Case Run 6...an Interesting Anomaly.....15MAY98
Before Run 5, which employed some G75-D catalyst received from Dr. Case, I modified the temperature sensors in the chamber so that two of them would be in the catalyst bed (now called Ttop and Tbtm(Tbtm)) and only one of them (Tgas) up in the gas space above the bed. During Run 5, the Tgas signal showed some erratic behavior that I initially interpreted as electrical noise. I have had electrical noise problems with these thermocouples in the past so it did not seem unusual. In my first report for Run 5, I omitted the Tgas trace.
For Run 6, I resolved to investigate the source of this noise. To my surprise, I discovered that it was not noise but rather a rapid oscillation in the actual temperature of the Tgas sensor. I confirmed this fact by monitoring the Tgas thermocouple with two independent electronic systems, one of which was a totally isolated battery-powered unit, thus immune to the kind of ground-loop noise problems that I suspected was causing the problem with the other electronics. Both systems showed the same curious oscillations.
Let's start with the entire record of Run 6 (please ignore the starting time and date…the system computer's clock is set wrong):
When Run 5 ended, I left the last charge of H2 gas in the chamber overnight. Run 6 was conducted entirely with that same charge of H2 gas in the chamber...the vacuum system was never energized. Hence, the pressure (Press) is plotted on a 0-100 psig scale.
The run starts with ~40 minutes of baseline and then 40 watts (Pin) was delivered to the heaters. At 1.5 hours, the heaters were reduced to 21 watts. Note that the pressure (Press) rises from about 35 psi at the beginning of the run to about 48 psi after 1.5 hours, very close to the last pressure in Run 5 indicating little or no leakage from the chamber overnight.
At about 1.7 hours, the oscillation in Tgas begins spontaneously. As you can see, it looks like a noisy signal on this time scale. Between hours 2 and 3, the oscillation becomes erratic. At about 2.4 hours there is a small gap in the Tgas trace caused by connecting the thermocouple to the independent electronics to confirm the signal behavior.
At hour 3, I reduced the pressure to 22 psi and was amazed to see Tgas rise to 140° C and the oscillations stop entirely! At hour 4.2 I began to explore the pressure relationship in detail starting at 70 psi and working down to zero in steps. As you can see, the oscillations only occurred when the pressure was 55, 48, & 43 psi. Above and below that pressure, the Tgas trace was relatively steady…and at a higher temperature that during the oscillations!
IMPORTANT NOTE: Throughout all these exciting temperature excursions, the Pin and Pout traces stay so close to each other that its hard to tell that they are two separate lines. In other words, there is no evidence of excess heat in this experiment.
Much of the erratic appearance of the Tgas trace is due to aliasing caused by the fact that the oscillation period is similar to the sampling interval (15 seconds) used by my data acquisition system. To examine the Tgas signal in detail, I employed our TEK THS720 digital oscilloscope. I connected the scope directly to the analog signal being fed from the thermocouple electronics to the data acquisition board in the system computer. This signal is scaled such that 1.00 volts is equal to 100° C. To see the oscillations better on the scope, I increased the vertical gain and then used vertical offset to bring the trace back onto the screen.
This trace was captured at hour 2.8 (see plot above) and shows the typical behavior of the Tgas signal during rapid oscillations. The vertical scale is 100mV/div, which corresponds to 10° C/div. The vertical offset is -10 div, which corresponds to -100° C. Thus we can read this trace directly in ° C. The center of the display corresponds to +100° C and the trace varies between about 127° C and 135° C. The horizontal scale is 10 seconds/div so the period of oscillation is about 15-20 seconds and not very steady.
This trace was captured at hour 3.5 and shows the spontaneous onset of the oscillations when the pressure was raised from 22 psi to 48 psi. The vertical scale is now 20° C/div (because the initial temperature was too high to display on the other scale) and the offset is such that 140° C is in the center of the screen.
This trace was captured at hour 6.0, when the oscillations look like occasional spikes on the plot above. In fact, they are occasional spikes. The horizontal scale is 20 seconds/div on this trace and you can see a solitary trapezoidal excursion in the Tgas signal that lasts about 1 minute. These excursions were about 6 minutes apart. This behavior appears to occur near the lower pressure limit for the oscillations. At higher pressures the oscillations are usually more rapid.
Now let's go back and take a look at Run 5, with the Tgas trace included on the plot:
After hour 2, when temperatures had stabilized and the chamber was at 50 psi, the oscillations in Tgas begin spontaneously. At hour 2.8, when the 3rd charge of H2 was put in the chamber, the oscillations changed from "rapid" to "occasional spike", yet the pressure was nominally the same for both charges. This indicates that other factors besides pressure can affect the oscillation patterns.
In fact, the oscillations essentially stopped when the chamber was filled with D2 (at hour 4.0)…and the gas temperature rose about 5° C. Later, at hour 6.2 when the chamber was again filled with H2, the oscillations resumed and the temperature fell.
However, it must be noted that the two sensors in the catalyst bed, where Dr. Case has his temperature probe, show essentially no sign of these temperature excursions.
This unusual Tgas behavior has only been observed in Run 5 and 6, both of which used the catalyst supplied by Dr. Case. However, that is not the only difference between these runs and the earlier runs. In both of these runs, only 20 grams of catalyst (all that was available) was placed in the chamber. This created a larger, taller gas space above the catalyst bed.
Perhaps the oscillations are due to an unstable convection pattern enabled by the increased gas space in these runs. If so, it is interesting that the instability is so sensitive to pressure and other factors (i.e. gas density).
Persons familiar with the Bernard instability, discussed previously on Vortex, are invited to comment on its possible application to this situation.