Using that delta we would have 2 watts of excess heat leaking into the system. My method of determining the .67 was not using a static measurement. I did trust your number of 41000 joules per degree C. as being accurate. Once that value is pegged, the time domain response follows an RC time constant very closely. So closely in fact that I obtain an error of less than .02 degrees C between what my model calculates and what is measured during the complete 6 hour period.
The error varies in an approximately equal positive and negative magnitude as time progresses. I see what appears to be quantization noise across the average value of temperature. Perhaps the thermometer output varies in accurate steps of .01 degree increments? My three pole IIR filter smooths out the more rapid variations to reveal a nice smooth sine like waveform having a relatively long period. In my calibration technique I adjusted the assumed value of the thermal resistance and viewed the resulting error magnitude of the coolant temperature. At each value of resistance I carefully adjusted the trim pump power term to get the least error. The best fit was found with the value of .67 ohms applied. I believe this result is reasonable but can adjust it if further data is made available. Dave -----Original Message----- From: Jed Rothwell <[email protected]> To: vortex-l <[email protected]> Sent: Fri, Jan 30, 2015 5:38 pm Subject: Re: [Vo]:Alternate Calculation and Calibration Method for Mizuno Report David Roberson <[email protected]> wrote: Do you have information about where the ambient temperature was during this long time period? For the entire 28 hours it is: Average 16.67°C, min 15.93°C, max 17.30°C The difference between the ambient and water settles to a much larger value than it did in the past. It is 1.39°C in this case. I think this changed after he put the tent over the experiment. It must have been in a pocket of warm air or something like that. He installed fans to make the air temperature more homogeneous. The temperature swings are much smaller than before, because of the tent. I have an extremely accurate measurement of the thermal resistance from the data you supplied which is .67 degrees C per Watt. That is what I got previously but I think it is changed. Or I guess I should say, I do not think the ambient temperature measurement is trustworthy to within a half-degree. After a lot of frustration, I decided to stop trying to derive the pump heat from on the basis of the difference between the water and air temperatures. I'm going to wait for additional calibration data and try and get it from that. - Jed
