A Simple Test for Large COP Claims in Pulsed Inductive Heater Circuits

[poynt99]

Indeed, Rose Ainslie is a special case, and stealing a phrase from our good friend TK, I can scarcely believe that I am still trying to teach her the finer points of challenging scope measurements. 

Last night I almost lost my mind when after arguing with her for an hour or so about how a current probe will give the same, but better results than a CSR measurement, I finally got out of her the fact that she assumed the current probe would have to be connected to a separate oscilloscope from the one taking the voltage samples, and hence her "objection" to the current probe approach!


I have hopes that the testing of Lawrence's device will have a much better and quicker outcome, and this will pave the way for more objectivity regarding testing of these devices in the future.

Have some faith Maestro  

.99

[poynt99]

As far as Free Energy Forums go, not only do I see no reason why several levels of discussion can not co-exist here in harmony, but so far on this site, I have seen more civilized, interesting, educational, inspiring, professional, cooperative, and respectful interactions than anywhere else on the web.

.99

[humbugger]

"Indeed, Rose Ainslie is a special case, and stealing a phrase from our good friend TK, I can scarcely believe that I am still trying to teach her the finer points of challenging scope measurements."

Perfect point to ease back into the the topic of this thread.  The whole idea here is to convince Rosemary et al that learning all the nuances of precise scope setups and complex math should not be a prerequisite to testing a claimed COP 17 device.

Anyone have comments, questions or arguments regarding what I've put forth here so far?

(It feels a bit awkward hinting to a Major Top-Dog Moderator and the forum Admin that they may be wandering off-topic slightly...especially when the comments are so excellent regarding the general state of things on the forum.)  

[poynt99]

POYNT

I had earlier edited that post to read "reasonable expertise is expected and assumed" as credentials often are not the measure.

It was not picked up in your reply.

It's ok, as I did not take the word "credentials" in the literal sense anyway. I knew what you were getting at, and I responded accordingly.

.99

[poynt99]

humbugger,

I fully support any method that cuts to the chase and is easy to implement. Yours and ION's suggestions fit the bill.

Be clear that I am not espousing the oscilloscope method as the only one acceptable for determining if a device is COP>1. It is one that can be precise if done correctly and with the correct probes. Many folks are eager to use their scopes to determine COP, but sadly with these type of devices, it is beyond the means of most to properly do so. But we are trying.

Carry on, I see no problems guys.

.99

[ion]

Sorry if I derailed the thread, perhaps those comments of mine belong in the "rant room", as it is more a cry of frustration.

You can understand the frustration, when folks like Tseung can "lift" the humble blocking oscillator from anothers website, claim it as their own invention, exalt it to lofty overunity device, and tie up so much time, then present no viable evidence that it is so, as many hold their breath for the professors blessing.

I have to wonder what the fuss is about and re-examine the definition of insanity.

Didn't mean to have it spill into this thread. I'll save the rest for the Rant Room.

[ion]

Here is the temperature profile of my styrofoam test chamber. I used a cheap indoor outdoor thermometer from Radio Shack, since it is a budget test absolute accuracy is not required, repeatability is. The unit reads out in tenths of a degree. I used F for greater resolution.

Temperature shown is rise over ambient. I used a 100 Ohm five Watt resistor 1%. Appropriate DC voltage was applied to produce the required power.

Looks fairly linear over the tested range. The temperature should be repeated several times with values recorded at each interval of power on the way up to max and back down, this done a few times to get good averaged data.

When the DUT and its load is put in the box, I will have no difficulty interpolating total power ie. that consumed by the DUT plus that delivered to the load. Each can also be tested separately.

[Grumpy]

Ion,

You replicated this and found 75% efficiency?

[MileHigh]

Ion:

Nice to see that graph!  You and Humbugger have posted some great ideas for thermal testing procedures and setups and as you know I am a fan of the thermal method myself.  The plot is a nice straight line which is exactly what it is supposed to be.

How long did you have to wait for the temperature to stabilize in your particular setup?  That's a critical thing for all experimenters to take note of.  Depending on the configuration, it might take one minute, five minutes, 15 minutes, etc.

Once you take a series of data points like you did you can plot the straight line graph that gives you temperature vs. power.  Your example clearly illustrates how easy this is to do.

Then you put the DUT into your thermal box like you stated and wait for the temperature to stabilize.  Then you take your stabilized temperature and compare it to your graph and you get the power output of the DUT.  It's so easy it's like taking candy from a baby.

The JT testing we have discussing is an excellent candidate for this procedure.  I made allusions to this in my recent postings but it helps tremendously to see the real thing "in action" like you clearly demonstrated.

MileHigh

[ion]

Ion,

You replicated this and found 75% efficiency?

No G, I tested at 75% efficiency the elementary blocking oscillator shown in the drawing "BO3 Test" of post 168 here:

http://www.overunityresearch.com/index.php?topic=538.150

I have not tested the Tseung version of the blocking oscillator as this has non-linear loading (LED's) and is extremely low power. When the signal is so far in the noise, it is easy to get bad data and interpret it as OU.

This should probably be tested using a very sensitive thermal method or DSO.

The device also has no practical application in it's current form that I can think of.

I have little faith in the Tseung knock-off as circuits like this are crudely thrown together with no attention to careful design or optimization. They are seen all over the web named "Joule Thief".

From MH

The JT testing we have discussing is an excellent candidate for this procedure.  I made allusions to this in my recent postings but it helps tremendously to see the real thing "in action" like you clearly demonstrated.

Thanks for the comments but it seems the Tseung device is so low power, I would have to use an expanded scale from 0 to 1 Watt max. I used about 15 minutes between readings, but greater accuracy could be had as I stated earlier by taking readings on the way up and on the way down, then averaging.

[MileHigh]

Thanks for the comments but it seems the Tseung device is so low power, I would have to use an expanded scale from 0 to 1 Watt max. I used about 15 minutes between readings, but greater accuracy could be had as I stated earlier by taking readings on the way up and on the way down, then averaging.

Yes indeed the Tseung device is really low in power.  Let's suppose we don't put the entire DUT inside a thermal chamber of some sort, but are going to test just the LED and the two resistors. This is perfectly valid since that's the claim and how it is likely going to be tested with the DSO.

So you take those three components and scrunch them together and gob them with thermal paste then wrap everything with some tape to make something about the size of the tip of your finger.  Then you put a thermocouple in contact with  the surface of the tape and tape that in place.  Then put the assembly into a a small box like a jeweler's box for holding a ring.  You also have cotton batting in the box so the thermal mass is in the center of the box surrounded by the cotton batting.  You place the box on top of a piece of Styrofoam and make sure that the ambient air currents are always the same.  You end up with a small box with to wires exiting for the load terminals and the two wires for the thermocouple.  That should be a stable and workable configuration that produces a decent temperature differential.

I am not sure what you mean by taking readings on the way up and then down and averaging.  The temperature vs. time curve would be a standard exponential curve that approaches it's asymptote after five time constants.  So I was alluding to making your final temperature reading after you are confident that about five time constants have passed.

MileHigh

[ion]

From MH

I am not sure what you mean by taking readings on the way up and then down and averaging.  The temperature vs. time curve would be a standard exponential curve that approaches it's asymptote after five time constants.  So I was alluding to making your final temperature reading after you are confident that about five time constants have passed.

Let me see if I can explain this. When characterizing the styrofoam test chamber with internal power resistor and stepped power input, wait x time constants, record data of point on asymptote, then go to next power level.

When the highest  (in this case 5 watt point) is recorded, reverse the procedure going back down each power step and record the data after x time constants. The real value  of temperature lies between the mirror image asymptotes, which is easily found by averaging.

This is not necessary, but if you desire to speed up the characterization, you can wait fewer time constants and still get acceptable results with the up/down method.

It's just a dithering method to get the actual value.

For best accuracy the DUT on the other hand should be left to soak for very many time constants before taking the final reading of temperature.

Your suggestions of how to test at yet lower power levels are well taken.

[MileHigh]

Humbugger:

I am starting to read your thread from the beginning and I noticed this comment:

Depending on how fast the DUT can raise the temp 30F, a higher number may be better here.  We will be timing how long it takes to achieve a given temperature rise, not the end temperature.

Using relative temperature rises of a fixed amount won't work here.  You can't ignore the fact that the temperature rise vs time is not linear so the slope of the curve is always changing.  So DUTs that run at lower power levels will take longer to create a 30F increase in temperature than DUTs that run at higher power levels.  The safest bet is to wait five thermal time constants.  Also, note that if the thermal test setup is already at a higher temperature than ambient at the start of the experiment, it's still going to take five time constants to be very very close to thermal equilibrium.

Another obvious point is that if the power output of the DUT is low, it may not be able to generate a 30F temperature rise.

I am not 100% sure that I am understanding exactly what you are saying, so I could be spouting a whole bunch of hot air.  I suppose my real point again is to measure a final temp when you are very certain that you are at thermal equilibrium.  I tend to be very conservative when it comes to things like this so at least that would be my strategy.

Here is the curve for reference (look at the top curve only where you can say temperature is on the vertical axis and time is on the horizontal axis:



Note also that the temperature vs. time curve is "approximately" an exponential curve.  The thermal resistance is not necessarily a pure constant as the temperature reading of the DUT increases.  However, we don't need to worry about this at all.

On the "fun" side of things, anybody know what the curve that describes the changing slope of an exponential curve looks like?  It's a sort-of trick question.

[MileHigh]

Another comment:

As the load gets hotter and hotter compared to the surrounding air, it also dissipates the heat at a faster rate due to increasing convection.  This means the temp rise vs power in the load is not linear and gets flatter as the temp gets higher.  A given increase in power will cause less and less temperature increase as the load gets far above ambient temp.  This trends toward a more error-prone measurement and may take a much longer time to adjust the control setup to exactly match the DUT's final temp.  Someone please jump in here and beat me up on this if I'm wrong, but I don't think so.

You are correct and alluded to this also when I stated in the previous posting that the thermal resistance may change as the temperature gets higher because the air convection may (or may not) have a non-linearizing effect here. (I'm not really sure)  However, all this means is that the thermal equilibrium temperature vs. power graph (like Ion posted) may not be a perfectly straight line.  So if you get the equilibrium temperature from the DUT you can still read off the equivalent power from the graph.  There are no real error effects here.

MileHigh

[MileHigh]

I am going to respond to Ion's quick test that he did with a blocking oscillator a.k.a. Joule Thief in his lab.  Note that he tested the pulsing output of the JT transformer and did not wrap an extra energy pick-up winding around the toroid as in the Lawrence Tseung configuration.

If you go back to his original posting you can see the block diagram.  He is using a large capacitor on the input, to convert the pulsing current load of the JT into smooth DC current that can be easily measured by a multimeter.  On the output of the JT he takes the pulsing current output of transformer's secondary winding and also converts that into smooth DC current so the output current and voltage across a load resistor can also be easily measured by multimeters.  The logic here is simple:  He is measuring the average energy output from the discharging secondary transformer winding in the JT.

While all the arguing was going on, I built up a crude blocking oscillator and tested its efficiency at 75.0% using the method in post #168 of the Tseung thread. This took less than an hour.

I drew up a two step method outlined in the attachment and had intended to post it early in the Tseung thread, but as the existing described methods were misunderstood by many, I realized the purpose of the thread seemed more about discussion.

This should be a very sobering moment for anybody that thinks a JT circuit in whatever configuration is a potential over unity device.  The numbers don't lie.

MileHigh

[MileHigh]

Well, I finally caught up on this rather interesting thread.  The Simpson 260 is just a tad before my time.  From the comments through I can see how it's equivalent for a disco DJ (vinyl era) would be the Technics SL-1200.  Never was a turntable so perfect!  It must have been a pure accident of design!  lol

An interesting comment about myself is that although I have spent many an hour on the bench it was doing digital design and debugging and the barest wisp of analog.  I have never investigated any over unity or free energy concepts in real life.  However, I still have a brain!  And of course I did do hundreds of lab experiments covering all of the basic building blocks that are discussed around here on a daily basis.  It is very easy for me to extend my "reach" into the mechanical domain also.   I can look at a YouTube clip (a rare occurrence nowadays) and pretty much figure out exactly what is going on most of the time.  I have pretty good analytical skills with respect to this subject matter in general.  I am very good at qualifying what's going on on the bench and what's going on between the stool and the bench.  I am also very adept at qualifying the "business" side of things.

MileHigh

[ion]

The actual numbers from the BO3 Test (attached) are as follows:

Input: 10.0 Volts @ .075 Amps =0.75 Watts input

Output: 7.5 Volts / 100.0 Ohms=0.5625 Watts output

Efficiency = Output / Input=0.5625 Watts / 0.75 Watts=0.75 x 100 = 75%

Joule Thief people will scream and yell that I am killing the operation of the circuit by adding the catch diode and capacitor on the output.

This test was meant to show that a simple blocking oscillator as a DC to DC converter with no optimization can yield 75% at the high input voltage of 10 Volts.

Efficiency will be less for the same circuit adjusted to operate at a lower voltage e.g. 1 Volt because semiconductor and copper losses will begin to predominate.

There are many things that can be done to improve the lowly blocking oscillator, such as optimizing base drive current requirements, anti saturation techniques for faster switching etc. I won't get into that here.

Unfortunately, I see very few if any in the JT crowd putting on their thinking caps and trying to really understand this single transistor oscillator, most are cut and try jockeys.

When I re- measure my test chamber for expanded scale between 0 to 1 Amps, maybe I'll throw the Tseung JT into the box and nail the coffin on this one.

[humbugger]

Sorry if I derailed the thread, perhaps those comments of mine belong in the "rant room", as it is more a cry of frustration.

You can understand the frustration, when folks like Tseung can "lift" the humble blocking oscillator from anothers website, claim it as their own invention, exalt it to lofty overunity device, and tie up so much time, then present no viable evidence that it is so, as many hold their breath for the professors blessing.

I have to wonder what the fuss is about and re-examine the definition of insanity.

Didn't mean to have it spill into this thread. I'll save the rest for the Rant Room.

Don't worry about derailing "my" thread.  Your interest and contributions far outweigh any minute distraction your voicing of frustrations adds.  I completely empathize.

[humbugger]

Here is the temperature profile of my styrofoam test chamber. I used a cheap indoor outdoor thermometer from Radio Shack, since it is a budget test absolute accuracy is not required, repeatability is. The unit reads out in tenths of a degree. I used F for greater resolution.

Temperature shown is rise over ambient. I used a 100 Ohm five Watt resistor 1%. Appropriate DC voltage was applied to produce the required power.

Looks fairly linear over the tested range. The temperature should be repeated several times with values recorded at each interval of power on the way up to max and back down, this done a few times to get good averaged data.

Looks exactly as I would expect for a closed and insulated small chamber.  One point I didn't probably make clear regarding my approach is that I'm tending toward open-air benchtop testing of the temperature.  That is why I expect that my curves would not be linear, with diminishing temp rise per watt as the air convection stirred up by a large delta between the load and ambient increases the rate of heat tranfer to the air.

[ion]

humbugger

The open air approach is certainly a good one with one caveat: it becomes difficult to capture all the heat output of the various components.

It is great when measuring a load resistor, provided the physical orientation is the same for each test so that convection currents remain the same.

As there has not been a real idea of what folks want to measure, it is sometimes difficult to make a choice of which method is best.

Each has it's place.

[humbugger]

Using relative temperature rises of a fixed amount won't work here.  You can't ignore the fact that the temperature rise vs time is not linear so the slope of the curve is always changing.  So DUTs that run at lower power levels will take longer to create a 30F increase in temperature than DUTs that run at higher power levels.  The safest bet is to wait five thermal time constants.  Also, note that if the thermal test setup is already at a higher temperature than ambient at the start of the experiment, it's still going to take five time constants to be very very close to thermal equilibrium.

I agree that the stopwatch-based method testing the time it takes for each setup to cause the load temperature to rise from ambient to some arbirary setpoint (above ambient but well below equalibrium) is not a good approach in general for OU claims...but I come to that conclusion for a completely different reason.

The method assumes that the black box immediately begins performing at its nominal claimed COP as soon as it is energized.  I don't want to hear arguments based on the idea that some unknown amount of warm-up time is required before a given DUT comes "up to speed".

Given a DUT that starts immediately at full COP performance, I don't see how using the rate of change of temperature well below thermal equilibrium (the steepest portion of the curve) would not correlate directly to power in the load.  If the DUT draws 10W from the battery and moves the load from 70F to 120F in ten seconds, for instance, and the control test ends up requiring 20W into the same load to get the same rate of rise, then the DUT has a COP of 2.  Is this not correct?

I think you'd agree that if thermal equilibrium takes half an hour to achieve (extreme case) and we are measurring how much it rises in the first ten seconds, for instance (extreme case again) we'd be operating in a pretty-darn-near linear portion of the curve.

[ion]

From humbugger:

Given a DUT that starts immediately at full COP performance, I don't see how using the rate of change of temperature well below thermal equilibrium would not correlate directly to power in the load.  If the DUT draws 10W from the battery and moves the load from 70F to 120F in ten seconds, for instance, and the control test ends up requiring 20W into the same load to get the same rate of rise, then the DUT has a COP of 2.  Is this not correct?

Yes with one caveat: I would guess the thermal masses (components) will change the rate of rise depending on whether they are emitting heat or absorbing heat.

e.g a small resistor will cause a sudden rate of heat rise only to be absorbed later by the more massive transformer core.

So I guess it depends on whether you are testing all the components in a box or just the load resistor.

Any thoughts?

[humbugger]

humbugger

The open air approach is certainly a good one with one caveat: it becomes difficult to capture all the heat output of the various components.

It is great when measuring a load resistor, provided the physical orientation is the same for each test so that convection currents remain the same.

As there has not been a real idea of what folks want to measure, it is sometimes difficult to make a choice of which method is best.

Each has it's place.

Yep, there are good arguments for putting the DUT and its LOAD in a chamber to measure the total heat produced, but I prefer to use a common load (chambered or not) and keep the DUT outside.  Then just compare how well the DUT heats the load for its given input power against how well a DC source and pair of fat wires heats the same load at an equal power.

I am focusing on claims that state "I have a magic box that heats a load using less power than just heating the same load with straight AC or DC power".

Given your correct comments regarding allowing the big load resistor to be in free air on the bench and ensuring identical airflow conditions, I just had to draw up this little cartoon.  Hope to hear the laughter over the roar of those two huge fans 

[ion]

We are in agreement, as this is what I posted early on. I only changed to total testing of DUT plus load resistor as there was some initial resistance to not including the measurement of power dissipated in the DUT itself.

I pointed out then, that this can always be arrived at as the difference between Pout and Pin., but there were further objections.

RE: the cartoon....good one!

[MileHigh]

Humbugger:

Given a DUT that starts immediately at full COP performance, I don't see how using the rate of change of temperature well below thermal equilibrium would not correlate directly to power in the load.  If the DUT draws 10W from the battery and moves the load from 70F to 120F in ten seconds, for instance, and the control test ends up requiring 20W into the same load to get the same rate of rise, then the DUT has a COP of 2.  Is this not correct?

Yes I agree with you here.  The one caveat is that you have to start from the same initial temperature for both trial runs and you assume that you don't also have to warm up other components in the setup as Ion stated.  Personally I am less comfortable in recording the initial rate of change of temperature as compared to waiting five time constants to achieve thermal equilibrium.  I suppose it really depends on the the time constants you are dealing with and such.

My feeling at this point is that this whole thermal analysis topic has been given quite a bit of coverage from multiple perspectives.  The burden of choosing an appropriate thermal analysis method and documenting, recording, and presenting the data should fall on the claimant, and not the contributors to the forum.  If you do a good job then your audience shouldn't be picking apart what you have presented.

MileHigh