Rediscovering Zaev’s ferro-kessor

[Vasik041]

Will be manipulating the gap with some magnets, not sure how to find out if the core is half saturated.

Hi Itsu,

I described the procedure here
http://www.overunityresearch.com/index.php?topic=3453.msg65653#msg65653

Regards,
-V.

[Vasik041]

I'm sure Vasik will have more to add to this but IMO you could use the formulae found at this link-
https://www.supermagnete.de/eng/faq/How-do-you-calculate-the-magnetic-flux-density

In my experience theoretical calculations for magnetic circuits are difficult and give not satisfactory results.
There are many factors which contribute to the issue.
For example small air gap make big difference, but you can't measure it precisely.

Regards,
-V.

[partzman]

Hi Vasik and Itsu,

In my attempts to find different PM/core combos I have discovered a source of measurement error in my previous tests using my circuit as posted. 

The error lies with the slow turn off time of the P-channel mosfet M2 in the "Vasik Test Schematic".  The turn off delay between M1 and M2 is only ~130ns but when the cursors are correctly lined up on the input measurements to account for this, the overall COP<1 in all the tests I'm currently running!  This error was also responsible for the apparent gains at the higher supply voltages due to the reduced input measurement periods which amplified the error.

I have created a gapped (.010") core arrangement wherein a PM is placed between two vertical ferrite poles and this assembly is placed across the core gap.  A separate ferrite pole is placed across the top of the two vertical poles that allows tuning of the amount of PM flux reaching the core.  See pix below.  In this manner the % of coil/PM flux can be set and optimized experimentally.  With this arrangement, I still can not achieve COP>1 with the above error correction.

Regards,
Pm

[Vasik041]

Hi Vasik and Itsu,

In my attempts to find different PM/core combos I have discovered a source of measurement error in my previous tests using my circuit as posted. 

The error lies with the slow turn off time of the P-channel mosfet M2 in the "Vasik Test Schematic".  The turn off delay between M1 and M2 is only ~130ns but when the cursors are correctly lined up on the input measurements to account for this, the overall COP<1 in all the tests I'm currently running!  This error was also responsible for the apparent gains at the higher supply voltages due to the reduced input measurement periods which amplified the error.

I have created a gapped (.010") core arrangement wherein a PM is placed between two vertical ferrite poles and this assembly is placed across the core gap.  A separate ferrite pole is placed across the top of the two vertical poles that allows tuning of the amount of PM flux reaching the core.  See pix below.  In this manner the % of coil/PM flux can be set and optimized experimentally.  With this arrangement, I still can not achieve COP>1 with the above error correction.

Regards,
Pm

Hi Partzman,

It's good that you found an issue with measurements.
Have you tried measuring BH curve ? Can you see offset created by permanent magnet ?
For me such core setups never worked, I guess because magnet too "far away" and its field get too dispersed...

Regards,
-V.

Edit: I tried draw how it supposed to be, please see attached picture.

[TinselKoala]

Hi Vasik and Itsu,

In my attempts to find different PM/core combos I have discovered a source of measurement error in my previous tests using my circuit as posted. 

The error lies with the slow turn off time of the P-channel mosfet M2 in the "Vasik Test Schematic".  The turn off delay between M1 and M2 is only ~130ns but when the cursors are correctly lined up on the input measurements to account for this, the overall COP<1 in all the tests I'm currently running!  This error was also responsible for the apparent gains at the higher supply voltages due to the reduced input measurement periods which amplified the error.

I have created a gapped (.010") core arrangement wherein a PM is placed between two vertical ferrite poles and this assembly is placed across the core gap.  A separate ferrite pole is placed across the top of the two vertical poles that allows tuning of the amount of PM flux reaching the core.  See pix below.  In this manner the % of coil/PM flux can be set and optimized experimentally.  With this arrangement, I still can not achieve COP>1 with the above error correction.

Regards,
Pm

Nice work PM. Carry on!
Timing issues plague efforts to make good power measurements in the kinds of devices that interest us. You've identified a timing issue related to the circuitry, but when I read this I also thought of timing issues related to scope probes, specifically active probes like differential voltage probes and current probes. As you know these will always have some lag time or latency in what they send to the scope, when compared to ordinary passive voltage probes. This is referred to in the literature as "probe skew". Most modern scopes will have some way, either automated or manual, to "de-skew" channel timing measurements to account for differences in latency between different types of probes connected to different channels of the scope. Probe latency usually won't be as much as 130 ns though! But it can be enough to screw up power calculations involving probes of different latencies.
Even the Rigol 1054z scopes have a manual setting to delay channels for de-skewing purposes. So if we are using an active current probe with, say, 30 ns latency in CH1 and a passive voltage probe with essentially zero latency in CH2, to get accurate power computations we would need to delay CH2 by that same 30 ns in order to synchronize the probes.
(This was one of the errors that Steorn made in their early Orbo magnetic motor testing; they failed to de-skew their active differential and current probe channels to have them properly synchronized.)

[partzman]

Hi Partzman,

It's good that you found an issue with measurements.
Have you tried measuring BH curve ? Can you see offset created by permanent magnet ?
For me such core setups never worked, I guess because magnet too "far away" and its field get too dispersed...

Regards,
-V.

Edit: I tried draw how it supposed to be, please see attached picture.

Hi Vasik,

I have not yet done a BH curve measurement with this arrangement but was rather using the change in inductance for determining the PM flux offset effects on the core.  With the advanced math on the scope, I can produce an integrated flux waveform that I would like to use as a source for the BH curve rather than external analog components, but I haven't found how the scope will allow this at the moment.  It seems to be fixed on channels 3 and 4 for the XY display.

Regards,
Pm

[partzman]

Nice work PM. Carry on!
Timing issues plague efforts to make good power measurements in the kinds of devices that interest us. You've identified a timing issue related to the circuitry, but when I read this I also thought of timing issues related to scope probes, specifically active probes like differential voltage probes and current probes. As you know these will always have some lag time or latency in what they send to the scope, when compared to ordinary passive voltage probes. This is referred to in the literature as "probe skew". Most modern scopes will have some way, either automated or manual, to "de-skew" channel timing measurements to account for differences in latency between different types of probes connected to different channels of the scope. Probe latency usually won't be as much as 130 ns though! But it can be enough to screw up power calculations involving probes of different latencies.
Even the Rigol 1054z scopes have a manual setting to delay channels for de-skewing purposes. So if we are using an active current probe with, say, 30 ns latency in CH1 and a passive voltage probe with essentially zero latency in CH2, to get accurate power computations we would need to delay CH2 by that same 30 ns in order to synchronize the probes.
(This was one of the errors that Steorn made in their early Orbo magnetic motor testing; they failed to de-skew their active differential and current probe channels to have them properly synchronized.)

Hi TK,

Yes, I have been victim of probe de-skew many times in the past as you well know.  In my case, it has been the current probes that are the biggest offender.  They can only be trusted with high frequencies or fast rise and fall time waveforms if checked against a good non-inductive CSR at the frequency of interest.

Tek has automatic de-skewing of probes it recognizes but the fact is, each probe is slightly different and each channel can also be slightly different from the rest, so to really get down to optimum accuracy, it's up to the operator to be sure all are synchronized as you point out.

Regards,
Pm 

[TinMan]

Hi TK,

Yes, I have been victim of probe de-skew many times in the past as you well know.  In my case, it has been the current probes that are the biggest offender.  They can only be trusted with high frequencies or fast rise and fall time waveforms if checked against a good non-inductive CSR at the frequency of interest.

Tek has automatic de-skewing of probes it recognizes but the fact is, each probe is slightly different and each channel can also be slightly different from the rest, so to really get down to optimum accuracy, it's up to the operator to be sure all are synchronized as you point out.

Regards,
Pm

Could you not just place all your scope probes on the same point of the DUT-say like the CVR,and make sure all of the scopes waveforms from each channel  are in phase ?


Brad

[TinselKoala]

Could you not just place all your scope probes on the same point of the DUT-say like the CVR,and make sure all of the scopes waveforms from each channel  are in phase ?


Brad

Usually, yes. In any case a reading from a current probe should be checked against a reading from a CVR if possible, for timing, frequency response, amplitude of response and even _direction_ of response, as it is easy to get the probe the wrong way round on the conductor and not realize it.

[partzman]

Could you not just place all your scope probes on the same point of the DUT-say like the CVR,and make sure all of the scopes waveforms from each channel  are in phase ?


Brad

Yes. 

Pm

[Vasik041]

Hello,

I got some free time and decided update my Arduino Due based core tracer. (New software can be found here https://github.com/vasik041/openlab/tree/master/ferd/core_tracer_v2).
While experimenting I found an error in calculations. I did not take into account gain in ADC channel which measure current.
After fixing the error I do not see COP > 100% any more It is a little sad, but good that one more mystery solved.
I don't mean that overall idea of device is wrong, but why it was strange that I can't loop it while measurements show big COP.

I think now anyone still interested in getting OU from ferrite + PM setup should concentrate on negative resistance setup.
It still showing some gain.

My apology to everyone who spent time trying replicate my bogus claims.

Regards,
-V.

[partzman]

Hello,

I got some free time and decided update my Arduino Due based core tracer. (New software can be found here https://github.com/vasik041/openlab/tree/master/ferd/core_tracer_v2).
While experimenting I found an error in calculations. I did not take into account gain in ADC channel which measure current.
After fixing the error I do not see COP > 100% any more It is a little sad, but good that one more mystery solved.
I don't mean that overall idea of device is wrong, but why it was strange that I can't loop it while measurements show big COP.

I think now anyone still interested in getting OU from ferrite + PM setup should concentrate on negative resistance setup.
It still showing some gain.

My apology to everyone who spent time trying replicate my bogus claims.

Regards,
-V.

Vasik,

No apology needed as far as I'm concerned!  I have been in this position myself and know how discouraging it can be, but it is just a part of the cost of searching for answers to solve our energy problems.

I am still researching possible OU from ferrite and PMs so I plan to download your software for the Arduino and give it a try.

Regards,
Pm

[Vasik041]

Simple experiment - flyback and magnet

Graphs show power supply current vs time.
Driver pulse duration 100us. Secondary coil connected to power supply through diode.

blue - coil's field aligned with magnet's field
yellow - no magnet
red - coil's field anti-aligned with magnet

Those small peaks at the beginning are very interesting

[ion]

Those small peaks at the beginning are very interesting

Hi Vasik

Indeed interesting. Wondering what happens if the diode is removed and just a resistive load is placed on the secondary?

Hard to account for the blue reversal unless the experiment was altered in some way e.g. reversed coil leads for blue test.

What do you think is causing this artifact?

Regards

[Vasik041]

Hi Vasik
Indeed interesting. Wondering what happens if the diode is removed and just a resistive load is placed on the secondary?
Hard to account for the blue reversal unless the experiment was altered in some way e.g. reversed coil leads for blue test.
What do you think is causing this artifact?
Regards

Hi Ion,

In these three tests the only difference is magnet presence and orientation.
I guess those peaks are "negative resistance" effect caused increased by magnet.
This need to be investigated further.

I am using "looped flyback" setup. It is very nice - I can see magnetization and demagnetization separately and power balance without complex measurements. Removing diode will complicate things.

Regards,
-V.

[ion]

Hi Ion,

In these three tests the only difference is magnet presence and orientation.
I guess those peaks are "negative resistance" effect caused increased by magnet.
This need to be investigated further.

I am using "looped flyback" setup. It is very nice - I can see magnetization and demagnetization separately and power balance without complex measurements. Removing diode will complicate things.

Regards,
-V.

Thanks for supplying the schematic. I get what you are saying about ease of measurement with the diode supply replenishment.

I was just wondering if the artifact went away when the diode was removed (diode capacitance?) I understand that it's removal would complicate power measurement, but power measurement can be a separate issue.

It is very good that you just flipped the magnet rather than swapping the leads to the coil, as swapping coil leads may have altered the stray capacitance coupling.

It will be interesting to see what you discover with further testing.

Thanks for sharing your work.

Regards
ion

[Vasik041]

Effect of magnet on BH curve

green - no magnet, red - coil's field anti-aligned with magnet, blue - coil's field co-aligned with magnet's field

Edit:

P.S. This is good illustration to avoid common misconception - field can be "neutralized".
No. Fields are independent in linear media.

[Vasik041]

Collection of BH curves obtained with curve tracer 

[Vasik041]

Tracer upgrade for shorting coils

[Vasik041]

Recent advances in the area: CW BH loop and "extra" output current

[partzman]

Recent advances in the area: CW BH loop and "extra" output current

Wow, this is interesting!  I and others I'm sure, would like to see the transformer/coil assembly you used to accomplish this. 

Regards,
Pm

[Vasik041]

This particular traces were obtained with N30 ring core and small ferrite magnet.
Core was sawn to allow magnet freely insertion.
Similar results obtained with ferrite rods (e.g. from magnet antenna).
In opposite to my "common sense" it appeared that open magnetic circuit is important

Regards,
V.

Edit: Other important factor is core thickness, if one take too thick core it will be very difficult to saturate it.
To see full BH curve in my case current about 20A-40A is required. If you take smaller core 5-10A will be enough.

[Vasik041]

I made some improvements to my negative resistance experiment.
It is easier observe NR if we add several impulses together.
He is a simple test schematic I use and example scope traces from it.
Effect is quite small, about 0.4milliW, but quite reliably reproducible now.

[partzman]

Recent advances in the area: CW BH loop and "extra" output current

Vasik,

I'm not sure if anybody is paying attention here but your clockwise BH curve indicates a power gain not loss.  Have you tried more turns on the split toroid to lower the large H field and if so, did you still see a CW BH curve?

Regards,
Pm

[Vasik041]

Have you tried more turns on the split toroid to lower the large H field and if so, did you still see a CW BH curve?

Hi Pm,

I tried many different setups. This CW BH curve presents mystery which bothering me for many years.
I still haven't decided whether it is real or not. One one side even people who not aware about FE produce
curves like this (see Fig-4.jpg attached). It is not fully CW BH but have two CW regions.
On the other side, I got these fully CW BH curves and it should give energy gain. But I can't see gain directly in flyback like setup. So, may be my measurements flawed? I build three different traces and all of them gives some CW BH curves. I even found old C. Steinmentz paper presenting such curves (pic.png). Also thermodynamics suggests that it should be possible. So, it should be real ?
To obtain such curve, core need to be driven into deep saturation. This requires currents like 10 or 30 amps.
This present significant technical difficulty. If we use longer coil, we got significant loses on a wire resistance and issues with parasite capacitance. Another issue is that curve is non linear, and non-linearity is different during magnetization and demagnetization, so analog integrator gives significant error.
This is interesting research topic but unfortunately my resources are limited, so I don't have all answers (yet).

Regards,
Vasik