Jim Mac's Figuera Thread

[Jim Mac]

Here is the results of the circuit build.

• Two 20N60C3 N Channel Mosfets
  
• 2 Channel Signal Generator 
  
• 1 DC supply
  
• and a Few Connection Terminals..
  

   

Using a Stock Wave Form to fire the gates, this is the output to 8Ω resistors..  (I have the scope leads reversed on purpose to keep the grounds bonded)

   

Not Too Shabby!

I started playing with arbitrary wave forms and I am sure I can get these waves much better, if not perfect with some tweaking.

The Mosfet selection is obviously very important.  And they are rated for 600V I think

[lfarrand]

If I raise the supply voltage, I must raise the gate voltage.  But the offset-  once set it stays there.

MOSFETs are quite sensitive to gate voltage and don't tolerate more than their datasheet maximums, so make sure the gate voltage doesn't exceed that otherwise they will let out the white smoke.

The MOSFET gate voltage is relative to the voltage on the source pin, so that's why it's specified as the gate-to-source voltage: Gate-Source Threshold Voltage - Vgs(th)(min) and Vgs(th)(max)

You shouldn't need to raise the gate voltage if you raise the supply voltage. The supply voltage will be whatever is on the drain pin, whereas the gate voltage is relative to the voltage on the source pin which would normally be 0V. Since the source voltage doesn't change if you increase the supply voltage, that means you don't need to change the gate voltage for different supply voltages. Once you've dialled it in then you can change the supply voltage as you please.

Learning curve here..  But I am pretty sure the Mosfet selection and Quality is Paramount.  I just ordered a 5 pack of these suckers

One thing I will say about MOSFETs (or any IC actually) is that counterfeit chips are big business in China and fake or rebadged devices are rife. I've even bought some from Amazon and they were fakes. For MOSFETs in particular, often the fakes will work similar to the real thing until you turn up the current. The reason for this is that the more expensive MOSFETs are able to handle larger currents due to a physically larger die area. Cheaper MOSFETs are rebadged and resold as their more expensive counterparts.

You might find that multiple MOSFETs from the same packet are inconsistent and behave very differently. Imagine blowing a MOSFET, replacing it with another seemigly identical one and then finding that your device doesn't work the same as it did before. How frustrating! I only buy them from a reputable source like Mouser, Digi-Key, Farnell, RS-Online etc. because I don't want to deal with inconsistencies. I want repeatable, consistent results.

Do yourself a favour and buy some relatively cheap Infineon 1200V SiC TO-247 MOSFETs from a reputable source.

[Jim Mac]

Thanks Ifrand again..  I will see if I can get this right..  But first I want to show outputs!  Looks Promising!

First-  According to conventional understanding, Open Circuit volts X Short Circuit Current=  Wattage can NEVER exceed Input Wattage.  

   

YES I understand that Open Circuit Volts and Short Circuit Amps is NOT a representation of actual output watts..  No need to point that out..  Just take it as it is for now..

This Circuit can output 4.5 TIMES the input when calculated with those values.  



I have NOT yet tried a real load yet..  But I will soon

[Jim Mac]

I tried some tests with the output and tried looping it. Guests cannot see images in the messages. Please register at the forum by clicking here to see images.

The output fills big caps just fine.  I filled a 12,000 uf cap to 40+ volts.  Then I connected the cap to the supply Input through a DC current meter.  Current flows from output to input.

But it will not self- sustain.  I disconnect the DC supply and the circuit peters out to a stop.  

I tried several other configurations and saw all sorts of interesting stuff.  Including one configuration that makes input amps go down when connecting a resistive load.

I think the answer is somewhere within this setup.  I have all sorts of things to try.

[Jim Mac]

Well I am still working on the wave and I can not get it right.  Some parameter is off..  Let me explain.

First, this is the same coil the below demonstrations are using.

   


Now I set my function generator on to a positive biased sinewave and feed the coil directly.  This is the scope hooked to the coil leads.  When I hold a magnet next to the coil I easily feel the pulsing field.

   

This above is EXACTLY what I need, but the wave needs to be created with resistance..  It is soo smooth, even going into a reactive coil, the wave / magnetic field is perfect.  

Now I hook up the Mosfet and tune the mosfet rig with a resistor.   This below is about as good as I have been able to get it.

   

From the surface, it looks almost decent...   

Now I take out the resistor and input that wave into the same coil as before.  And I get this...

   

ABSOLUTE CRAP!!!!

So the AIM is to achieve a similar wave into the coil as the function generator does.  BUT the wave MUST be made with varying resistance.

********Scratching my head...****

[unimmortal]

The MOSFETs look like they take a moment to get off the ground...

The coil really craters that pulse. I'd rewind the coil and try to get it near perfect for what you are trying to do. Jungle wiring creates all kinds of weird anomalies.

[Lasco]

hello Jim;

I have just received cores from soft Iron 9 very low carbon content and going to experiment again,

I'm also scratching my head but I strongly believe that resistors are used only for starting this device, once is motor running, secondary will feed the primary with a nice sinus wave regardless.

selection of wire thickness for the primary is my problem and wire for the secondary is another, however, I'm in the configuration of all primaries "N" are in series and all primaries "S" are also in series  All secondary's "Y" are parallel 
therefore output must generate on one secondary coil  voltage that is going to be able to excite 7 primaries in series and have an extra current left to be used with other load 

I started with a thin wire but it did not work due time constant of the solenoid - the wire must be thicker for the reduction of the time constant 
now I'm on 2.5mm diameter copper wire

yes you are correct Jim ----lots of head-scratching

[Jim Mac]

Cool Lasco-  I will be watching your thread.  The more methods covered the better..

~~~~~~~~~~~

So I resolved 1 problem..  The skitchety part of my wave was because the current draw of the gate.  Apparently the function generator can not supply the in-rush of current needed under operation.  4uF of AC caps in parallel with the function generator seems to have corrected it..

I still have another issue I am working on with the sinewave, but I will cover that a bit later..

Anyway,  now I can produce Full AC 2 polarities with this mosfet.  And when I balance the wave on either side, the input drops to Zero amps input.  My fluke had 3 decimals to read current and it registers as 000...

Now I am going to show a video of some effects now.  I will say now-  some assumptions and commentary on how it is working is WRONG in the video..  But I will cover that later once I get a bit further..

Bur for now-  watch this sucker on the CUSP of Self-Sustaining...  I turned the supply off and it kept running over 2 hours..  And it would have went even longer but I killed it trying to switch caps while it was going..



So..  I will say this..  The output from the rectifier was NOT charging the cap like the video claimed..  The output did not have enough voltage to do that.

[Jim Mac]

Ok, well the output of the last post is miniscule.  

I tried several methods to control this mosfet gate to use decent input power but keep hitting problems.  I can get a near perfect wave with a resistor as a load, but when I attach a coil as a load the wave form changes for the worse. I suspect something is happening with the reactance of the coil affecting the gate ground..  It may be worth exploring if optocouplers can be used to isolate the gate ground from the drain but I don not know how this would wire up.  

As I experiment further, I also ordered some stainless steel powder.  I am hoping it works better than the graphite powder in resin to make the resistive ring idea.  

Basically, I really want to find a solution to produce a perfectly smooth positive biased wave using pure resistance and nothing else. Whether it is a mechanical brush or a solid state, either way would be good for now.

So we will see what I come up with

[Jim Mac]

In the next few days, I plan on doing several tests to find a good recipe to make the resistance ring.

First Attempts:

I tool JB-weld Steel Stick which is a fast curing high strength putty epoxy.  I mixed 1 sample with Graphite powder,  and mixed a second sample with iron filings.  

These will be cured in an hour or two. I will report resistiveness at that time.  I am expecting POOR results.