FanGen - HV Generator from 6V

[Slider2732]

Where to begin ?
Some of you guys may well see this device and notice that many elements of today's experimental methods are incorporated within it - that is the idea.
It's a high voltage generator, from a 6V, low current power supply.
The name is derived from 'fan' and 'generator'. I believe 'fangen' is a slang term for awesome, cool, great and that it means collect or catch in German.

After reading up on and viewing YouTube videos of Ed Leedskalnins' Coral Castle work, I linked that with Imhoteps' Bedini PC fan, my own White Crow pulse motor ideas and many other techniques, that could be implemented quite easily at zero build cost.
The device meets the intended aim of delivering HV from a small power source and is to go forward with wireless energy experiments, single wire rectifying, magnetic resonance work and much more.

It's a PC 12V fan, with drive electronics remaining to power it from 6V, with which it revs to a good usable speed due to inertia from the magnets and no fan blades.
6x neodymium magnets surround the fan rotor and, because the internal weak magnet material has been left in situ, enables the magnets to be attached and moved readily...they just clip to the sides and remain in those places under power spins. Once the correct placements have been ascertained then standard superglue will hold them in place.
There is a 24V relay inserted into one side of the fan casing, this is the fired coil.
A Hall sensor (DN6851) was taken from another fan and forms a Hall sensor based coil collapser. Collapsing the coil just before a magnet reaches it was part of RomeroUK's work. In this case, a 2SC3200 very standard NPN transistor handles the switching output of the Hall sensor and shorts the coil at the correct moment. Simply holding the sensor to the side of the casing allows finding the best spot...superglue can again be used to hold that position.
There is a small neodymium magnet at the rear of the coil.
Operation below and above 6V shows less efficiency. At voltages lower, the circuit runs erratically or not at all. At voltages above, there appears to be no significant extra HV output.
The magnets are derived from old CD-Rom drives, there being 2x neo's in the laser assemblies.
Current used by 12V PC fans is typically approximately 130mA at 12V. I would welcome a replication that could show the actual figures of DC power used and AC power output.

Performance details:
Without the coil collapsing, the output is barely strong enough to light an LED to medium brightness,
Without the coil magnet on the back, the output is insufficient even with coil shorting, to light the neon bulb.
With the coil collapse effect and magnet on the rear of the coil, all heck breaks loose
HV generation occurs at the merest RPM, right from the beginnings of device spin up.
The magnets are arranged NSNSNS. If all face the same direction, they halt the operation of the motor.
Output, from my lackluster Sperry SP-6A meter, appears to be 8V AC at 25ma, yet it lights the neon and there is a very noticeable kick of AC if a finger shorts the coil output.
The Hall sensor trigger indicator LED also performs as BEMF guard for the 2SC3200 transistor Collector output, for the returning spike of HV.
A video uploaded yesterday showed a bizarre effect of a heartbeat effect. That is something that Ed Leedskalnin replication machines seem to have a trait of. I fully presume the heatbeat within the video yesterday to be a low voltage condition...but this project did indeed go on to run very well and to run as intended..be born and grow as it were. The LED/Hall sensor section and the fan were two seperate circuits at the time, connected only via their shared power leads...yet the heartbeat oddity came from both.

I have no schematic, but do hope that the FanGen fires up some peoples imaginations at least...and would aim to produce a schematic. However, i'm historically poor at such things and would welcome those of experience to draw one up, if the circuit and effects are deemed worthy.


Attached are a couple of pics. One shows it generating HV and the second is of the rotor and magnet placings..
Here is a video of this first FanGen in operation.

YouTube - ‪FanGen - HV generator‬‏

[Slider2732]

Here are some output figures, replacing the neon bulb with 4x1N4001 diodes as bridge rectifier and a 2200uF 25V electrolytic capacitor.
Power input is from a Universal AC/DC Adapter, bridge rectified with a KBJ406G and 4700uF 35V electrolytic capacitor.

6V input - 3.9V 26mA
7V input - 4.4V 42mA
8V input - 5.6V 55mA
9.7V input - 8V 70mA

I would welcome any advice on extracting especially more milliamps from the rectification

[Slider2732]

Just an update on this, while it seems to bear relevance to the RomeroUK/Muller replications.
How can a generator system increase its speed automatically, when a load is introduced ?

My intention from this FanGen learning system, has been to remove the drive elements effects of such a 'BEMF' collection system from the energy collection section. Thus, the original drive circuit of the fan was retained.
That allowed me to mount the Hall sensor and 24V relay coil as a separate circuit.

On charging a very dead 12V battery earlier, I decided to see what would happen if a 12V bulb were used as an extra load. The rotor increased in speed
Also, if the battery was disconnected, the rotor slowed and the neon lit, upon reconnection the rotor sped up and the neon would extinguish again.
My theory, is that while the motor control circuit cannot be a factor, the interactive relationship of the shorted coil to the magnetic flux certainly can be. More load = changes in magnetic flux, described as a lessening of combative forces, producing a speed up of the rotor.

Here's the short demo video:
YouTube - ‪Motor speed rise with load‬‏

[Farmhand]

Hi Slider, Cool demonstration, I think what is happenning there is the BEMF is holding the fan motor speed back, then when you connect the light bulb the LED' dims because the BEMF is removed more fully which allows the rotor to spin faster.

THe input current to the fan should rise a bit as it speeds up.

The same thing happens with an SSG if the load is removed or the load impedance increased less BEMF is removed so the rotor slows, then when the load is replaced or increased the rotor speeds up and the input power to the drive coils increases.

Getting the BEMF looped to run the motor is the trick.

One thing to remember with looping the BEMF or CEMF is that it must be returned to a higher voltage then regulated down to supply voltage.

Attempting to loop the BEMF back directly to the supply voltage will cause all the BEMF to be removed and A destruction loop is created which will destroy the device if it is powerfull enough.

This is why the recovery battery must be in series with the run battery, swap the 12 volt battery for a six volt one then try charging 24 volts this will show the effect if it is happening how I envision it.

Try it and see.

Can you make a rough drawing or sketch so we can see what is happening ?

Cheers

[Slider2732]

Thanks Farmhand, very useful information.
Tests are severely limited at the moment by a lack of batteries and test gear. I have 1x 6V lead acid and 2x 12V car battery, along with the salvaged 12V puffed out one shown in the latest video. All were formerly junk. Only working meter is a Sperry SP-6A analog needle type.
The 6V was rejuvenated by a pulse motor (my original White Crow looper that has a record of 21 minutes from 3x 1F supercaps. Long story, but after successfully charging the battery on its own it was removed from the work table, where it had been simply hotglued on to and has taken til yesterday to even partially retune !)). The 12V was partially rejuvenated on the same circuit. but current is always a problem, as the circuit is only diddly and draws approx 10mA. Am about 2.49A on the wrong side it seems for a proper recharge lol
I do try to video any of everything possibly useful though and am using all this to learn....been doing this only a couple of weeks and really value such replies.

The input current rising is something to test. At least I can do that

Circuit wise, i'll draw one up certainly. May not have time til this evening, but here's what happens.
It's based on a standard 12V fan with the blades removed and 6 small neo's fitted to the rotor. Then a 24V relay coil sits within the outer case and a Hall sensor is at its side where the effect of BEMF is greatest. The Hall triggers the base of a 2SC3200 NPN transistor, which sends a negative spike to the coil. The other side of the coil is connected to positive of the supply. You may expect that to be the negative, to short the coil, but i'm finding a bigger neon lighting ability on this one by connection to the positive. An LED lights to show firings from the transistor and also limits the BEMF returning to possibly destroy it. As the Hall next switches, a magnet passes and dumps the energy through the coil, lighting the neon. A couple of small neos sit behind the coil and greatly assist neon lighting.
With any load attached the neon doesn't light.

I originally wrote that the circuit only works properly at 6V. That's been found to be incorrect. It will work at higher voltages, if the rotor is given time to speed up between increments from the wall adapter 3V-12V source.

[Farmhand]

Ok I think I see, it's cool little setup. I meant to say swap the charge battery to a 6v or a 24volt one if the one you are charging now is 12volts that should show a difference both ways, the 6 volt one should cause more input current and higher rotor speed than the 12v and the 24 volt should be slowest with least current draw. The lower impedance is more like a short and a higher impedance is more like an open circuit. There is a sweet spot or an impedance match which should give best power transfer.

If I understand correctly you are just switching the generator coil ? Yeah if you are switching the generator coil for BEMF spike's then the negative of the gen coil should go to the positive of the charge battery. The rotor magnets are charging the coil. And when the coil is switched the BEMF disapears if the load is low enough impedance, This is shorting the generator coil. If you can switch it at the top of the generation wave you should see the best effect. You should have a diode from the coil to the battery aswell I think.

Anyway it is a very interesting effect. Good job, seems like that could be very useful.

Cheers

[Slider2732]

Thanks for your notes again
Whole thing really, for this and for the White Crow type, is a free source of learning...something solid and readily available to build from household bits, that gives the same results time after time. The original circuit that you might remember I thought I was having troubles with, charged the 6V batt by itself and so i'm chasing self running again...a very odd situation to be in !
Latest on that original circuit, is that a 3 magnet rotor setup has just given a 7m 41sec run from a 1F cap and while a white LED was connected as the load.
Yes, if the impedance matches, the load becomes a non issue, even on these sort of pulse motors.
However, all is about trial, move a magnet a bit, trial again.
The FanGen is a 6 magnet rotor, which I maybe erroneously thought would be 6x power per revolution, taking friction into account. However, off the other circuit running so well on 3 magnets, the dwell angle (or some other automotive term) seems more important..time for the coil to collapse at speed and then energy be drawn from it. The 10 spikes per pulse Arduino setup I saw on YouTube earlier is something to really think about. Perhaps a 555 simply kept at a frequency dependent on rotor speed, then switched in and out with the Hall sensor.

Quick question - if I have a 7min 41 sec run from a 1F supercap charged to 4.5V, with a few mA load of the LED through a 3K resistor, what's the simple equation for efficiency ?
Yep, that's how fresh I am at this


Got to get 8 kittens ready for spaying collection and then i'll draw a diagram of the FanGen circuit. I was kindly pointed to a Freeware diagram package but would need several days to learn it.

[Slider2732]

Perhaps I ought to learn Eagle !
Rubbish pic attached, but it shows the simple FanGen circuit

Running voltage is 6V to 12V. The LED seems fine, doesn't over brighten and the 2SC3200 transistor has now run for a few hours of tests. A 1K resistor was on the positive rail for the DN6851 Hall sensor, but seems unneeded. Similarly, a 100ohm resistor in series with the LED was thought to be limiting needlessly and so now is also omitted.

[Slider2732]

A few things were trialed today.
The transistor was swapped to a more powerful 2SD667, resulting in no effect...quite surprisingly everything ran exactly as before. I had wondered about increased Collector current and yet nothing different was seen.

Solid state thoughts moved along a little, with the noting that at a certain spot, when a magnet pole flips to the opposite pole the neon would light. Movement only has to be around 10 or so degrees of an otherwise stationary rotor.
The solid state idea, is of a coil that has a biasing magnet on one end. An oscillator powers up the coil to create an opposite magnetic force, then a seperate transistor fires and shorts the coil momentarily. The collapsing field produces the HV spike, without anything physically moving anywhere.
If a 555 timer circuit were used to control the oscillations, the duty cycle would be where to tune.

Talk in the RomeroUK/Muller replications thread has recently turned to HV caps having benefits in the coil area. Various methods and various sizes of AC caps were tried out. The best results have been had with a duo, a 0,22uF @ 100V and a 0.1uF @ 250V.
In fact, the 0.32uF total is only 0.01uF different to Romero's machine.
The effect, has been to both brighten the output (more voltage) and to light both sides now of the neon. Previously just 1 side of the neon bulb was lit.

Here's a video to show the findings:

YouTube - ‪FanGen - HV cap addition‬‏

[Farmhand]

Hi slider I was confused before, I thought you were still using the pulse motor to drive it and collecting the bemf there too. Some of what I said won't apply but the load on the driving motor should reduce if the gen coils are harvested for BEMF instead of normal generation. Cool setup for testing this stuff out, nice and cheap and easy to change.

Have you seen my mini three pole video, I made a three pole Bedini energiser from a fan rotor and chassis.
YouTube - ‪AlternateFarmhand1's Channel‬‏
The coloured led's are lit from a pickup winding (extra trigger winding). The white Led's are lit from the generator coil. It can do 12 000 rpm with 30 volt's input, when the bearing were new it could. haha

Those are nice neon flashes you get just moving the rotor. It does make sense the higher voltage caps collecting the spike's better, a high frequency cap should help too because of the fleeting nature of the spike.

Cheers

[Slider2732]

From around 3 minutes in your video, there's the unmistakable note of some builds and some motors that is as pleasing to the ear as perhaps a Rolls Royce Merlin engine...just something about all the coils and the rotor speed doing things in a harmony of sound.
Well, subscribed to your channel anyway, somehow i'd missed it !

Sorry for the confusion, my fault entirely. I'm focusing on this one for Muller type experiments, but referenced another build for efficiency results.

Got a strange video uploading to YouTube at the moment. It's of the FanGen with a load test. The rotor isn't spinning, yet the load fan starts to after some time.
Several caps are on the output, through a FWBR. Then a 12V PC fan after that. The FanGen rotor wobbles, reminiscent of the Ed Leedskalnin machine heartbeat way, constantly firing and refiring the Hall sensor, giving HV output. The wobble is lack of rotor drive power and subsequent Hall firing , that's what's been realised for the heartbeat thing.
What happens, is that the rectified output is not quite enough to start the output fan running, but, each time that fan nearly starts and wobbles to s stop it allows the caps to charge a little more.
After about 1min 45secs in the video, the cap charges are enough to finally get the motor spinning. Meanwhile, the FanGen rotor remains pulsing away as before.
Quite mesmerizing as a little effect and perhaps just shows that builds ups of charge over time are dynamic and not simple on/off switches.

Here's the video:

YouTube - ‪FanGen - cap charge start up of external fan‬‏

[Slider2732]

After scrabbling to find other batteries to test, it occurred to me that old laptop battery packs might be a good idea.
I found a set, from within a mid 1990's 120MHz laptop (oh yes, THAT old) and took the casing off. A few of the batteries were noticeably sulphating themselves on the ends, but 4 of them seemed ok, outwardly.
So, they have been trialed.

Ni-MH 1.2V per cell, reading 0V and 0mA

The connections from the HV output of the FanGen have been:
Negative of the batteries to the left side of the coil, which also has the transistor pulse connection.
Positive of the batteries to the right side of the coil, through a 1N4007 diode.

Initially, the rotor hardly turned at 6V, but slowly over a few minutes the speed increased, the whole thing gathered speed.
Charge rate was initially quite quick to 2.2V and then slowed off some. After about 25 minutes things changed to a good charge rate, as seen in the video. After approx an hour, the pack was charged and WORKED
It didn't take more than 5 minutes though to reflatten them to below 3V with a small motor and so they went back on charge. This time, recovery to around 4.6V only took about 15 minutes, with the rest of the time of 1 hour spent crawling up to and past the 4.8V capacity.
Power has been better this time and the resting voltage is higher by approx 1 volt after 30 minutes.
It would seem that cycling these a few times should see them return to useful service.

Very very pleased

YouTube - ‪FanGen - NiMH rejuv + charge up‬‏