Alan - the manufacturers for spec re power and velocity cannot directly be compared with the actual power and velocity because the "rated power" and velocity will be under some back pressure load condition maybe different from that in Mizuno's system. So this graph can be over-estimating the disparity. But I agree it still does not make sense, which is evidence in favour of the airflow being over-estimated.
What assumption is made about back pressure in the fan law curves?
I am stating the (obvious) fact that even if the anenometer results have some consistent error, the adjusted results (comparing active runs with a known power out calibration) should not be affected.
If the anemometer had a consistent error, it would not agree with the other anemometer, or the smoke test, and the calibrations at low power would not show an energy balance very close to zero. The "known power out calibrations" would reveal the consistent error.
However, I also point out, that reliance means it is important to establish that conditions when the control reactor and the active reactor arun are identical (so that in these two different cases the calorimeter losses are the same).
No, that is nonsense, as I have pointed out many times, and as anyone who has used a calorimeter will know. You do not need to make the active and control reactor identical. The calorimeter reacts the same way with a 50 kg reactor as control, or a 20 kg one, a 300 g one, or a resistance heater placed by itself in the calorimeter. The data points fall right on top of one another. You cannot tell them apart.
There is no reason the losses from different sized reactors would be different. The surface temperatures of the control reactor sitting next to the active reactor is the same in all cases, at the same power level, so the inner wall temperature must be the same as well.
Mizuno did, actually, make a point of using identical reactors as control and active. I almost wish he hadn't, since you are now using this as a bogus reason to claim his more recent tests are invalid. He did not do this because he thought that different reactor sizes or shapes would cause a different response in the calorimeter. That is impossible. It is ludicrous. He did it so that he could switch from active to control midway through the test without opening the calorimeter chamber.
With 100 W of input, 84 W are recovered in the stream of air. That is the same with any sized reactor or a bare resistance heater. So that is 16 W lost from the walls. Always 16 W. Never more, never less. So what do you imagine is happening? During one of the R19 excess heat runs shown in Table 1, on May 15, 2019, with 100 W of input , 110 W were recovered from the air alone. You don't even need to consider losses from the walls. Accounting for them gives 131 W of input, but if you have some doubts about the losses being correct, feel free to ignore them. It is not possible to recover 110 W from the air with 100 W input. It could be no more than 84 W. There is no magical way to make the air warmer than it would be. The simplest approximation is mass of air * specific heat * degrees K. By that approximation the air is much warmer than it can be from 100 W. There is no way it could be an error in the air flow rate, input power, or temperatures (with a 7 deg C Delta T). Instruments do not make errors on that scale.
(Mizuno's actual equation is more complicated but the answer is close to that.)
What assumption is made about back pressure in the fan law curves?
Do not make assumptions. Measure the values with different instruments and techniques. You will find that the manufacturer's specifications are correct, but it is good idea to check. The manufacturer knows about fan law curves, so you don't have to worry about them. The manufacturer does not speculate or rely on theory. They measure actual performance. You can measure it too.
This paper is what I refer to: J. Condensed Matter Nucl. Sci. 29 (2019) 1–12
Ah, the ICCF21 paper: https://www.lenr-canr.org/acrobat/MizunoTexcessheata.pdf
Yes, Fig. 8 shows variance. I believe it is caused by those damned ambient temperature changes. They make precision impossible.
Yes, there is variance for when you measure things for a short time. You have to do long measurements in quiet conditions to get something like Fig. 9.
The new paper Fig. 7 shows the effects of ambient temperature changes. Here is the new version of that. This like trying to conduct an orchestra in a boiler factory:
Just for the record, here's a chart of the Mizuno M19/R20 callibration points, and the Manufacturers spec.
I do not know what to make of this, but I will trust the two anemometer readings, and the calibrations.
If you get a blower and an anemometer, I suggest you measure the actual values and compare them to the specs.
paradigmnoia bought a hotwire anemometer...and a fan, and measured 5m/s or so.
he did spec calcs too where the spec conflicted with his measurement
They measure actual performance. You can measure it too.
Wow THHnew and AF collaborating with Paradigmnoia to measure actual performance..
Paradigmnoia can write a thesis on it maybe..
As long as I am at it….
A.) Basics of air flow calorimetry - Jacques Ruer SFSNMC [email protected]
“Any fluctuation of the inlet temperature results in an error of the heat flow measurement. The heat storage capacity of the whole system introduces a time lag on the readings”
I find the change in inlet temp on the calorimeter as described in the 2017 paper (I believe, see following) to be very suspicious.
B.) In the spreadsheet “Mizuno 2017 120 W input excess heat”, the delta-T seems to show an abrupt jump when then input power is turned off (~21,500 sec) with a concomitant decrease in noise level. (The deltaT plot is on the spreadsheet. Note the change at the ~21,500 sec point.) This seems to be primarily due to outlet T which shows some significant negative deviations from the expected values starting at that point. See below. The inlet T starts decreasing when the power is turned off. The outlet T is very messy. This all says ‘electrical noise’ to me, but that’s just me…
Oops! Got inlet and outlet reversed in the title of the plot in my last Figure. Outllet > inlet. So noise is in inlet, not outlet.
There could be a problem with this calorimetry. I have been looking for one for months. Until someone replicates, I cannot rule out that possibility. But we are not going to find it by inventing bogus nonsense. For example, by saying we cannot know what the average is across time, when the same value to within a few percent is measured for months at a time with the anemometer, and when the power going into the blower does not vary by even 1%. As I said, the average of 4.1 repeated thousands of times is 4.1. The average across the face of the outlet orifice, right up to 3 mm from the edge, is also 4.1. Granted, that is the smallest unit this anemometer and the other one could measure, but that's what it was. Measured many times. Checked several times a day for months. Disputing the "average" when the value hardly changes is not how you do science. Any fourth grade kid knows that when the same number is repeated, that is the average. Why does THH even bring this up? What is the point?
Mizuno and I carefully checked through the data, graphing it and looking for unexpected variations and problems like this. Of course we might have overlooked something! I would never say that's impossible. I have made too many mistakes to say that. But when I tell you, "yeah, we thought of that, and here is why we do not think it is a problem" you should either say: "okay, cross that off the list" or: "ah, but maybe your method does not work, for thus and such a reason." Don't ignore what we say and keep repeating the same thing.
For the last few years, I have been over some of these spreadsheets time after time, looking for problems. I look for problems that occur to me, and problems suggested by many other helpful people. I have no objection whatever to people saying "What if the flow is not mixed? How do you know it is turbulent, and not laminar?" Good question! The person who peer-reviewed the ICCF21 paper brought this issue up first thing. So Mizuno added the traverse test, and we went over the details of that test carefully. (By the way, that reviewer did an excellent job, pointing out many weaknesses and potential problems.)
It is very helpful to suggest real potential problems. What is NOT helpful, ever, is to say that 4.1, 4.1, 4.1, 4.1 repeated 14,784 times might not average 4.1. The spreadsheet tells me it does average 4.1. I checked. I checked many times.
It is also not helpful to suggest a potential real problem, such as "the flow might be laminar" but then when we show you we thought about that, and checked for it, to keep insisting it might be a problem. If you do not think the traverse test and calibrations are sufficient to disprove the flow is laminar, okay. But you must then suggest some other way to prove or disprove your hypothesis. Make it falsifiable. Tell us what we should do to check for this. We may have already done it.
The outlet T is very messy. This all says ‘electrical noise’ to me, butthat’s just me…
This all says ‘electrical noise’ to me,
except the time period is many hours.. not a few seconds
As I recall KS gave up on this in 2017... saying he refused to analyse the spreadsheet data by mcp delta T.
because he didn't trust it
Good to see that KS managed to at least graph the data this time
If KS still has the data there on his PC can he regraph the socalled electrical noise area
with a time scale which has a range of 10 minutes..
and then post it back here..on this thread
So noise is in inlet, not outlet.
Yup. That's where the noise is, every time. It gets so bad that Mizuno cannot conduct experiments for 2 or 3 months a year.
Believe me, we know about it. I have often made graphs like the one I uploaded just now, with the ambient temperature fluctuating and dropping over time. A change of 0.1 deg C comes out as 2 W with this calorimeter. You can see how big a problem that is. Here's the graph again:
It is a little hard to measure from this image but if you print it out you will see that some of those short-term fluctuations exceed 0.1 deg C, and you will see the calibration constant is about 0.055 deg C/W, as I said.
There are, however, actual fluctuations in the anomalous excess power that do not correlate with ambient noise. I look very carefully to be sure of that. Sometimes the magnitude exceeds the possible effect of the temperature changes. Sometimes it does not correlate with ambient changes. (As far as I can tell, anyway.) Here is a recent example, with the new screen in the R20 reactor. I will present this at the conference in two of the slides. I don't know what to make if it, but let the audience decide:
The cycles at the beginning are 1.5 to 3 hours long, and much larger than ambient changes can explain. They correlate with spontaneous loading and deloading of the sample:
That is not under Mizuno's control. He didn't make it happen. He says it happened "自動的" (jidouteki) meaning "automatic, spontaneous." I think a calorimetry error is ruled out.
It is a darn shame the new mesh is not producing 250 W, but at least it is doing something new and interesting. It may improve. You can't leave the old one in there indefinitely. It has to come out and be analyzed, or you will never learn anything about the nature of the reaction.
The R19 with the mesh shown in Table 1 is still intact, still working, and still producing roughly as much heat. A little less, but just about as much. It was last tested on July 18. I will show those results at the conference.
QuoteA change of 0.1 deg C comes out as 2 W with this calorimeter. You can see how big a problem that is. Here's the graph again:
A temperature-controlled-liquid cooled Seebeck and/or mass flow calorimeter would not be as sensitive to environmental temperature changes if correctly designed, made and calibrated. I've never used air flow calorimeters but, based on personal experience with a variety of Seebeck designs, I'd guess the difference would be very significant. There is no reason why a liquid cooled Seebeck "envelope style" calorimeter can not also supply data from the coolant which can be used for mass flow calculations. And yes, to do this and keep the walls hot enough for Mizuno's purposes could be harder than the air calorimeter- probably is.
QuoteIt is a darn shame the new mesh is not producing 250 W, but at least it is doing something new and interesting. It may improve. You can't leave the old one in there indefinitely. It has to come out and be analyzed, or you will never learn anything about the nature of the reaction.
Vehemently disagree. You should never take down a working experimental setup if you expect it to be excruciatingly difficult to set it up again. Rather, leave the working device alone and try to make a new one that works and when you have done that, analyze it- the new one. IMO, the analysis can wait. Confirming Mizuno's startling results to a skeptical scientific and investor community that has ample resources to analyze, duplicate and exploit is can not. Obviously, my priorities are not yours or Mizuno's. But I have to point out that neither of you has gotten very far attracting large investment or interest outside the LENR "community" (aka usual suspects).
But I have to point out that neither of you has gotten very far attracting large investment or interest outside the LENR "community" (aka usual suspects).
You have no evidence at all that that is the case, you just made it up because it is difficult to refute without naming names.
ETA (in magnified font): But I have to point out that, as far as I know, neither of you (Jed or Mizuno) has gotten very far attracting large investment or interest outside the LENR "community" (aka usual suspects).
TIFIFY
I'd be delighted to hear otherwise. As per JedRothwell ,Mizuno says he has no secrets about the current device so I have no idea why he would not reveal new large support if he had it. It also would make no sense that Jed describes him as working with the budget of a church mouse, if he was, in a figurative sense, well endowed. But hey, have at it.
(aka usual suspects).........
(aka usual suspects).
I guess the repetition is for emphasis..
Sounds like they are criminals ... is it just SOT aka MY's sensationalist habit to make it sound like LENR is full of shady types?
ETA (in magnified font): But I have to point out that, as far as I know, neither of you (Jed or Mizuno) has gotten very far attracting large investment or interest outside the LENR "community" (aka usual suspects).
TIFIFY
I'd be delighted to hear otherwise. As per JedRothwell ,Mizuno says he has no secrets about the current device so I have no idea why he would not reveal new large support if he had it. It also would make no sense that Jed describes him as working with the budget of a church mouse, if he was, in a figurative sense, well endowed. But hey, have at it.
Surely you would agree that both have been as open as they could be? If not, how would you be more open in their shoes? From what I see, Rothwell should be commended for being here most days, answering in detail whatever he can. Mizuno the same, as he opted to get his results out ASAP to get it vetted, instead of keeping it a secret for years. Since then, he has been busy with colleagues, experiments, visitors, filling Jed in, and trying to run a business.
And really, at this stage would you expect "large investments" to be pouring in? And if he did, would you expect him to reveal it?
Patience....grasshopper.
QuotePatience....grasshopper.
I'm no f'n grasshopper. Get that through your skull. And outside, maybe, of flying, you're not a sensai.
Display MoreThere could be a problem with this calorimetry. I have been looking for one for months. Until someone replicates, I cannot rule out that possibility. But we are not going to find it by inventing bogus nonsense. For example, by saying we cannot know what the average is across time, when the same value to within a few percent is measured for months at a time with the anemometer, and when the power going into the blower does not vary by even 1%. As I said, the average of 4.1 repeated thousands of times is 4.1. The average across the face of the outlet orifice, right up to 3 mm from the edge, is also 4.1. Granted, that is the smallest unit this anemometer and the other one could measure, but that's what it was. Measured many times. Checked several times a day for months. Disputing the "average" when the value hardly changes is not how you do science. Any fourth grade kid knows that when the same number is repeated, that is the average. Why does THH even bring this up? What is the point?
Mizuno and I carefully checked through the data, graphing it and looking for unexpected variations and problems like this. Of course we might have overlooked something! I would never say that's impossible. I have made too many mistakes to say that. But when I tell you, "yeah, we thought of that, and here is why we do not think it is a problem" you should either say: "okay, cross that off the list" or: "ah, but maybe your method does not work, for thus and such a reason." Don't ignore what we say and keep repeating the same thing.
For the last few years, I have been over some of these spreadsheets time after time, looking for problems. I look for problems that occur to me, and problems suggested by many other helpful people. I have no objection whatever to people saying "What if the flow is not mixed? How do you know it is turbulent, and not laminar?" Good question! The person who peer-reviewed the ICCF21 paper brought this issue up first thing. So Mizuno added the traverse test, and we went over the details of that test carefully. (By the way, that reviewer did an excellent job, pointing out many weaknesses and potential problems.)
It is very helpful to suggest real potential problems. What is NOT helpful, ever, is to say that 4.1, 4.1, 4.1, 4.1 repeated 14,784 times might not average 4.1. The spreadsheet tells me it does average 4.1. I checked. I checked many times.
It is also not helpful to suggest a potential real problem, such as "the flow might be laminar" but then when we show you we thought about that, and checked for it, to keep insisting it might be a problem. If you do not think the traverse test and calibrations are sufficient to disprove the flow is laminar, okay. But you must then suggest some other way to prove or disprove your hypothesis. Make it falsifiable. Tell us what we should do to check for this. We may have already done it.
Looking at the discrepancy between theory and measurements there led to the insight that the flow after the fan had not fully developed into its eventual state - because the pipe was not long enough. That is helpful.
I think your concern about averages is because you are not distinguishing between spatial (across area) average and time average. You will notice that I'm not questioning the traversal data which is why I reduced that component of the airflow uncertainty. However, we now thanks to AF have two things:
(1) HWAs do over-measure average velocity in turbulent (time-varying) flow
(2) The fan calibration results show the fan working significantly better than the manufacturers typical flow data graphs - or else Mizuno is operating the fan above its max operating voltage of 13.8V. (If so, he should expect the fans to stop working). Even then the difference between spec and measured airflow is pretty large.
So that introduces a completely different reason to question the air speed element of the absolute power out calculations.
WRT using the calibration data to validate the airflow measurement we have two related matters:
I'd expect that the calorimeter efficiency measurements at low powers (5% loss) could be used to bound airflow measurement error as long as the setup and blower power used there remains the same as that used for real.
Finally WRT "3 power meters can't be wrong"
The problem is that nowhere are we told what these power meters measured. Was it is input to the heater? Or the input to the PSU that drives the heater (as Jed said)? And for which elements of the recorded data was this done?
Jed has a lot of "we've tested this, we've tested that, it can't be wrong". My long experience tells me that in any complex system that type of informal validation can go wrong, because the exact relationships between the various bits of validation sometimes lead to surprising results. I can't say that is a problem here, but the various potential problems, and the various unanswered questions, leave that open.
Jed's insistence on absolute measurements here that do not have full validation is unnecessary. Robust and careful use of control (calibration) data from an identically sized and situated reactor would be a route to knocking on the head all these complaints.
The paper would need to state the methodology used showing precisely what varies and what remains the same between the calibration and active tests, detailing the calibration results with the test results. My understanding at the moment is that the two runs are done using heat emitters (reactor bodies) in a different position in the enclosure where the airflow could be very different. All that is needed here is another calibration - replacing R19 by an R19-shaped reactor in the same position as R19 was, with all other elements of the system identical. And the calibration done with the same setup as the active run.
So the question marks are:
NB - Kirk's point about room temperature change was highlighted on this thread very early on (echoing GSVIT who did some work). Room temperature changes matter here because the reactor thermal time constant is large, so a change in the input air temperature will not immediately propagate to the output temperature. Therefore any slope on the input will lead to a consistent error in the output. This issue matters for calibration and active runs. The Mizuno spreadsheets do track input and output temperature I believe, which would allow this to be investigated. It needs to be done.
Now, Jed is confident that none of these questions are problematic, which may well be true. In most cases I'm sure he will be right. All that is needed is the evidence of that to be presented, one by one. Assertions that one part of the calorimetry must be Ok because otherwise the error would show up somewhere else do not cut it.
For replicators: please note these issues, and address them all, one by one. Or - rely only on control versus active differential data, in which case the air flow measurement uncertainty does not matter but the other things do. Many of these issues are just methodological, keeping consistent and clear records, making the equipment, setup, timing of control versus active runs clear by time stamping results and calibration, and detailing precisely all equipment or measurement changes.
For replicators: I've pointed out the importance of cross-checks. Here is another one that would deal with the issues about airflow. Have a similar tube on the calorimeter input orifice. Measure air velocity, similarly, in that. The differences will be a very different flow, uninfluenced by fan turbulence. If as Jed thinks HWAs are completely accurate there will be no difference.
THH
I'm no f'n grasshopper. Get that through your skull.
Sounds like the sensei has a rebellious student
And really, at this stage would you expect "large investments" to be pouring in? And if he did, would you expect him to reveal it?
I would answer no and no... anyway I think Jed is focussed on the presentation at Assisi..technical stuff.. not so much of the non technical
