This plot was constructed from a data file that Mizuno transmitted to me as an email attachment. For convenience, I have color-coded the traces using the same colors as my recent plots. As noted on the two vertical scales, flow rate and cell current (Icell) are plotted using the right-hand scale. Power, voltage and temperature use the left-hand scale. The horizontal scale is 10 minutes/division and covers about 96 minutes. His data points are taken at 10 second intervals.

The excess heat signal is clearly evident from the fact that the Pout trace runs well above the Pin trace for the entire run. Note that the excess appears "instantly"...there is no noticeable "loading period" like other CF cells have exhibited.

As far as I can tell from the data file, there is no calibration factor applied to the Pout data...it is simply flow rate multiplied by observed delta-T. The only "fudge factor" I can see is a 1.1 that is applied to the current measurement and Mizuno explains that was derived by comparing the measurements made with his data logger with a Yokogawa power analyzer.

I have no idea why he changes the flow rate while the run is in progress. At least the reported Pout is about the same for each flow rate.

Note that the Pout trace is rather noisy at first and then smoothes considerably, especially at the last flow rate value. The absolute value of the delta-T is getting larger and larger as the flow rate is decreased so the noise appears to be a more-or-less constant magnitude fluctuation in the temperature readings.

Click here to download w90430-1.zip. It's about 200 kb and the unzipped file is in Excel 97 format. The graph you see above is an embedded object that I put in this spreadsheet. Otherwise it's just as Mizuno transmitted it to me. Most of it is self-explanatory but here are the temp probe assignments:

T1 = room temp

T2 = outlet water temp

T3 = cell temp

T4 = inlet water temp

T5 = room temp in a different place

There are some columns in there concerning vapor pressure which are for info only. That data does not go into Mizuno's output power calculation.