264) Measuring electric input


Ludwik Kowalski (10/14/05)
Department of Mathematical Sciences
Montclair State University, Upper Montclair, NJ, 07043

The topic of possible systematic errors, in measuring electric energy supplied to a cold fusion cell, is often discussed on the CMNS list. The purpose of this item is to show interesting extracts from recent messages, to comment on them, and to ask additional questions. Note that item #258, and the appendix in unit #261, were also devoted to this topic. As usual, authors of messages from which I quote are not identified. I refer to them as researchers: R1, R2, R3, etc. I will be happy to substitute these Rs by real names, if requested. Likewise, I will remove a quote if the author asks for this. Please let me know. This issue, by the way, was raised by two young journalists on the CMNS list. They feel that quoting without permission is not consistent with journalistic ethics. I am not a journalist but I understand their concern. Therefore, from now on, items of that kind will not be posted before showing the draft to the list members, and waiting a day of two for possible requests.

Inserted on (10/16/05):
The above paragraph triggered an interesting discussion about privacy. It jad nothing to do with this unit; it had everything to do with ethics. Many old-timers feel that quoting them, even anonymously, is not appropriate. Before showing anything that was posted on the CMNS list to outsiders one must have a permission. I will start followong this rule. But this unit was composed befor the topic of privacy was discussed. That is why I decided to post it. Only one quote was removed because I now know that the author is against anonymous quoting without permission.

R1:
. . . The biggest concern in my view is when folks are simply measuring a voltage and current points and multiplying them, and assuming this represents accurate AC wattage. This is another source of potential systematic error in cold fusion experiments which I've seen time and again. Any good engineer who knows AC power issues will look at the results of that experiment, see the simplistic approach used, and discount the results.

My comment:
In a standard a.c. setup, i.e. when the waveforms are sinusoidal and when the phase factor, cos(phi) is known, the power is I*V*cos(phi), where I and V are effective values read from two instrument. The phase factor can be measured by using an oscilloscope. What R1 has probably in mind is a situation when v(t) and i(t) are not sinusoidal, or when cos(phi) is time dependent. In unit #261 I was addressing a situation in which v(t) remains constant and only i(t) is highly irregular. In that case the approach we used seems to be reliable.

R2 (referring to the approach described in item #261):
Note: R2 did not want to be quoted. The essence of the message was that the v(t) and i(t) waveforms contain frequencies excedding the limit of the instrument, 82 MHz. The sample rates must thus be higher than 100 Mhz. He worned us about possible errors associated with the assumption that v(t) was cinstant.

L.K.:
My reply was that we were cautious; voltages were sampled at the rate of 2000 per second and mean values were calculated (and recorded) five times per second. That was like using an oscilloscope. The near constancy of v(t), when i(t) is highly irregular, was possible because a capacitor of 0.01 F was used, as illustrated in Figure 1.The issue of the sampling rate was discussed in length in item #258. My conclusion was that the frequency of sampling can be as low as one wants, provided the number of random samples is sufficiently high.

R3:
I think all this talk about the difficulty of measuring power only applies to people who are working on a shoestring and do not have enough money to buy a top-of-the-line power meter. I do not think there is any difficulty measuring power when you use professional instruments costing thousands of dollars. With regard to glow discharge experiments, it should be noted that Mizuno measured power with both the Yokogawa and with an ordinary computer-based meter. The results agreed to within 1 percent, and the excess heat was considerably larger than this. The Yokogawa PZ4000 specifications are here: <http://www.yokogawa.com/tm/pdf/bu/pz4000/pz4000spec.pdf>

R4:
It is important when trying to measure power deliver to a varying load that the source impedance of the supply be much lower than the load. The power industry never has to worry about this as their line impedance is zero compared to what we have connected. If not we blow the breaker. We can be fooled by top of the line watt meter if there is not a correct match between the source and the load. If the match is incorrect the watt meter is measuring power that is dissipated in the source and in the load.

In the Mizuno Experiment the load impedance is low and changing rapidly. The impedance of the source must be significantly less than the changing load impedances. If it is not then getting correct power measurement is much more difficult because the impedance of the source must be added to the equations.

R4 (in an earlier message)
I've been interesting in CF since 1989 when fellow engineers at Eberline (Radiation Detection Equipment) in Santa Fe. NM tried to duplicate the P&F experiment without success. Advanced Energy is a manufacture of precision power supplies for the semiconductor industry. These supplies are ideal suited for testing the Mizuno plasma excess heat experiment. The supplies are fully programmable with voltage, current and power regulation. The supplies can deliver 20KW at voltage to 800V. They can also measure Joules delivered to the target at 1% accuracy, sample rates on the order of 1ms. The supply can be programed to run many different types of waveforms. It is also able to suppress arcing within the plasma. I'm setting up an experiment here in either Fort Collins or Boulder any one in the Denver area that is interested in helping is welcomed. I'm basing the experiment on the one done by Fauvarque and Clauzon. This is the web site for the Pinnacle power supply spec.

<http://www.advanced-energy.com/products/ProductDetail.asp?PID={4D7A014D-32FF-4653-AADC-0B9FEB0C9279}&cat=Cat&subcat=Power%20Systems>

R6:
We are making good progress. I agree, many NAE exist. However, I suggest a common mechanism operates in all of them. Based on your model, part of the common mechanism would be the creation of band states that cause wave formation. Other conditions would include an environment that can dissolve sufficient D and an environment that can properly host the resulting He wave. When all of the many required conditions are combined, very few chemical systems would appear to qualify. The BIG questions is, how does a person identify such a system in advance based on your model?

LK:
I am commenting on information posted by R4. Ability to suppress electric arc, for example, in a Mizuno-type cell described in item #261, might be very important. Uncontrollable arcing was the main cause of our difficulties. Let me speculate about one possible consequence of being able to eliminate arcing. The cell receives a measured amount of electric energy, E, and a measured amount of thermal energy, Q, is produces. The coefficient of productivity, COP, is defined as Q/E.

Many different processes occur in the cell, such as dissociation of water molecules, glow discharge, arcing, etc. We do not know which process creates NAE (nuclear active environment) mentioned in the message above. But we can imagine that

E=E1+E2+E3+ . . ., Q=Q1+Q2+Q3+ . . . and COP=COP1+COP2+COP3+ . . .

where 1, 2, 3, etc. refer to individual processes. Suppose that only process #1 (glow discharge) creates NAE whem E1=50 kJ and Q1=200kJ. This would make COP1=4. Also suppose that the second process, arcing, does not generate any excess heat; for that process E2=5000 kJ and Q2=5000 kJ. To shorten this speculation I will ignore other processes. Not able to measure E1, E2, Q1 and Q2 we measure E=E1+E2=50+5000=5050 kJ and Q=Q1+Q=200+5000=5200 kJ. Then we conclude that CPU=5200/5050=1.03. This shows how much would be gained, in this hypothetical situation, if arcing could be suppresses. It should be much easier to argue that the COP=4 is real than to argue that CPU=1.03 is real. Furthermore, assuming the test lasts 5 minutes, the power without the arc would be 167 W while the power with the arc would be 16,833 W. Designing a cell for 5 minutes of electrolysis with arcing is certainly much more difficult than designing a cell in which the arc is suppressed.

R1:
I've read the posts regarding measurement of "watts in" in complex AC/DC high-voltage systems. . . . R5 and I spent many months investigating this issue and found that it is somewhat of a black hole. We're of the opinion that such measurements cannot ever be accurate or dependable, regardless of the sampling rate. The problem is that in high-voltage, high-frequency systems, the power factor of the AC current is quite unpredictable. Voltage and current are NOT the only variables that lead to wattage. Power factor and phase are a major issue, and they are extraordinarily complex. > . .

Another problem is that conversion efficiency of the power supply. . . . Sometimes the power supply is 60% efficient, sometimes 50% efficient, and you cannot predict that efficiency at any moment. Our solution was to abandon the entire concept of "measuring" AC power as a product of voltage and current. We simply don't try to do it. We adopted a simpler and we think more accurate approach. We have designed our systems so that the power input to the system is always low-voltage DC (6 to 40 volts) -- this can be be accurately and reliably measured as a power input.

Then we convert that DC to high-voltage AC within our apparatus, and the converting electronics are enclosed in a separate and unique calorimeter. This allows us to precisely measure the energy loss of the electronics. By the way, this is also the method that the vendors of wattmeters use to calibrate and test their wattmeters. Ultimately, measurement of heat output via calorimetry is the most "true" form of watt meter. So why not just use that method to start? The above method has proven to be very simple, reliable, and robust for us. But, your mileage may vary.

L.K.:
I know that R4 and R5 work with a gas glow discharge cell. Could their approach be used in working with Mizuno-type electrolytic plasma cells? I do not think so. Here is my reply to the above message. Suppose you begin with a d.c power supply whose voltage is very stable, for example, 60 V. The voltmeter shows volts and ammeter shows current feeding the next step. It the current is 5 A then P1=300 W. The next box is a high voltage power supply. It receives electric energy at the rate of 300 W and uses it to generate, for example, 500 V. You apply this to your gas glow discharge cell. Right?

At what rate is the cell receiving electric energy? The rate is P2=P1-X, where X is part of P1 that is converted into heat in your high voltage supply. You must know X to calculate P2. That is why your entire high voltage power supply is located in a calorimeter. That box has only four wires connected to it, two at the input and two at the output. Input receives electric energy at the rate P1 and the output delivers it to the cell at the rate P2. After waiting many hours the temperature of the calorimeter stops rising and you calculate X. Suppose it is 50 W. You know that P1=300 W because your d.c. is 5 A. Therefore, you know that P2=250 W.

That is a good method when the glow discharge is very stable. But suppose it is not stable because things change inside your cell. You are aware of changes because the d.c. is going down or going up. Your ammeter reacts to all changes at once but your calorimeter takes hours to react. Suppose that the d.c current changed from 5 A to 10 A. Thus P1 becomes 600 W, for exactly 10 minutes. Then the current becomes 7 A and P1 becomes 420 W. Your experiment ends 30 minutes later. How many joules of electric energy was supplied to your cell during the last 60 minutes? I do not think that the answer would be correct if you used the same X at all P1. That is a limitation of your method. I suppose it is highly reliable when P1 remains constant during the entire experiment. To conduct an experiment at a different cell voltage you would have to wait many hours to find the new value of X. Do you agree?