Dave-- When I first saw the ripples in the models predictions, I thought that it would have to be explained--that the instrumentation signal may have some unknown input noise. This needs to be resolved as you suggest.
Bob Cook ----- Original Message ----- From: David Roberson To: [email protected] Sent: Friday, January 30, 2015 5:50 PM Subject: Re: [Vo]:Alternate Calculation and Calibration Method for Mizuno Report I agree that the pump power is not going to screw up the calculations as long as it is constant. It looks like just any other constant signal that will be subtracted off by your technique. The same protection is given to the errors in ambient measurements such as an offset. Once you hold the true ambient constant you will have to deal with the leakage of the signals that occur before you make the last measurement. Take a look at my calculation of how much is going to escape through the thermal resistance. I can clearly see the tilt downwards with my model. We could compensate for the signal droop if you wish. We know the time constant and can effectively boost signals that have been drained as a result of its operation. That will likely double the number of joules you measure from the pulse train. First, lets get the ambient variation under control and then we can clean up the loose ends. It looks like the thermometer is not the source of the ripples. Maybe I am seeing the HVAC cycles on and off. The period is in the ballpark of 1000 seconds. The odd thing about it is that it is appearing upon the coolant measurement itself as well as the ambient. The large coolant thermal capacitance should filter them from that waveform. We might be seeing an issue with noise getting into the electronics instead of the calorimeter. Dave -----Original Message----- From: Jed Rothwell <[email protected]> To: vortex-l <[email protected]> Sent: Fri, Jan 30, 2015 7:52 pm Subject: Re: [Vo]:Alternate Calculation and Calibration Method for Mizuno Report David Roberson <[email protected]> wrote: Using that delta we would have 2 watts of excess heat leaking into the system. From the pump? That is plausible. ~1.5 W from the mechanical work of the impeller, and ~0.5 W from motor heat conduction. However, I hope you now agree that this will not affect the adiabatic calorimetry, given that the temperature was so stable for 28 hours. I was astounded to see it was so stable! Any heat pulse or anomalous heat will be on top of this baseline. Perhaps the thermometer output varies in accurate steps of .01 degree increments? I doubt that. It is coming from a 1980s vintage HP A/D interface. I do not think that was limited to 2-digit accuracy. Watch out for your spreadsheet settings. I think I uploaded it set to round to the nearest 0.01. Change the setting and you will see more digits appear. I do not know how many are significant. It does not take much to throw off the air temperature measurements. Move a fan or a space heater a little and bang, you have a half-degree change. It is fortunate the reactor is under such heavy insulation. - Jed
