Posts by Paradigmnoia

    I opened up my calorimeter to check for damage after a hour or so of open circuit operation of the heater. It looks like brand new inside the SS cylinder other than the fibreglass AC insulation is a bit faded near the cartridge. Re-crimped and tightened up the connectors anyways. Thinking that this is an ideal time to run some open resistor tests (no SS tube)...

    Such an error would be clumsy and thus unlikely but not impossible. That is why having a replicated experiment or at least a second set of eyes review all components in Mizuno's rig is important as he could have accidentally left off an important detail. That said, do we see anywhere that the active vs control run have used a different blower power circuit?

    I would expect that even if the current sense resistor in series was inserted in the active runs, the blower voltage should have been measured across the blower motor, not the blower + resistor, and the sense current measured across the resistor; so that no guess needs to be made on the blower effective resistance. Better than a current sense resistor would be to put a current sensor in series as exists in a common digital multimeter (but one that collects data for the digital data recorder) and not use a sense resistor as a crude current measurement device.

    It is ESSENTIAL that the control run and active run have as little changed as is possible between them. This includes any electrical component and ideally the reactor itself. I have always been uncomfortable that the calibration reactor is different than the control reactor for the initial published Mizuno R19 tests; and that there is essentially no control for the published R20 tests. Any change between control and active must be documented as fully as possible for conclusive proof of excess heat.

    Without fan voltage and current measurement , (without a MAF, etc) it is nearly impossible to determine if the air flow rate changes between experiments. The air flow measurement in the Mizuno calorimeter is one of the weakest points of that arrangement, IMO. Reduced airflow nets a higher delta T, all else being equal. It is the easiest way to fool the calorimeter based on my tests. The Albiston calorimeter fixes the weakest link with the MAF sensors.

    The current sense resistor in the Mizuno calorimeter is in series with the fan in order to make a voltage signal for the data logger. In the Saito calorimeter photo, there appears to be no blower fan sense resistor, instead there is an adjustable DC power supply. In that case it is possible to feed and measure the reactor and calibration units the desired blower voltage and current directly. To match the Mizuno calorimeter characteristics, Saito would have to set the fan voltage to about 9V, not the total 10.45 V that the Mizuno calorimeter gets because the sense resistor is in series. However, it appears that the calorimeter is not fully assembled in that photo: the RTDs are still in a bag beside the electronics, and there is a blower fan gasket on the RH side of the calorimeter.

    A current sense resistor in series with the blower performs some very useful functions. The voltage to the blower fan is proportional to air flow, and the current sense resistor voltage can show that the fan is in fact always on.

    Although not fully tested by myself, it appears so far that changing the reactor/calibration device resemblance does not affect the outlet-inlet air temperature dT, if everything else remains the same. The calorimeter captures (most of) the input power no matter what the configuration of the heat source is. A 200 W resistor 10 cm long or a 20 x 50 cm steel cylinder with 200 W input react the same at steady state. It is quite difficult to get a 5 C dT above calibration without adding real heat.


    Total air volume is in the 450 SLPM range for a blower voltage of 12.5V. I don't try and control blower power, but I assume it is fairly constant.

    Thanks. I was running some calculations and that is pretty close to what I came up with. (472.8 SLPM at 100% recovery, estimating the outlet air temperatures). That means that the overall airflow is about 2/3 of what Mizuno was running through his. (I realize it is not a replication of his calorimeter.)

    I woke up two nights ago wondering if somehow the Mizuno blower fan was being run at 10.45 V for calibration but only 8.5 to 9 V for the excess heat runs due to inserting the 3 ohm sense resistor only in excess runs. Of course that might be possible once, but is very unlikely for many different runs to keep doing that.

    I'm not sure I understand your question. I don't do anything to account for the MAFs small diameter other than I require four of them to allow sufficient air flow because of the back pressure they create. If I had a stronger blower I could probably get away with two MAFs. All of the calorimeter exit air passes through the MAFs. I hope that answers your question.

    Sorry, I was confused about what I was looking at.
    The end of the outlet tube is capped off and instead the MAFs all independently vent the outlet tube.

    Hot wire traverses suck to do, in my opinion, even if they work well when done properly. Doing it each time, for each temperature range is a total pain. I thought I would fix that with a vane anemometer the same diameter as the outlet. Turns out vane anemometers have high flow resistance at anything approaching their diameter. This is corrected by a ‘K factor’ (normally used for grilles) referencing a true velocity measurement. AKA back to square one, comparing to hot wire traverses. I think the K factor is constant for all velocities once worked out, (when installed in the same arrangement) but I have still to confirm that.

    However the outlet restriction is severe, so even the K factor can only be used for reference not permanent installation when attached to the outlet. There is so much restriction that the original air flow is highly unlikely to be flowing through vanes, no matter the normalization correction factor. The alternative is to mount the vane anemometer in the air flow at some cm scale distance from the outlet, which is essentially arbitrary, but can be adjusted to a few mm from where the blower fan current just starts to drop or outlet tube pressure rises (indicating a restriction) and set the K factor from there (if required).

    So, back to the MAFs, they are small diameter compared to the outlet tube, so how do you deal with that?

    Great work, and good idea repurposing the freezer. I wish I thought of that and now I really want a door on mine...

    Did you standardize to STP mostly because the mass air flow sensors report in STP?

    If you have more details on the mass airflow set up I would appreciate it, if you have the time.

    The cylindrical devices of Mizuno’s (50 cm long) are clearly longer than the cruciform reactors, so the Saito calorimeter box appears to have the same external dimensions as Mizuno’s calorimeter box. I estimate the cruciform reactor to be about 40 cm long, consistent with the description in the 2018 ICCF power point presentation (page 5).

    72 W and 50 W calibrations done. 50 W was fairly stable, much better than expected.
    Seems comparable to the figure 7 plot calibration, discussed above. One It cools down I’ll check the real data.

    Edit: For some unknown reason the input power turned off for an hour, right at the beginning of steady state for the 72 W run, and then started right back up again. So the 72 W delta T results are questionable, but at least 4.5 C. The 50 W run worked out fine, and has a solid 3.95 C delta T.

    Probably the heater cartridge connections are getting oxidized and will need inspection.

    If that delta T (blue, excess) on slide 7 was from my calorimeter, (which is not exactly the same and possibly slightly more efficient), it would represent about 210 W input. For the 50 W calibration trace, the delta T seems low but extrapolates OK from fan heat data. I can’t run as low as 50 W stably with the present heater arrangement, but serendipitously I can hold 30V which is 72 W.

    That is no surprise. I remember long discussion with a plant fitter in years gone by about speed variations of the industrial version of a turbo vacuum pump, he couldn't understand why it would speed up when there was a good vacuum and drop when there wasn't - he was sure there was something wrong with it. I tried explaining that if the turbo-fans aren't moving much air then they are doing little work, when they are pumping air there's more load. He could never see it.

    I also sort of found this out already when doing the voltage steps with the fan outlet unhindered a few months ago. At some higher voltage, the fan current plateau-ed as it seemed to reach the peak efficiency RPM of the fan.

    I reset the fan and series 3 ohm resistor total voltage to 10.45, which puts the blower fan at 9.05 V, and exactly 1.40 V for the resistor drop.

    The taped on (still) vane anemometer reports 6.30 m/s compared to 7.06 previously. That probably will be still 2 m/s high compared to the hot wire results if I took the vane off again. The hot wire anemometer can’t test the outlet properly with the vane blocking up the opening, which is now three wedges. I can stick another 65 mm ID tube onto the vane to fix that up. I also don’t think I should spin that vane for countless hours (unless it has magnetic bearings), but if I burn out the bearings working out why it seems to read high maybe that is covered by warranty...

    Also I variably blocked the air outlet to check the fan current draw under adverse conditions and it actually drops. Almost half the normal current consumption when the air outlet is blocked completely. Voltage remains the same.


    As noted, the spreadsheet has many more columns. This is just a sample made to fit the slide. The columns on the left show data as recorded, without any adjustment.

    Can you confirm if the Saito blower voltage is the voltage directly to the blower, or if there is a sense resistor in series included in the 10.57 V (and whatever resistance that resistor is, if any)? The 10.57 V and 0.48 A is consistent with a 3 ohm sense resistor in series, at least as far as my San Ace B97 blower measurements go. Mizuno had 10.42 V to the blower including the 3 ohm sense resistor, so the final blower voltage was lower, based on the equations we went through a couple months ago. I don’t think it affects the experiments much, as long as things are consistent, but it means that I should put my fan voltage back to where it was to do a more correct calorimeter replication.

    Is that a new version? I see it had the bubble foil installed photo I have been dreaming of.

    Anyways, on Slide 13, Calorimetry, I think the “Blower A” column is actually the voltage drop of the 3 ohm sense resistor, so current is actually 1/3 of the values in that column.

    Nothing to it. It is based on the calibration. That level of input should produce a 10 deg C temperature difference. It is 15 deg C instead. In other words, the blue line is actually measured; the green line is extrapolated from the input power measurement.

    If I might be so bold, why not show a calibration Delta T instead of a calculated one?