new oscillator circuit shared By Nelson Rocha

evostars

How much power is that purple line?
« Reply #225,  »Last edited
I Did another test, or better said, I played around, trying to get some result. Seems my Load on L3 (111 Ohm) is a bit too much. It might restrict the current in L3.

I stacked the coils: L1 is bottom, then L3 is placed on top of L1 but, with 15mm spacing. L1 and L3 are equal size 10M 0.75mm2 bifilar coils.
L2 is an equal mass, but 0.4mm2 aroun16 meter (see other post) bifilar pancake coil, and is placed on top of L3.

I removed C1 from the circuit.
C2=10nF c3=635nF c4=2.3uF c5=25nF C6(parallel over L3 and Load)=121nF

Power consumed by the supply: 0.87A  11.2Vdc =9.744 W

Newfile18 shows:
yellow= current probe 100mV/A   20mV/DIV
Blue= 10:1 L3 outside rim              50V/DIV
Purple= Math: yellow X Blue (power)  200mV per div
Purple in the bottom says: max=352 mV

rough guesses:
yellow = 60mVpp equals 0.6AMPs peak to peak
blue= 185Vpp

now Im going to ramble, because my head is foggy:
0.6A/2 devided by square root of 2 = 0.212132034  A rms
185/2 dev by sqrrt of 2=  65.407 V rms
If the amps and volt are in phase (which they are not but near) it would give 13.86W rms consumed
that would be more that the 9.8 at the supply.
But this is very speculative.

I think it better to conclude,  my load is much to big. the 111 Ohm is part of a 2500W heat Fan. A bit too much I would say....
So I will get another Load. A Car light bulb wil probably get to much voltage. So... Maybe back to the pure resistive load of 2x8Ohm in series (50w power resistor)  But thats no fun either... I need a "working" load. I need to see/feel the power.
maybe a 15W lightbulb again.



 Newfile18.jpeg - 118.58 kB, 800x480, viewed 5 times.


Matt Watts

Re: new oscillator circuit shared By Nelson Rocha
« Reply #226,  »Last edited
Quote from evostars on November 13th, 01:40 PM
I need a "working" load. I need to see/feel the power.
Maybe it will work; maybe it will not, but here's what I'm going to use for the SERPS device I'm working on...

A pair of 24 volt, 150 watt, universal input power supplies.  I may place a bridge rectifier in front of them, but I don't think it's necessary--pretty certain their input stage already has a FWBR.

These are about 90% efficient, so I can monitor the straight DC output power and have a pretty good idea how much input is available.

My thinking is:  If I can't energize a power supply like this, I'll never be able to loop the device anyway, so might as well configure the system to do what I want it to do straight away.

 pmc-24v150w1aa_SPL.jpg - 52.3 kB, 448x324, viewed 0 times.


evostars

Re: new oscillator circuit shared By Nelson Rocha
« Reply #227,  »
Quote from Matt Watts on November 13th, 02:45 PM
Maybe it will work; maybe it will not, but here's what I'm going to use for the SERPS device I'm working on...

A pair of 24 volt, 150 watt, universal input power supplies.  I may place a bridge rectifier in front of them, but I don't think it's necessary--pretty certain their input stage already has a FWBR.

These are about 90% efficient, so I can monitor the straight DC output power and have a pretty good idea how much input is available.

My thinking is:  If I can't energize a power supply like this, I'll never be able to loop the device anyway, so might as well configure the system to do what I want it to do straight away.
FWBR? whats that.
Keep in mind if your power is high frequency ac you will need fast rectifier diodes

Matt Watts

Re: new oscillator circuit shared By Nelson Rocha
« Reply #228,  »Last edited
Quote from evostars on November 13th, 03:44 PM
FWBR? whats that.
Full Wave Bridge Rectifier -- A square of diodes.
Quote from evostars on November 13th, 03:44 PM
Keep in mind if your power is high frequency ac you will need fast rectifier diodes
Yes.  I have some ultra-fast rectifier diodes (MUR1550) I can use if I need them.  Same thing Ruslan used on his device.

haxar

Re: new oscillator circuit shared By Nelson Rocha
« Reply #229,  »
Quote from Matt Watts on November 13th, 07:20 PM
Yes.  I have some ultra-fast rectifier diodes (MUR1550) I can use if I need them.  Same thing Ruslan used on his device.
These are the same diodes (MUR1560) that Stan used on his VIC.

evostars

Re: new oscillator circuit shared By Nelson Rocha
« Reply #230,  »
new load 188 ohm (cold and off) 15W 220V regular lightbulb.

its higher than the 111 ohm of the heat fan resistive wire, but those are measurements when cold and turned off. resistance changes when they heat up.

I would prefer a low voltage 12V automotive 15W bulb, but I think I need some higher voltages now.

photo shows stacked coils from the side. l3 in the middle under L2(top) L1 is the bottom and us distanced by a 15mm roll of insulation tape

 IMG_20181114_150850.jpg - 254.04 kB, 1280x720, viewed once.

 IMG_20181114_150903.jpg - 276.82 kB, 1280x720, viewed 0 times.

Re: new oscillator circuit shared By Nelson Rocha
« Reply #231,  »
turned out, that the previous measurements didn't have a spike in  L2

going to decrease c6 (l3)
Re: new oscillator circuit shared By Nelson Rocha
« Reply #232,  »
removed 77nf from L3
lamp starts to glow
c6 is now 46nF

lamp glows when the sinus is pure. no more disruptive discharge in l2 perfect  sine. slightly lower voltage than l3.

Re: new oscillator circuit shared By Nelson Rocha
« Reply #233,  »
So... I can make a spike in L2
but the lamp is bright when I tune it just a bit lower freq then where the spike starts.

for instance 100khz gives the biggest spike. tune down to 88kHz and the spike disappears in L2 leaving just the resonant sine. at this point, the lamp is most bright.

at the same time c5 and c6 influence the frequency where the lamp is brightest.

c5 tunes L2
C6 tunes L3
but L3 and L2 are coupled.

Nelson suggested removing the capacitor c6 from L3 and only look at the load. tune it to max current with c5...
Re: new oscillator circuit shared By Nelson Rocha
« Reply #234,  »
at 53 khz my power supply goes nuts.
the left channel is off. zero volts and amps. but... then it clucks the relais inside and it shows voltage. for no reason.
something is triggered

both are independent. and normal maximum range is 32 V

 IMG_20181114_155641.jpg - 200.36 kB, 1280x720, viewed once.

Re: new oscillator circuit shared By Nelson Rocha
« Reply #235,  »
removed c6
lamp glows when l2 is max resonant without discharge.
L3 has 125V sine but distorted after peak
Re: new oscillator circuit shared By Nelson Rocha
« Reply #236,  »
In the document Radomir (forum) shared, is described what Nelson (partly) uses in his circuits to get HV impulses. I don't fully understand it all. But it seems a diode can create a kind of water hammer effect.
the diode is closed by a intense high reversed current pulse,  at the same time  It is charged up by a back emf of a coil. This creates a kind of water hammer effect.
Very Very interesting. This gives good insight. Like the inductance of the coil is influencing impulse rise time. I already noticed this when changing L2 to a higher inductance (more windings) coil.
Re: new oscillator circuit shared By Nelson Rocha
« Reply #238,  »
http://www.avtechpulse.com/papers/thesis.pdf

complete thesis on same subject


study time. learn how this works in nelson circuit. seems he also uses 2 diodes in series increasing the effect
new info on d1 and d2
« Reply #239,  »
D1 and D2 are ultra fast Diodes Mur1660CT .
gate drive IC dead, probably need zener, mercurius retrograde?
« Reply #240,  »
like the title of this post says. my IXDN604PI gate driver
http://www.ixysic.com/home/pdfs.nsf/www/IXD_604.pdf/$file/IXD_604.pdf
 is dead.
I tried reproducing the original spike setup. but saw it didn't switch.
 Gate - Source had only 5 V difference. needs minimum of 10V I use 12.7V (battery)

specs say: Vcc +0.3 V absolute max.
I remember seeing overshoot, so I'm going to put a 10v zener on the input of the dual channel ixdn604 (pin 2 or 4)

I will also try the other IC channel, maybe that one still works. I have a spare but these are not cheap...

I remember seeing a lot of ringing on the gate. much more than in the beginning when everything worked fine.
Re: new oscillator circuit shared By Nelson Rocha
« Reply #241,  »
My new video, about high side switching:


https://youtu.be/Asa2RuubKWM

Matt Watts

Re: new oscillator circuit shared By Nelson Rocha
« Reply #242,  »
Just a couple of little comments evostars:

Keep in mind a typical buck converter isn't isolated.  The battery is isolated, but due to its size (mass), it has quite a bit of parasitic capacitance.  What this means is the battery now floats on the high side.  It is actually connected to the source pin of the MOSFET--same place where the back EMF voltage will be seen when the transistor begins to shut off.  So if you think of the battery as one side of a large capacitor plate and the environment as the other, there's a good deal of charge moving around each time a voltage spike is created from the coil.  How fast this charge moves around is uncertain due to the capacitance.

My rule of thumb has always been to remove as much mass as possible from a high-side gate driver system--keep any and all conductive elements to a absolute minimum.  To do this I use tiny isolated DC to DC converters, placed as close to the gate driver and transistor as possible.  This minimizes the conductive elements that will be subjected to the high voltage back EMF pulses and lessens the chance of cross conduction from actual ionization.  With my Universal Switch when voltages begin to exceed about 350 volts, I can actually see blue arcing between the heatsink and the circuit traces.  This was unexpected when I designed it, but now I know when switching coils that are able to produce large back EMF, this design doesn't work properly--less conductive mass and better insulation is needed to deal with these voltages.

Switching is a real art.  Switching coils intentionally to produce high voltages is even tougher.


evostars

Re: new oscillator circuit shared By Nelson Rocha
« Reply #243,  »
Quote from Matt Watts on November 16th, 12:44 PM
Just a couple of little comments evostars:

Keep in mind a typical buck converter isn't isolated.  The battery is isolated, but due to its size (mass), it has quite a bit of parasitic capacitance.  What this means is the battery now floats on the high side.  It is actually connected to the source pin of the MOSFET--same place where the back EMF voltage will be seen when the transistor begins to shut off.  So if you think of the battery as one side of a large capacitor plate and the environment as the other, there's a good deal of charge moving around each time a voltage spike is created from the coil.  How fast this charge moves around is uncertain due to the capacitance.

My rule of thumb has always been to remove as much mass as possible from a high-side gate driver system--keep any and all conductive elements to a absolute minimum.  To do this I use tiny isolated DC to DC converters, placed as close to the gate driver and transistor as possible.  This minimizes the conductive elements that will be subjected to the high voltage back EMF pulses and lessens the chance of cross conduction from actual ionization.  With my Universal Switch when voltages begin to exceed about 350 volts, I can actually see blue arcing between the heatsink and the circuit traces.  This was unexpected when I designed it, but now I know when switching coils that are able to produce large back EMF, this design doesn't work properly--less conductive mass and better insulation is needed to deal with these voltages.

Switching is a real art.  Switching coils intentionally to produce high voltages is even tougher.
I wonder if the back emf is also able to charge the battery, it seems my battery hardly drains. but it could also be the load being so small.
What doe you mean with "cross conduction from ionisation"?

And the dc to dc can be isolated, but still the negative would be connected to source, right?
Re: new oscillator circuit shared By Nelson Rocha
« Reply #244,  »
Circuit still isnt switching properly, put in 12V zeners on in and out of the gate driver,
replaced mosfet
but only 7.5V between gate and source. should be 12V

must be my switching print....