loss of hydrogen would also still occur - but at a very very reduced rate and probably so small as to be Irrelevant.
Imagine the Celani wire is operating in a vacuum. With a thermocouple on a quartz glass outer tube that are both totally transparent at all wavelengths i.e. can only absorb/measure conducted heat. Then no temperature rise will be recorded at that thermocouple (assuming no electromagnetic absorption in the surrounding air). But if some mica is in contact/close with/to the wire and you also record the temperature of the mica you will notice the temperature rise in the mica.
@Ron B: as much as I would like the excess heat to clearly trigger, I think this might have something to do with a sudden change in ambient/room temperature conditions occurred a few minutes ago.
This all costs money and time though…
The mica reading will include conduction, convection and radiated heat from the wire. The glassout will record far less radiated heat as you have shown (much to my surprise at these low temperatures for quartz glass!). Thus if a large proportion of heat loss is via radiation, glassout will not record it - whereas the mica temp will. Mica is a very good infrared emitter (absorber) – radiant fires are made from mica.
This is why I suggested monitoring the mica temp, as, it might more readily show up an effect – even though it is subject to more noise and calibration spread.
RSS feed for comments to this post