Overunity SSG - 1.23 COP! (Corrected again)

[John_Koorn]

Hey Aaron, those numbers look a lot better. Thanks heaps for posting the correction.

If you have time could you do the same test with the coil in repulsion mode? It would be great for people to see what the performance difference is between the two modes.

John K.

[Aaron Murakami]

Hey Aaron, those numbers look a lot better. Thanks heaps for posting the correction.

If you have time could you do the same test with the coil in repulsion mode? It would be great for people to see what the performance difference is between the two modes.

John K.

Hi John,

I'm anxious to do that when the kits are ready. I want to run the same machine and change out the wheels with a different magnetic configuration for comparison with an ABA test.

[Aaron Murakami]

This shows three meters showing the same thing - current sensing resistor is actually 0.65 to 0.7 ohms:

Certainly NOT 0.25 like I assumed just because it is written on the resistor.

[Aaron Murakami]

Ok everyone, am correcting the readings again.

Was brought up in a youtube post about considering the resistance of the leads themselves that could be added to the current sensing resistor.

I'm not going to consider anything that Cen Tech Harbor Freight meter says and will only rely on the Fluke's because of their quality.

So - the 123 Scopemeter showed me 0.65 ohms and the Fluke multimeter shows me 0.7. Averaging those shows me 0.675 ohms.

The leads of the Fluke multimeter have a resistance that bounces back and forth between 0.1 and 0.2 but hangs at 0.1 for about 75% but I'll just call it 0.15 ohms resistance to be conservative even though it is slightly less on average.

So if the direct current sensing resistor reading of that is 0.7 and we subtract lead resistance of 0.15 ohms that leaves 0.55 ohms as the actual reading of the current sensing resistor according to this meter.

The 123 Scope Meter accounting for using ground connected to the probe cable or the COM ground between both A & B ports, it comes out to be right at 0.5 ohms as the real reading and not 0.65 which doesn't account for lead resistance or other possible interferences.

So between both meters - lead resistance, etc... I'm going to pin the resistor at 0.55 + 0.5 / 2 = 0.525 ohms.

Using 0.525 ohms to calculate the input draw, here are the new numbers:

Time	Battery Volt	Resistor Volt	Resistor	Current Amps	Watts	Duty Cycle	Watt Sec/Sec	Watt Sec per Min	Running Time Min	Total Watt Sec
8:42	18.83	0.0347	0.525	0.066095238	1.2445733	100.00%	1.244573333	74.6744
8:50	18.62	0.0346	0.525	0.065904762	1.2271467	100.00%	1.227146667	73.6288
9:00	18.53	0.0346	0.525	0.065904762	1.2212152	100.00%	1.221215238	73.27291429
9:14	18.2	0.0344	0.525	0.06552381	1.1925333	100.00%	1.192533333	71.552	32	2289.664
To be conservative, we will use the starting voltage and current to calculate draw so there will be a higher number to beat AND will use the higher duty cycle at the end!
	18.83	0.0347	0.525	0.066095238	1.2445733	100.00%	1.244573333	74.6744	32	2389.5808
Estimate of actual averages
8:42~9:14	18.515	0.0346	0.525	0.065904762	1.2202267	100.00%	1.220226667	73.2136	32	2342.84

That means the draw is 2342.84 joules from the input.

The output still remains at 2883.96 joules - that is pretty dead on and won't change. I'm using the resistance of 3 X 100 ohm 10 watt power resistors and a current sensing resistor and I can tell you at that resistance at about 300 ohms plus or minus a couple ohms, that isn't going to change the reading at any significant level.

2883.96 / 2342.84 = COP 1.23
If estimating 25% in mechanical, then the estimated total COP is 1.54 but I'll leave it at 1.23 since I didn't measure the mechanical work of that wheel spinning at 4500 RPM.

So, forget what the resistor is rated at. Measure the resistance of your probes by making good contact with each other and keep them shorted. Whatever that is, subtract that from the reading you get when you measure the resistor because lead resistance will add to that resistor.

Keep in mind this is a one pass test and not multiple passes. I'll post something else later.

Any comments, questions or suggestions are welcome.

[Aaron Murakami]

One thought I have is to use a high enough resistance current sensing resistor so that any invariance in lead resistance, etc... will have no significant relevance. And, will be most practical for anyone that doesn't have a $300 precision micro-ohm meter, etc...

If the current sensing resistor is 10 ohms, a 0.1 or 0.2 ohm lead resistance won't make a difference that is big enough to take you out of the ballpark.

1.23 cop from 1.6 is about a 23% difference because of the lead difference factor I experienced.

With a 10 ohm resistor, a 0.1 or 0.2 ohm lead resistance difference is insignificant like 1-2% difference.

Using super small resistors for current sensing is so that you don't limit the overall current significantly to effect the machine's operation. So perhaps using the largest resistor you can but small enough to allow as much current as you're ever going to need may be a good idea. Using a larger resistor will simply make it easier, more accurate and more practical for the average builder who can use this method with any good quality multimeter to see exactly what the input vs output is.

[min2oly]

Some great thoughts Aaron.
I say this light heartedly and with no malice.... I only want to see everyone succeed.
this measuring stuff, when we really dive into the details, can tend to muddy the water under the bridge so to speak. Many have gone down that road only to be held to the spotlight and never come out quite right in the end.

your first post was on 12/10. today is 12/20
by now you could have completed at least 5-10 complete cycles using John K's spread sheet. Batteries themselves do not always tell the whole truth right-off-the-bat, put them through their paces and you can shake the truth out of them :-)
there might be more energy in there that is not being accounted for...
Kind regards,
Patrick

[Aaron Murakami]

Some great thoughts Aaron.
I say this light heartedly and with no malice.... I only want to see everyone succeed.
this measuring stuff, when we really dive into the details, can tend to muddy the water under the bridge so to speak. Many have gone down that road only to be held to the spotlight and never come out quite right in the end.

your first post was on 12/10. today is 12/20
by now you could have completed at least 5-10 complete cycles using John K's spread sheet. Batteries themselves do not always tell the whole truth right-off-the-bat, put them through their paces and you can shake the truth out of them :-)
there might be more energy in there that is not being accounted for...
Kind regards,
Patrick

Hi Patrick,

I'm only interested in laying out one single protocol that anyone can do with a common multimeter that has both voltage and resistance.

I could have done 5-10 cycles with this same method too in the same period of time. If you like that method, go for it. I'm personally not going to measure the energy it took to charge the input battery with a wall charger because I'm interested in just the input output of the SG/SSG in and of itself for the sake of proving a point. I'm not saying this to take away from that measuring method. I'm just sticking with what I've used for a long time. I accounted for all these things in the past but it has been almost 5 years since I really did any measurements like this with that method - so I'm taking a refresher course so to speak.

This method is extremely accurate - just need to plug in the right numbers. I have no worries about it. The only thing that beats it is the integrated detailed power analysis using a data dump of single waveforms - in my opinion.

Batteries do lie all the time. Can have the voltage but as we all know, that doesn't show what is actually in there. But measuring how many real joules are burned on the output shows exactly what the truth is regardless of how many cycles they've gone through - and this method definitely shows it.

This single pass test - one charge and discharge - is only a drop in the bucket. By rotating the batteries from back to front, COP's around 10.0 are possible. Not with this small SSG because it can't even push those 18v tool batts. I should be using 6v 1.7 ah batts with this model to be closer to what it is designed for.

I know a batt charged with spikes is not really compatible with a conventional charger, but I'm not so sure it is an absolute rule that they can't run an inductive load by being put on the front. 3 swaps and you're already over 2.5 COP for a decent build. I have no idea why this is not recognized very well, but it is no different than a bouncing ball where you add up all the Fd on each bounce upwards.

If you recover only 90% on each pass and swap back and forth and just keep adding the input work on each battery until they're down to the starting voltages, it is WAY "overunity." I know I'm not the only one who has done this.

[min2oly]

Hi Patrick,

I'm only interested in laying out one single protocol that anyone can do with a common multimeter that has both voltage and resistance.

Got it :-)
Thanks for sharing your experience with this, I look forward to reading more and replicating the experiment.
Kind regards,
Patrick

[forelle]

Hi Patrick,

I'm only interested in laying out one single protocol that anyone can do with a common multimeter that has both voltage and resistance.

I could have done 5-10 cycles with this same method too in the same period of time. If you like that method, go for it. I'm personally not going to measure the energy it took to charge the input battery with a wall charger because I'm interested in just the input output of the SG/SSG in and of itself for the sake of proving a point. I'm not saying this to take away from that measuring method. I'm just sticking with what I've used for a long time. I accounted for all these things in the past but it has been almost 5 years since I really did any measurements like this with that method - so I'm taking a refresher course so to speak.

This method is extremely accurate - just need to plug in the right numbers. I have no worries about it. The only thing that beats it is the integrated detailed power analysis using a data dump of single waveforms - in my opinion.

Batteries do lie all the time. Can have the voltage but as we all know, that doesn't show what is actually in there. But measuring how many real joules are burned on the output shows exactly what the truth is regardless of how many cycles they've gone through - and this method definitely shows it.

This single pass test - one charge and discharge - is only a drop in the bucket. By rotating the batteries from back to front, COP's around 10.0 are possible. Not with this small SSG because it can't even push those 18v tool batts. I should be using 6v 1.7 ah batts with this model to be closer to what it is designed for.

I know a batt charged with spikes is not really compatible with a conventional charger, but I'm not so sure it is an absolute rule that they can't run an inductive load by being put on the front. 3 swaps and you're already over 2.5 COP for a decent build. I have no idea why this is not recognized very well, but it is no different than a bouncing ball where you add up all the Fd on each bounce upwards.

If you recover only 90% on each pass and swap back and forth and just keep adding the input work on each battery until they're down to the starting voltages, it is WAY "overunity." I know I'm not the only one who has done this.

Hi Aaron,
please can you post a picture of the H-wave from this machine and tell where you put the probes.
Thanks

[Aaron Murakami]

Forelle,

The current sensing resistor is between the input battery negative and the negative going to the transistor.
The + of the probe is going to the side of the resistor pointing to the transistor.
The - of the probe is on the side of the resistor that is closest to the battery negative.

I'll have to take a pic of the waveform across that resistor later, but the waveforms aren't going to really be any different from any other SSG.

[min2oly]

Forelle,

The current sensing resistor is between the input battery negative and the negative going to the transistor.
The + of the probe is going to the side of the resistor pointing to the transistor.
The - of the probe is on the side of the resistor that is closest to the battery negative.

I'll have to take a pic of the waveform across that resistor later, but the waveforms aren't going to really be any different from any other SSG.

can i suggest a large CAP in parallel with the supply battery... and the current sensing resistor in between the battery and the cap?
this might help smooth out any anomalies, and the ckt can interact with the cap as the primary...
kind regards,
Patrick

[Aaron Murakami]

Another thought I had was why couldn't I figure out how many joules a cap holds at a given voltage

Please use this: Electronics 2000 | Capacitor Charge / Energy Calculator

[lis_wang]

Why did you use power unit as 2.614 watt seconds per second instead of 2.614 watt?







Hi everyone,

First of all, yes, I know the word "overunity" is an oxymoron - you can't have more than everything. But obviously this word has persisted for so long and it is known that it's intended meaning is for a device that is over 1.0 COP and that is why I used it.

This old dinosaur, my first ever SG from 12 years ago gives a 2.38 COP on the output battery and that doesn't even include mechanical work. That is just one recent test.

Circuit is MJl21194, both trigger and spike diodes are 1N4007 and base resistor is 60 ohms with a 1k 10 turn precision pot in series. Rotor is pink roller skate wheel from a $2 pair of roller skates from the good will with bearings in fair condition. I'm actually using neos - 3/8" thick double stacked neos...1 recessed into the wheel flush and the 2nd on top of that. Magnets are every 90 degrees and it is running in the enhanced mode described in Bedini SG - The Complete Beginner's Handbook. The power windings is 23 awg and trigger is 26 awg. The power winding is 4 ohms so that tells you how long it is. I built it originally with a MPS 8099 and followed the instructions in the original SG diagram posted on Keelynet WAY back.



For this particular test, I used different batteries than are shown in the pictures. In the pictures are 12v 7ah gel cells that were from an electric scooter. I used to charge those up with a bicycle wheel trifilar SG and used to drive it down to John Bedini's shop when I worked at a pulsed light healing device company down the street about 10 years ago or so. They're not in perfect shape but it is a miracle they're even half way good considering the agony and torment that I've put them through.

Anyway, for the COP test I did yesterday, I used 18v nicads from a Black & Decker Grasshog trimmer. They are rated at 1200 mah or 1.2 Ah.

Looks like this:



That one is actually a 2.0 Ah model, but the ones I have are 1.2 Ah that have been heavily used for the last 5 years. 90% of the time I charged it up, I used my 1AU Tesla Charger.

The input battery was fully charged up and so was the secondary battery. The secondary battery, I put a fixed C20 load (300 ohms) across the secondary to drain it overnight. I was going to stop it at 18.0 volts exactly but didn't catch it until it was at 17.41 volts.

I then disconnected it and hooked up the run battery to the SSG in the pics above and got it up to the fastest speed for the least draw. With a coil this small, you don't need the neon for protection and can run it without a secondary battery hooked up.

Anyway, after I had the base resistance where I liked it so it would be the fastest with the least draw and the lowest duty cycle, I hooked up the output battery.

I used a Fluke Scopemeter 123 for the test. It is 2 channels. I had a 0.25 ohm calibrated current sensing resistor on the ground line of the input battery and had both channels across that resistor.

Input battery went from 18.83 to 18.20 and I ran it with a battery on the back end for 32 minutes.

Before I give you the real numbers, I want to give you numbers that will handicap it to show the greatest draw - more than it actually drew so I have a bigger number to beat with the draw down test on the secondary battery.

I will use 18.83 volts (starting voltage and NOT average voltage) to calculate draw for entire running time. Voltage across the 0.25 ohm resistor was 0.0347 volts when the run started. So again, I'm using the HIGHEST numbers to show what it drew to be conservative. If I used lower numbers, it would be easier to beat so let's see what this shows us first.

0.0347 volts / 0.25 ohm resistor = 0.1388 amps of current. 0.1388 amps X 18.83 volts = 2.614 watt seconds per second. 2.614 watts X 25.3% (using the largest duty cycle towards the end, again to handicap the results to the max) = 0.6612418 joule seconds per second "burned" from the input X 60 seconds = 39.67 joule seconds per minute X 32 minutes of running time = 1269.58 joule seconds burned from the input.

The volt reading across the resistor was done using DC Mean instead of RMS since at these relatively low speeds it is accurate. If we're running in the mhz or something, then we'd definitely want to use RMS. The Frequency was about 300hz, which is 18000 cycles per minute divided by 4 magnets every 90 degrees = 4500 RPM at the start just to give you an idea of what the wheel is doing. Anyway, 300 cycles per second is very much in the slow range to use DC Mean on a scope to measure the voltage across the current sensing resistor.

At 32 minutes of run time, the output battery was disconnected and a C20 load was applied. 1.2 Ah C20 rate is a 60ma current draw. The battery voltage I used was 18.0 v / 0.06 amps = 300 ohm load. I used 3 X 100 ohm 10 watt power resistors with the 0.25 ohm current sensing resistor in series on the ground side of the string.

The starting voltage was 18.7 and it took 45 minutes to go down to 17.41 v where it was drained to before it was charged up.

AGAIN - to double handicap the numbers in favor of conservative numbers, I'm going to calculate the total draw using the voltage of 17.41 (when the battery was drained) so it will show that I drew the least amount from the output battery.

at 17.41 volts, the voltage across the resistor was 0.015 volts. 0.015 volts / 0.25 ohms = 0.06 amps. 0.06 amps X 17.41 volts = 1.0446 watt seconds per second and of course we leave it at that since the fixed resistive load is at a 100% duty cycle.

1.0446 watt seconds per second x 60 = 62.676 watt seconds per minute X 45 minutes until it hit the 17.41 volt goal = 2820.42 joule seconds burned from the output recovery battery, which was charged from the input.

2820.42 joules on the output battery / 1269.58 joules on the input battery = 2.22 COP and that does NOT include any mechanical work.

We used the highest possible numbers to show a large input and the lowest possible numbers on the recovery battery to show a small output - handicapping it in both directions for the benefit of the doubt.

Using the real averages, averaging the average of the different geometrical ramp downs on the voltage grahps, the average voltage was 18.515 with a voltage across the current sensing resistor of 0.0346 volts. 0.0346 / 0.25 ohms = 0.1384 amps X 18.515 = 2.562476 watts X 24.90 average % duty cycle = 0.638 watt seconds per second X 60 = 38.28 watt seconds per minute X 32 minutes = actual joules burned on input of 1225.07.

Using real averages for draw down test on output battery, considering actual averages of both geometrical ramp downs of the graph, 18.25v for 14 minutes at 0.0155 volts across resistor = 0.0155 / 0.25 = 0.062 amps X 18.25 v = 1.1315 watt seconds per second X 60 sec x 14 minutes = 950.46 joule seconds burned.

Then the second part of the graph average is 17.6 volts with 0.015 volts across 0.25 ohm resistor = current of 0.06 amps x 17.6 volts = 1.056 watt seconds per second x 60 seconds = 63.36 watt seconds per minute X 31 minutes at this average = 1964.16 joule seconds burned.

950.46 + 1964.16 joules burned on the output until battery got back down to 17.41 volts = 2914.62 actual joules burned on output battery.

2914.62 divided by 1225.07 = 2.38 COP and that still doesn't include any mechanical work added to that. With mechanical work, will be about 2.8.

Even with handicapping the input and output for the worst case scenario, the COP is still 2.22, but using the real averages, it is 2.38 COP and that is without any mechanical work added to the equation.

The output battery obviously will recharge itself a bit and some skeptics will grip about C20 being too low of a discharge. Obvoiusly if we use a C50 rate, we could probably wind up with a COP of 5+, but C20 IS realistic. And with these numbers, I could do a C10 rate and would probably still beat 1.0 COP easily.

Is my little SSG "overunity"?[/QUOTE]

[Aaron Murakami]

watt seconds per second

Why did you use power unit as 2.614 watt seconds per second instead of 2.614 watt?

Watts without time has no energy.

Watt second is the rate and then we have to multiply by time to know what the energy usage is.

[pontozero36]

Dear Aaron,

Great numbers!

The only thing i never understood about those COP figures and that's the only thing keeping me in the category "believer" from moving to the category the category "know for sure". (apart from the fact i never built one)

When you measure the resistance of the current sensing resistor it has to be 0.7 ohms +- an error (half the precision of the meter). When you measure volts the same thing. You're using an indirect current measurement. The error in this current measurement has to be propagated from the inicial ones through some formulas i don't remember now but should be in any experimental physics textbook. The error in the watts measurement again have to take in consideration the error for the bat voltage and the calculated propagated error for the current, using the same formulas. At the end, you should come with something like COP = 1.6 +- some error. If that calculated error is, for instance, greater or equal to 0.6, then it's, at a first glance, inconclusive "overunity". Anyway, never knew why anyone hasn't bring this up before.

As i see the sg, anyway, any conclusive battery COP > 0,5 calculated the way you did (and having taken the mentioned error calculations in consideration) already beats conventional science and that is mindbogling for me already! Just want to have better arguments to fight the skeptics...

I'll see if i can get this through when i build one myself, wich i'm waiting a better time and gathering funds. Hope i will manage to do it in 2014.

Once again, nice job and thanks for all the work you've been doing for the cause!