ok john,

Here's a video of the small light, hi/lo, before I tried coil.

Pictures by Thugugly - Photobucket.

I know this is not much but it's as far as I have gone. The cap across the 8-5 pin of opto was .1uf/50v.

I hoped it would work, as I won't be home for 1-2 weeks, I can't wait till I'm back, i'm sure john you will have it working, and I can save some money

Peace and good luck JohnStone

Thug

Thanks for tip kEhYo77, This is same behavior I got when I tried the ir2101, thought I would give it one last try with opto's, still have a few left, and two dc/dc converters, so I'm not quite done yet, Go JohnStone Go.

So far the best hi/lo driver was the homemade n-fet, driving the hi pfets, it worked for 24 v, vnice, but i can't get these factory drivers to switch 24v without frying something.

BTW, Nice setup.

I would love to know more about your shorting technique, you should start another thread, sounds really interesting.

Peace

Thug

Maybe we should be using a higher coil resistance to drive hi/lo. I measured the incandesent light (il) resistance (5.7 ohms), will unhook the bifilar (1.2 ohm), and use a (2.4 ohm) coil, and add (3 parallel,10 ohms resistance) to bring it up to the (il) resistance, and try once more before i go. I could at least have a few more converters for when i get home... I'm sure JohnStone or Dana it have it working by then.

I think i will use a higher gate resistance as well, 2.2 k, Larry mentioned that he thought it might need to be higher as well, I think. It worked with the homemade shifter.

But not tonight

Thug

G'Day Larry,Thugugly
I have loaded some detailed pics in my PhotoBucker here for a clearer view

Sun 8 July 2012 pictures by Kogs1 - Photobucket

In the above pic it is the same one I used but was an old pic where I had left out 2 jumpers I had included the jumpers in the cuit under test
@ Larry
There is only one pot in my circuit in the photo at the bottom there are 2 blue wires they go to each side of the 50k pot the brown one goes to the wiper.
The green wire at the bottom is running under the board to connected to the earth bus that runs along at the top of the picture along the underside where all the earths are it connects to the regulator earth.
At the left the red wire comes from the regulator and the brown wire goes to the input of the Mosfets.

Larry If I leave this oscillator out of the Machine and remove the coil and large Diodes from the circuit and replace the CLF with an 36v bulb then connect the output from the regulator direct to the input of the mosfets the 36v bulb lights up bright

Please Larry in your circuit LM339 /4 comparitor it shows R5 as 38k the only resistors I can get here are 38k3 would these be OK

@ Thugugly
The pic above is my circuit exactly except it does not show two blue jumpers though they were in the circuit when tested . The copper wires are woven through the board to each component and where there is a line where you do not see the wire as it would connect with/crossover the other wire it actually runs under the board I have tried 3 different 555 timers

I do appreciate Your help
Regards

G'Day AmpsVolts
Try here
Super Speed 9-18VDC Hobby Motor | eBay

Regards

Thanks Ian for your consideration,
Checked the link out but others beat me to it.. will keep looking
Cheers - Waiting for the next UFO installment

Found the motors in the UK at Miscellaneous components

Ian I thought you said your oscillator worked except when you connected it to the FET board.

As I tried to explain to you before... if you connected the you circuit the way you explained here (of course I don't know if by the input of the FET you mean the gate or the drain), you should not have gotten the light to light. You need something to turn on the FET. The input of the FET is the drain and the output of the FET is the source. If you did hook it to the gate, did you try to disconnect it to see if it would turn off the bulb. You should be able to turn the bulb off and on by connecting and disconnecting the regulator. Try this and let me know.

Yes that resistor would work... the original resistor was 45K.

Regards
Larry

Hi Thug,
I am so sorry you smoked silicon because of my circuit being still under contruction. Thanks for daring to pioneer! As Larry mentioned some protection diodes might be necessary. I plan to investigate this matter by measuring the voltage spikes on all terminals with 12V and a resistor and coil in series as load. Tender approach is a MUST if you love your circuit.

@ALL: Some know how regarding increased gate resistance:
A gate can be driven via a higher gate resistance at low frequency and low load - only. The restrictions above are to be obeyed! Why?
Different from transistors our FETs make use of the whole silicon chip if full opened only (full conductive DS-channel). Below full conduction - the usable conduction channel is being tied up by the electric field of the gate (imagine a flexible hourglass being choked at the constriction area). We get a higher resistance and additionally the increased heat losses focus an the choked area only. We obsereve a multiple positive feedback! Different from transistors the conductive channel increases its resistance at increased temperature. (Different from transistors FETs are happy if operating in parallel!)
If we drive the rated DS current (IDS) and higher frequencies the FET will give smoke signs of pain! The silicon will be punched by molten areas.
Usual FETs are not built for analog operation. Special makes are available for this task where special precautons are made in order to dissipate the heat generated inside.

We learn that for initial testing with low stress in our circuit higher gate resitstance might be helpful in order to detect problematic behaviour but later on the switching action shall be as fast as possible.
The small gate resistance serves for dampening resonant oscillations between gate capacitance and inductance of the wire. Those oscillations will give poor switching property because the FET shows multiple transitions while oscillating from on to off state -> RF disturbance, losses!.
You should know that FETs are extremely fast semiconductors. Their limit in switching originates from the gate capacity and the aibility to pump electrons there to in order to generate the electric field. There are circuits out there where quite normal FETs show switching speeds in the picosecond range! This explains their inclination for HF oscillations.

Sorry, for complexity. If we leave proved and dead safe areas some conditions start to become essential!

Summed up: The low side switch seems to be proved and can be built by others but not the high side switch.

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

For all you brave pioneers in opto matters:
Corrected circuit for high side drivers:

• DC/DC converter -> pin 1/2 now correct
• Protection diodes added -> D1/2 shall be of of low capacitance type . Normal 1N4007 will do for lower frequencies.

@Larry: What was your idea of inserting protecting diodes?

ADVICE: Please start with low stress on your circuit: voltage, load, duty cycle, frequency. Increase stress conforming your expertise you learned from operating your circuit.

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

Hi John
Do you have a secret source for the DC-DC converters. Everywhere I check has no stock or 25 min. order. Maybe a cross ref #???
Dana

External diode parallel with FET internal diode. Suggested earlier by someone else and I read in a couple of other places that it is a good idea as the internal one is barely sufficient. If you have cut the drain pin off your FET then it will be difficult to install.

Regards
Larry

G'Day Larry

When I say I connected the output from the Oscillator or the output from the Regulator I presume I am connecting to the input of the mosfet circuit when I connect to the Mosfet circuit Gate Bus
I actually connected to the Mosfet circuit Gate bus, when I connect the Oscillator this way No Light ON when I By pass the Oscillator by connecting the Regulator to the Mosfet circuit Gate Bus the the light comes on Bright and when I disconnect it goes Off so I am the Regulator and the Mosfet circuits are working

I have purchased of Ebay on of these

Square Wave Signal Generate Generator Module * Frequency / Duty Ratio Adjustable | eBay
it will probably take a week or 2 to get here then when I connect it Up everything will work fine or perhaps

Regards



Still working at it

In Europe we have a big ditributor Conrad - Ihr Online Shop für Elektronik, Computer, Multimedia, Modellbau & Technik. They sell DC/DC-converters from TRACO, TDK Lambda, HN-Power, RECOM.
Other manufacturer: Aimtec

You need to know that you get two kind of DC/DC convertes:
1. no isolation in/out = 3 pin
2. isolation in/out minimum 4 pin

We need the types of second item with minimum 1000V isolation capability.
I used RECOM.

There are other sources for bare PCB type (open frame / no case around). @Larry: you got such a simple bare bone DC/DC converter. Do they sell what we need here?

Every region all over the world sells components differently. So there is no general source. @all: please contribute sources for UK, US, AU ..... I don't know where you all sit in our global village and contribute.
rgds John

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

Nice module!
1.
Please note that they use the CMOS type of 555 timer. The output is source/sink about 8mA (not 200mA like NMOS type). No problem so far but you need to account for that.

2.
Can you share the schematic? My schematic posted originates from a bought kit as well.

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

Yes, right! But it is not needed generally. It depends on the FET and the quality of the internal diode. A UF4007 or similar seems to be a good choice (initially). I will insert it and test it as soon I get some spare time. Might be too tiny if you decide for heavy monster coils.
@ALL: If you test a new or upgraded setup, start with less stress first! Lower voltage, medium frequency, low duty cycle, resistive load or coil with resistor added. Measure voltage, currents if possible. I recently posted a peak detector for those not owning an oscilloscope.

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

Hi All

Here is my video of something related to this thread.
Pulse Motor Generator Coil Shorting Test

A Hall sensor triggers a MOSFET transistor to activate the driving coil in attraction mode to the rotor. The stack of magnets behind the driving coil is in opposition to the rotor.
When the driving transistor shuts off, the magnetic field from this stack pushes the rotor away from the TDC. When there is no power applied the rotor is affected by cogging only a little as the stack of magnets is pushing away while the iron/magnetite cores of the generator coils pull to the TDC. The rotor consists of 6 neodymium permanent magnets (N50), of which all the poles are oriented with their NNNNNN outwards.
Two generator coils that are connected in parallel are being shorted constantly many times per cycle using two MOSFET transistors connected source to source with bypassing diodes. This pair is being driven from a small variable frequency/pulse width square wave signal generator/MOSFET driver.
Two neon bulbs are connected across the generator coil pair being shorted.
There are moments when they put out quite a light show of purple flashes there.

I am not sure what is planned for the two motors but If we are going to convert one to a motor and one to a generator these are really nice coupling devices to connect the motor shafts together.

Jameco Electronics - Electronic Components Distributor

Les

Johnstone,

Last night I had to try again, so I took apart the opto circuit and rebuilt it one more time, only hi side, as per your your update.

1. C1=.1 uf 25v, C2=.1uf 50v, C3=.1uf 50v, C4=10uf 50v, C5=10uf 50v
R1= 10 ohm 3 watt, also used 20 ohm on opto pin 3-ground.
Diodes across fets (2 parallel), and D2 were RFU10Tf6S-ND, no D3.

Dana converters at digikey ( p/n 102-1366-nd) (mfg p/n vbsd1-s12- s15-sip)

2. I first used 3,parallel, 10 ohm resistors in series with small incandescent
5.7 ohms, in series with single coil 2.4 ohms ( a little over 10 ohms total).
Light lit and flashed,

3. removed resistors, lit and flashed fets barely warmed.

4. removed small light, connected cfl across single coil, did not light, but nothing fried.

5. re-connected bifilar 1.2 ohm, cfl did not light, 2 parallel fets got quite warm.

So, no fried parts, I still am not sure why it fried last time, no cfl but I'm sure that we can get get that working, I was only switching 12 v.
I think this circuit will work. Can't wait to get home. Will try more fets, 36v, hi and lo.

Very noisy at opto on scope, good luck John.

Peace
Thug

Everyone,

I just thought I'd touch base. I am still waiting for the CF tape to come. Dana and I have come to the conclusion that the tow doesn't work as well. Dana says that he read that the RE doesn't travel very far into the CF, so I am going to focus on surface area contact.

So far the most I've been able to get from a CF secondary coil (not attached to the circuit) has been 169.4v at 200mA of 24v input to the FO Device system. This is while running a NE-30 on the system output (a small load as the larger the load, the lower the CF coil voltage reads, because the CF coil is feeding the load on the output of the system). But I get 166.5v at 100mA input and 133.6 at 50mA. So obviously the lower input is much more efficient.

I also found that running a 120v LED (2W = 25W replacement bulb for the energy savers) that I got at Wal-Mart for $7 on the output could be lit at about half the brightness at 10mA. AT 40mA it was about the bulbs normal brightness, but that's still a ways below 2W. So we can run CFLs and LEDs cheaper than conventionally.

BTW, I am using two K2837 MOSFETs on this system. Do they work as well? Any difference?

Also, I am running my combined UFO circuit (combined from Cinan and MadScientist's post's) and the Frequency and Duty Cycle control has been very good. I have replaced the 100k pot and 100k resistor with three 100k pots and a 1k pot all in series which gives me 30 turns to cover a 300k range plus 10 turns of very fine tuning over another 1000 ohms. The Duty Cycle pot is 50k (10 turn Bourns). Bourns (the 3590 series if I remember right) is the smoothest pot I've found and the prices are reasonable.

I hope to have time to catch up on what everyone is building. (fat chance!)

I did buy the Radio Shack motors. Can't wait to see what UFO's got for us!

Enjoy,

Bob

If you keep blowing FETs it seems like a good idea to me. Many designers use them for a bit of insurance.

This would be nice Les, but the motor has a 2mm shaft and the coupling you indicated is for 3.13mm shaft.

Oh brother I'm sure scoring high over here.... My pictures and web sites didn't make it on my other post and now......

yup! I thought they had more variety there....Sorry 'bout that
We can always fall back to what DadHav did, and use a little piece of rubber tubing. Still, Since I don't know what UFO has planned it may all be a moot point anyway....

I finished my coil today. It is a two inch diameter three strand of #18 in parallel at 1.8 DCR. my digital Meters sure have a hard time reading that coil.
I had to set the meter at 100hz to get any kind of stable reading, and even then it was showing a negative resistance. No, not negative like that, just the negative sign on the meter. -1.8034... Odd though!

Why three strands?....Tesla 3, 6, 9, .....

Ran with half the power of my previous coils and the 300watt light(about 23 actual watts) lit up very easily. Found a lot of resonant points where it would kick on through the ranges. I found the easiest thing to do was set it to a 50% duty cycle then bring up the frequency while watching current. When I see a sudden surge of current I back off a bit and then tune the duty cycle. It seemed to work very well that way to find the points where the cfl would kick on.
Sometimes it would get so bright I would get worried and back off a bit.

It all ran in a range from 400-600 ma I could push it to about 1 amp but she would be gone... I hooked it up to my pulser circuit and it dumped faster, way, way, faster than any other coil I had tried. I didn't get the scope hooked up to it today to see but I am sure the difference is something like from 3hz with my other coils to probably 80hz. If it's important I can measure it though...

This cap dump circuit is a very good measure of output because the dump takes place at the same voltage so the faster it dumps the more energy your getting. And yes the input current for all my other coils was pretty much in the same range.

I also tried a few core materials from welding rod, small iron rods, to metglass and didn't see any real observable difference other than a drop in current of a milliamp or two for the Metglass.

Bob, Dana, Still no where close to the low current you guys are talking about though. My mjl's held up nicely as always. they would get a little warm at those times when "She" wasn't there but just about to come in. and if I tuned to just past "When 'She' would leave" then the transistors took a big hit. I have a feeling this is where you guys are having trouble with the fets. There is some kind of kickback at that point. I think that is what may be the fet killer. and remember she goes backwards through stuff so even with diodes it may still be pretty hard to keep the fet's safe. More testing needed though....


Sorry for such a long report but pretty cool stuff today....
Les

When you said the transistors took a big hit, exactly what was it (volts and amps) and what was the duration?
BTW the diodes are to protect from the kick back. External diode should be faster than the internal diode and very fast recovery.

Larry

Hi Larry,
I had two things happen, If stayed in the "bad zone" my trannies would get warm. when I would come out of resonance on what I would call the upper side (increasing frequency or pulse width) And the light would go out. My voltage would momentarily drop to almost zero. So I should say the power supply took the hit but in that same sense the transistors had to pass that surge current which would warm them up as well. After getting back in resonance the transistors would get cool again. I wish I had a pulse meter and I could measure the surge current and give a much better Idea of what is going on. It is odd to me that the higher current is happening when the light is off and things are apparently out of resonance.

Certainly we know that when we play with these coils we can drop impedance very low. I would guess in terms of instantaneous current we could hit 300 amps or more...? because of the duration we aren't talking high wattage per say but certainly enough to take out some components along the way...
In all honesty Larry, I think if I had fet's in there I wouldn't be testing tomorrow, I would be rebuilding my circuit. And I am a big fan of Fet's...

I think the hardest thing for these devices Fet's and transistors are cutting off those high surges and that may be the area causing most of the problems with letting the smoke out...

Anyway just my WAG Larry.....

Les

Les_K
One of the things that we are doing at the same time that we are turning the pots is trying to watch all the gages and meters and trying to analize at the same time. There is a lot going on and how we analize at that time gives us our memory of the events. By using the RadioShack 22-812 digital meter which is set to record its data at a vary fast rate, on each area of study, and do it several times each area, we might be able to catch a better idea of all the activity on the data log and see just what we are up against. By getting this type of data from several areas on the circuit, something may make sense. Just a thought.
When working on the CF I thought it would be a good idea to compair against the same coil in copper. I built a small coil to insert into a core by starting with welding rod core of 3/4 inch diameter. Then rapped thin tape onto that and added two 1/8 inch tall ends. Rapped 1300 feet of 34 gage copper onto the thing and set it into the primary core with diodes attached. 284 volts.... Keep in mind that primary battery was 12 volts, 4AH. I just do not think that there was any amps. Several things may have been at play here. Yes, the wire was small, but this is also the same way I wound my primary coil the gets so much power, much more than any other coil I have made in any configuration. By using a spool core we have a small distance of plastic to go thru and I think that this space between wire and core is killing maybe 1/2 the power. It is this space that is also killing the power on the Tow that is covered by a thin caseing of heat shrink. The larger flat carbon is only covered by thin cellophane and is much stronger. This may be due to almost direct contact. When we are also going to be getting into building some more coils at the tail end, we may take advantage of this direct contact idea as what we need there is strong forces. Again, just a thought.
Dana

Hi Dana,
I have heard some talk about the air gap between the windings and core I hadn't thought to much about that, very interesting. My coil is not litzed on the primary. and I have 400 turns. I tried a small coil inside that was made of #23 wire I don't know how many turns or anything just a spare one that was around. at 12 volts in I only got about 150volts out of the secondary no power. But what was the real surprise is that it seemed to have no effect on the primary, even when shorting the secondary out. At this point My primary is air core and that might be what is making the difference. I am thinking about doing something like UFO's ufo coil. I may use the welding rod but bring it up around the outside of the coil on the ends like that and see if that does anything about catching that vortex on the end. I would like to build a 6000 turn secondary and slip that inside and see what happens. I'll have to see about getting some more wire for that. In a regular transformer the real trick is getting all the energy into the core so it is transferred to the secondary as efficiently as possible. So I am still trying to grasp the difference here between a regular transformer and what UFO is doing. I am hoping as we begin to play with the motors we can get a better picture.
Andrew (armegdn06) has some great vids on youtube I'll see if I can find them and pass them along here. I think they are quite relavent.

Les

Ok, this is his cute video of showing the transfer of energy from one coil to another.

Energy Propagation - YouTube

I would highly recommend watching his whole series I think his work really helps understand what UFO is doing here. on one of his coil setups you will see a very large center coil made with copper pipe that looks like something Dollard would do...

Les

Just rememberd to have red somewhere that tin coated copper wire produces a 4 fold magentic field. Normally magnet wire is not tinned. There are sources for tinned enameled wire but unfortunately I did not find a source selling less than tons of material. But most PVC insulated wires are tinned.
So please consider and post it if you used such wire. It might be an important detail for success.

__________________
Experts spend hours a day in order to question their doing while others stopped thinking feeling they were professionals.

Hi John

You can buy it from RS in the UK, I asked them yesterday if they did bare copper, and this is what they offered - a coincidence! However it is not insulated.

Buy Hookup & Equipment Wire Tinned annealed copper wire,24swg 91.3m RS 1288P online from RS for next day delivery.
Plus this which is insulated:
Buy Hookup & Equipment Wire 1.0mm Panel Wire UL-CSA-HAR 1015 Brown Lapp 4160303 online from RS for next day delivery.

Regards

John