the epoxy acts as the dielectric of the capacitor. L2 and L3 are the plates of a capacitor.
epoxy is much better then air.
the L2 impulse, discharges the capacitor, which produces a current in L3.
at the same time, L3 is a coil that is resonant, and this discharge current, amplifies the L3 resonant current.
Radiant coil capacitor
Ah I see, thanks for clarifying. Are you using speaker wire as-is (with plastic coating) within the epoxy, or are you using bare copper wire? If using it with the plastic coating, then doesn't this also act as a dielectric of sorts? I presume the most effective solution would be to strip the plastic coating, wind the copper coil using the bare wire and then encase that in epoxy, right? This of course presents further problems such as maintaining a uniform distance between the windings.
most important is to get the air bubbles out, as corona discharges introduce losses.
I also tried sunflower oil, but it's very very messy, although it is a good dielectric.
increasing the dielectric constant (K) increases the energy stored in the coil capacitor, thats why a ceramic titanium (dioxide) would be most ideal. But without air
Impulse on primary must be a harmonic of the secondary coil (L3 no cap)
I need to get rid of some ridged thoughts.
I always assumed I needed the impulse on L2.
But what If I dont need it?
What if the impulse of L1, is perfectly absorbed into L2.
No more ripples, but also no more impulse visible on L2.
Just a perfect sine wave.
Then with parametric exitation, when both the positive and negative impulse of the 2 L2 coils is perfectly absorbed into L2, creating a perfect sine wave, then L2 as primary would hold massive amounts of energy.
this is what I constantly see. When L2 is perfectly tuned resonant, the impulse "disapears" as it is absorbed into the L2 primary coil, forming a sine wave, without impulses.
but due to the tricks in time, from the L1 coil, the energy present in L2 by parametric excitation, is much higher, then what normally would occur. L2 is at a much higher energy level, then the power supply would normally be capable of.
So the L3 becomes resonant, also at a much higher level.
So... the question then becomes... do I need that impulse on L2?
How many forms can L2 take?
Ideally I still think, we at least need to be at pure resonant frequency of L2. not detuned above or below.
And, still I think, L2 should have the impulse present at that tuning. (on not by detuning slightly above resonance).
I say this, because L3 has the most energy when L2 is tuned perfectly resonant.
And not when L2 is slightly detuned, above the Fres. to get the impulse visible.
So L2 really need to be tuned by the series capacitor.
I always assumed I needed the impulse on L2.
But what If I dont need it?
What if the impulse of L1, is perfectly absorbed into L2.
No more ripples, but also no more impulse visible on L2.
Just a perfect sine wave.
Then with parametric exitation, when both the positive and negative impulse of the 2 L2 coils is perfectly absorbed into L2, creating a perfect sine wave, then L2 as primary would hold massive amounts of energy.
this is what I constantly see. When L2 is perfectly tuned resonant, the impulse "disapears" as it is absorbed into the L2 primary coil, forming a sine wave, without impulses.
but due to the tricks in time, from the L1 coil, the energy present in L2 by parametric excitation, is much higher, then what normally would occur. L2 is at a much higher energy level, then the power supply would normally be capable of.
So the L3 becomes resonant, also at a much higher level.
So... the question then becomes... do I need that impulse on L2?
How many forms can L2 take?
Ideally I still think, we at least need to be at pure resonant frequency of L2. not detuned above or below.
And, still I think, L2 should have the impulse present at that tuning. (on not by detuning slightly above resonance).
I say this, because L3 has the most energy when L2 is tuned perfectly resonant.
And not when L2 is slightly detuned, above the Fres. to get the impulse visible.
So L2 really need to be tuned by the series capacitor.
I keep forgetting...
I need to check for ground currents
I need to check for ground currents
I need to get rid of some ridged thoughts.
I always assumed I needed the impulse on L2.
But what If I dont need it?
What if the impulse of L1, is perfectly absorbed into L2.
No more ripples, but also no more impulse visible on L2.
Just a perfect sine wave.
evostars
efficiency comparison test of L1 and L2 close/not coupling
« Reply #331, on December 24th, 2021, 03:13 AM »
In addition to my latest video, I decided to test the efficiency of the series resonant L2 coil, being close coupled to the L1 impulse generation coil versus not being coupled at all.
Since coupling changes the resonant frequency, I decided to tune to an equal amount of resonant current (10A pp) and devide the input power by the frequency, so I could see the amount of energy per kc/s of each setting.
I used a small capacity is series of L2 of 12nF, this prevented the production of ripples.
attached are SCR63 which is the close coupled setup
and SCR65 which is the not coupled setup.
SCR63 close coupled data:
input power: 21.6W (0.47Ax46.0V)
F res= 167.9Kc/s
power input per 1000 cycles=0.1536W
SCR65 not coupled data:
input power: 35.3W (1.1A x 32.12V)
F res=98.4 kc/s
power input per 1000 cycles: 0.3587 W
the close coupled system, is much more efficient. it needs 43% less energy for the same amount of resonant energy.
The impulse is not fully absorbed into the resonant system, and the impulse voltage is higher at the lower frequency of the uncoupled system. but it cannot account for the losses, as the impulse energy is fully restored into the capacitor by the usage of the body diode. (video explains circuit)
My latest video, which explains the efficient use of resonant energy:
https://youtu.be/Pa2MipWGSKQ
Since coupling changes the resonant frequency, I decided to tune to an equal amount of resonant current (10A pp) and devide the input power by the frequency, so I could see the amount of energy per kc/s of each setting.
I used a small capacity is series of L2 of 12nF, this prevented the production of ripples.
attached are SCR63 which is the close coupled setup
and SCR65 which is the not coupled setup.
SCR63 close coupled data:
input power: 21.6W (0.47Ax46.0V)
F res= 167.9Kc/s
power input per 1000 cycles=0.1536W
SCR65 not coupled data:
input power: 35.3W (1.1A x 32.12V)
F res=98.4 kc/s
power input per 1000 cycles: 0.3587 W
the close coupled system, is much more efficient. it needs 43% less energy for the same amount of resonant energy.
The impulse is not fully absorbed into the resonant system, and the impulse voltage is higher at the lower frequency of the uncoupled system. but it cannot account for the losses, as the impulse energy is fully restored into the capacitor by the usage of the body diode. (video explains circuit)
My latest video, which explains the efficient use of resonant energy:
https://youtu.be/Pa2MipWGSKQ
Some ideas of combined resonance
I tested L4 with a series capacitor to ground.
Since it is fed by L3 with an AC signal, it can be come series resonant again.
the L2 primary haveing the lower frequency sine combined with the higher frequency impulse,
could be used to induce parallel resonance in L3 using the L2 s8ne (lower frequency)
while the impulse is used to excite the L4 series resonance?
the combined resonance of L3 and L4 than could produce power somehow?
the l4 l3 interaction could be a form of parametric excitation
I tested L4 with a series capacitor to ground.
Since it is fed by L3 with an AC signal, it can be come series resonant again.
the L2 primary haveing the lower frequency sine combined with the higher frequency impulse,
could be used to induce parallel resonance in L3 using the L2 s8ne (lower frequency)
while the impulse is used to excite the L4 series resonance?
the combined resonance of L3 and L4 than could produce power somehow?
the l4 l3 interaction could be a form of parametric excitation
In my last video I showed how to give more power to the impulse by placing a parallel capacitor over L1, which makes the parallel resonance cycle longer in duration.
But from the parametric excitation videos, I also learned, the excitation needs to be twice the frequency of the resonant system that is to be excited.
So the parallel resonant impulse duration of L1 ,
needs to be twice the frequency of L2.
Since this will demand more power from the supply, we will need to increase the voltage supply.
As this already is quickly limited, we need to lower the frequency of L2.
Then we can pump in more current into the system (which isn't at its 5A power supply limit yet).
Once these conditions are met, the ripple should also be completely vanished. due to the slower ripple.
Then L2 is parametricly excited, by the current AND voltage of the L1 impulse.
giving L2 massive resonant energy.
L3 then can be tuned to this frequency (keeping the L1 and L2 ratio of 2:1 in balance). into LMD mode.
Also. The aluminum foil floated at a set distance. I should measure this distance, and divide by 2. That should be the ideal distance between L2 and L3.
Divided by 2, because the counter force of the foil is equal to the force of the coil impulse. hmmm... but doesn't L3 also produce this counter force, by its induced resonant current? eh... no as it then is AC, not DC?
hmm...
This isn't even the coil I am using as L1. so it probably will be different anyway. Also depending on the power.
So.... leave it alone.
But from the parametric excitation videos, I also learned, the excitation needs to be twice the frequency of the resonant system that is to be excited.
So the parallel resonant impulse duration of L1 ,
needs to be twice the frequency of L2.
Since this will demand more power from the supply, we will need to increase the voltage supply.
As this already is quickly limited, we need to lower the frequency of L2.
Then we can pump in more current into the system (which isn't at its 5A power supply limit yet).
Once these conditions are met, the ripple should also be completely vanished. due to the slower ripple.
Then L2 is parametricly excited, by the current AND voltage of the L1 impulse.
giving L2 massive resonant energy.
L3 then can be tuned to this frequency (keeping the L1 and L2 ratio of 2:1 in balance). into LMD mode.
Also. The aluminum foil floated at a set distance. I should measure this distance, and divide by 2. That should be the ideal distance between L2 and L3.
Divided by 2, because the counter force of the foil is equal to the force of the coil impulse. hmmm... but doesn't L3 also produce this counter force, by its induced resonant current? eh... no as it then is AC, not DC?
hmm...
This isn't even the coil I am using as L1. so it probably will be different anyway. Also depending on the power.
So.... leave it alone.
parametric excitation.jpg
- 40.97 kB, 741x478, viewed 8 times.
I've decided to let the dual alternating polarity impulses go for now.
instead to focus on the single negative impulse in combination with the extra coil.
I want to see if I can get power between ground and the l3 l4 series connection.
I wonder if I can ground L4 with a capacitor, getting it to resonate at the L3 connection.
then having 2 resonant coils, but one at a higher frequency, that responds to the impulse.
L3 and L4 need to be harmonics. So I made a new L4 coil with equal copper mass, but 0.75mm2
maybe L4 needs to be counter rotated,
to add up.
the idea is the l3 and L4 need to be constructive, adding up voltages at their series connection.
this is then again rectified to dc.
maybe L3 provides current and L4 voltage.
(coming in phase at series connection).
need to play with this a lot.
instead to focus on the single negative impulse in combination with the extra coil.
I want to see if I can get power between ground and the l3 l4 series connection.
I wonder if I can ground L4 with a capacitor, getting it to resonate at the L3 connection.
then having 2 resonant coils, but one at a higher frequency, that responds to the impulse.
L3 and L4 need to be harmonics. So I made a new L4 coil with equal copper mass, but 0.75mm2
maybe L4 needs to be counter rotated,
to add up.
the idea is the l3 and L4 need to be constructive, adding up voltages at their series connection.
this is then again rectified to dc.
maybe L3 provides current and L4 voltage.
(coming in phase at series connection).
need to play with this a lot.
I'm just re-reading Gerry Vassilatos' Secrets of Cold War Technology and it seems that Tesla was using a single-turn copper helical coil with a capacitor and disrupter when experimenting with Radiant Energy in the early stages at least.
This provoked some shower thoughts and I was wondering what advantage a bifilar pancake coil capacitor has compared to simply using a helical coil with capacitor in an LRC circuit? Is it that the coil capacitor provides both inductance and capacitance in one component?
This provoked some shower thoughts and I was wondering what advantage a bifilar pancake coil capacitor has compared to simply using a helical coil with capacitor in an LRC circuit? Is it that the coil capacitor provides both inductance and capacitance in one component?
I'm just re-reading Gerry Vassilatos' Secrets of Cold War Technology and it seems that Tesla was using a single-turn copper helical coil with a capacitor and disrupter when experimenting with Radiant Energy in the early stages at least.
This provoked some shower thoughts and I was wondering what advantage a bifilar pancake coil capacitor has compared to simply using a helical coil with capacitor in an LRC circuit? Is it that the coil capacitor provides both inductance and capacitance in one component?
evostars
Re: Radiant coil capacitor
« Reply #337, on December 31st, 2021, 05:43 AM »Last edited on December 31st, 2021, 05:45 AM
new L4 coil, again series connected to L3.
L4 shows a resonant frequency of 3x 217.155= 651.465 kc/s (see attachent, orange)
So I need to tune L3, to a sub harmonic, to get its maximum amplification.
now I can tune L3 perfect resonant, but then there are a lot of ripples on L2, and a high power draw.
So I prefer to tune above the resonant frequency, where there is a single impulse on L2.
This draws less power, and still amplifies L4 and L3.
This tuning should be again a perfect sub harmonic of L4 (L3).
lets see If I can make that work
4th sub=162.9
5th sub=130,293 kc/s This one should be possible.
6th=108.58
7th=93.07kc/s this is also interesting
L4 shows a resonant frequency of 3x 217.155= 651.465 kc/s (see attachent, orange)
So I need to tune L3, to a sub harmonic, to get its maximum amplification.
now I can tune L3 perfect resonant, but then there are a lot of ripples on L2, and a high power draw.
So I prefer to tune above the resonant frequency, where there is a single impulse on L2.
This draws less power, and still amplifies L4 and L3.
This tuning should be again a perfect sub harmonic of L4 (L3).
lets see If I can make that work
4th sub=162.9
5th sub=130,293 kc/s This one should be possible.
6th=108.58
7th=93.07kc/s this is also interesting
SCR66.PNG
- 34.91 kB, 1280x824, viewed 12 times.
If I enlarge the parallel capacity of L3, to tune down to a sub harmonic of L4, then L4 wont form a single sine wave any more.
Another thing, is I think I should tune the duration of the impulse half wave to the L4 resonant frequency.
this means the ultra fast L1 should be slowed down again, by a parallel capacitor.
or, L4 should be made into a higher resonant frequency. this can be done by making the 2 windings of L4 bucking.
Done enough work for this year.
Happy holidays.
See you next year
:cheerleader:
oh... ps... L4 changes its frequency when L3 is tuned down...
Another thing, is I think I should tune the duration of the impulse half wave to the L4 resonant frequency.
this means the ultra fast L1 should be slowed down again, by a parallel capacitor.
or, L4 should be made into a higher resonant frequency. this can be done by making the 2 windings of L4 bucking.
Done enough work for this year.
Happy holidays.
See you next year
:cheerleader:
oh... ps... L4 changes its frequency when L3 is tuned down...
SCR67.PNG
- 35.44 kB, 1280x824, viewed 13 times.
evostars
Re: Radiant coil capacitor
« Reply #339, on January 1st, 2022, 04:34 AM »Last edited on January 1st, 2022, 04:41 AM
to sinq up the impulse to the L4 resonant frequency, I need the impulse to be a quarter wave only.
L4 period=4,605uS/3
so impulse needs to be 4,605/12=383nS
impulse is a half wave, so L1 resonant frequency would be 1/383nSx2)=1.3Mc/s
So I probably, only need a few hundred pico farads parallel to L1 to get that duration. This will lower the voltage of the impulse, so I can push the power of the supply up, to give more power to the impulse, but... the frequency is probably so high, that my power supply is alreadyu at its voltage limit of 64V. So I would need to lower the frequency again, to another sub harmonic of L4.
Then L2/ L3 should be tuned to a sub octave of L4, I need to check which, but probably it will be the 4th not sure on that one
will play... again.. a lot
in the end L4 should again form a single sine wave again on its sub harmonic, equal to the L2 L3 frequency.
but then L4 get maximum gain from the impulse also.
L3 gets induced by L2 by its (current) sine wave,
while L4 gets kicked by the impulse of L1 which is superimposed on L2
L4 period=4,605uS/3
so impulse needs to be 4,605/12=383nS
impulse is a half wave, so L1 resonant frequency would be 1/383nSx2)=1.3Mc/s
So I probably, only need a few hundred pico farads parallel to L1 to get that duration. This will lower the voltage of the impulse, so I can push the power of the supply up, to give more power to the impulse, but... the frequency is probably so high, that my power supply is alreadyu at its voltage limit of 64V. So I would need to lower the frequency again, to another sub harmonic of L4.
Then L2/ L3 should be tuned to a sub octave of L4, I need to check which, but probably it will be the 4th not sure on that one
will play... again.. a lot
in the end L4 should again form a single sine wave again on its sub harmonic, equal to the L2 L3 frequency.
but then L4 get maximum gain from the impulse also.
L3 gets induced by L2 by its (current) sine wave,
while L4 gets kicked by the impulse of L1 which is superimposed on L2
oh now I see
I just did a test, and it turns out that the L1 coil should not be slowed down... it's the L4 coil that need to speed up!
lets see If I get it right.
The impulse (half wave) should be a quarter of the L4.
so one full wave of L1, should be a half wave of L4.
So L1 should be twice the frequency of L4.
But how so I speed up L4? I don't have any capacity connected to it that I can decrease to speed it up.
the only thing would be, to reduce its amount of windings. And then if the frequency would be to fast, I could slow the L1 down a bit with some parallel resonance.
Damn... but then If I got it tuned, I would need to retune again, with the rectifier circuit connected to L4. So I better hook that up right away, before I waste time with tuning it all.
Still... very excited again, to test this tuning theory, and see if it hold up under load.
I need that load anyway, as L4 is already way above 4kV, need to keep it down to below 2k pp or else the rectifier diodes need to be doubled
I just did a test, and it turns out that the L1 coil should not be slowed down... it's the L4 coil that need to speed up!
lets see If I get it right.
The impulse (half wave) should be a quarter of the L4.
so one full wave of L1, should be a half wave of L4.
So L1 should be twice the frequency of L4.
But how so I speed up L4? I don't have any capacity connected to it that I can decrease to speed it up.
the only thing would be, to reduce its amount of windings. And then if the frequency would be to fast, I could slow the L1 down a bit with some parallel resonance.
Damn... but then If I got it tuned, I would need to retune again, with the rectifier circuit connected to L4. So I better hook that up right away, before I waste time with tuning it all.
Still... very excited again, to test this tuning theory, and see if it hold up under load.
I need that load anyway, as L4 is already way above 4kV, need to keep it down to below 2k pp or else the rectifier diodes need to be doubled
No I will keep this larger L4 coil.
and tune the L1 impulse by adding parallel capacity.
The ripples probably vanish with proper tuning, as then the aether can flow flawlessly.
if the impedances are not matched, the impulse expieriences resistance, and L1 cant fully discharge, reducing its voltage, like a stutter.
I have seen the ripple present with only one L1 coil, so it isn't a reflection.
hmm I still don't fully understand it.
but thats ok.
I suspect when the energy is able to flow into the load, then the ripples will vanish.
hopefully.
I could also enlarge the L2 L3 distance to get the frequency down.
and tune the L1 impulse by adding parallel capacity.
The ripples probably vanish with proper tuning, as then the aether can flow flawlessly.
if the impedances are not matched, the impulse expieriences resistance, and L1 cant fully discharge, reducing its voltage, like a stutter.
I have seen the ripple present with only one L1 coil, so it isn't a reflection.
hmm I still don't fully understand it.
but thats ok.
I suspect when the energy is able to flow into the load, then the ripples will vanish.
hopefully.
I could also enlarge the L2 L3 distance to get the frequency down.
ok, I hooked up the DC rectifier module (2 fast mur8100 diodes and 2 large capacitors, with ground in the middle)
, and yes I can still tune it into a single sine wave (needed to retune it).
When I put the load on it (42W halogen resistive wire lamp) then I can get it brightly lit.
by again retuning it, but not efficiently.
the L4 into the rectifier signal, is looking like a clipped sine wave (looks like a square wave) which is to be expected.
But if I look at L3, it isnt a proper sine wave anymore.
L2 current wave also doesn't look proper.
So now it comes down to proper tuning, while being blind.
Blind, because I can look properly at the L4 signal, as it is clipped by the lamp load.
and if I would remove the lamp I would have to retune it again.
basically, because the lamp+cap+diodes are like adding series capacity, tuning L4 down.
So I dont know If the impulse of L1 is good. it looked good, 400nS but probably now should be a bit slower.
So I'll ad a few 100pF to L1... and the rest is just... guessing.
With only the rectifier and the cap no lamp load, the L4 had a frequency of 583kc/s
while without , l4 was around 621kc/s, or even higher dpending on the parallel capacity of L3.
So... How will I tune this?
I guess,
by pure instinct.
and luck
and... intuition.
it's only 4 coils ...
, and yes I can still tune it into a single sine wave (needed to retune it).
When I put the load on it (42W halogen resistive wire lamp) then I can get it brightly lit.
by again retuning it, but not efficiently.
the L4 into the rectifier signal, is looking like a clipped sine wave (looks like a square wave) which is to be expected.
But if I look at L3, it isnt a proper sine wave anymore.
L2 current wave also doesn't look proper.
So now it comes down to proper tuning, while being blind.
Blind, because I can look properly at the L4 signal, as it is clipped by the lamp load.
and if I would remove the lamp I would have to retune it again.
basically, because the lamp+cap+diodes are like adding series capacity, tuning L4 down.
So I dont know If the impulse of L1 is good. it looked good, 400nS but probably now should be a bit slower.
So I'll ad a few 100pF to L1... and the rest is just... guessing.
With only the rectifier and the cap no lamp load, the L4 had a frequency of 583kc/s
while without , l4 was around 621kc/s, or even higher dpending on the parallel capacity of L3.
So... How will I tune this?
I guess,
by pure instinct.
and luck
and... intuition.
it's only 4 coils ...
Have you come across this Tesla coil driver? https://github.com/Netzpfuscher/UD3
My initial impression is that it's a computer to automate Tesla coil tuning.
In the features list it advertises that 'The coil sweeps through the given range and finds the resonant frequency'
I came across it here: https://highvoltageforum.net/index.php?topic=188.0
Perhaps this could be of some use to automate the tuning process and reduce manual effort each time? I imagine that pressing an 'autotune' button would be much simpler and less error prone than trying to manually tune.
My initial impression is that it's a computer to automate Tesla coil tuning.
In the features list it advertises that 'The coil sweeps through the given range and finds the resonant frequency'
I came across it here: https://highvoltageforum.net/index.php?topic=188.0
Perhaps this could be of some use to automate the tuning process and reduce manual effort each time? I imagine that pressing an 'autotune' button would be much simpler and less error prone than trying to manually tune.
Have you come across this Tesla coil driver? https://github.com/Netzpfuscher/UD3
My initial impression is that it's a computer to automate Tesla coil tuning.
In the features list it advertises that 'The coil sweeps through the given range and finds the resonant frequency'
I came across it here: https://highvoltageforum.net/index.php?topic=188.0
Perhaps this could be of some use to automate the tuning process and reduce manual effort each time? I imagine that pressing an 'autotune' button would be much simpler and less error prone than trying to manually tune.
he has done great stuff.
creating huge currents and then interupting it.
but at very low speeds.
the sweep to get the Fres isn't new to me.
Maybe in the future I can use something similar.
A slayer exciter does also auto tune.
but first I need to get all the coils tuned with the right capacity.
Since I not only have a dual resonant tesla coil, (which is easy) but also an impulse coil and an extra coil to tune...
for the L1 impulse duration parallel capacitor, I again need a parallel series combination.
to get high voltage capabilities.
In theory the 2000Vdc capacitor can handle around 630V ac at 50 c/s
and at higher frequencies it can work at higher voltages.
but i want it to be able to handle the 3500V of the max impulse voltage, although I might not be able to achieve max impulse voltage.
so series caps, which also reduces the capacity.
then there is the impulse current. which also is high, so I need parallel caps also.
this also increases the capacity again.
I think the amount of capacity should be tuned without a load on L4. only the rectifier and dc caps attached to L4.
then slightly detuned above the frequency, L4 breaks up in 4 cycles again, which can be measured to determine the impulse duration.
to get high voltage capabilities.
In theory the 2000Vdc capacitor can handle around 630V ac at 50 c/s
and at higher frequencies it can work at higher voltages.
but i want it to be able to handle the 3500V of the max impulse voltage, although I might not be able to achieve max impulse voltage.
so series caps, which also reduces the capacity.
then there is the impulse current. which also is high, so I need parallel caps also.
this also increases the capacity again.
I think the amount of capacity should be tuned without a load on L4. only the rectifier and dc caps attached to L4.
then slightly detuned above the frequency, L4 breaks up in 4 cycles again, which can be measured to determine the impulse duration.
due to the high frequency of the L1 impulse,
I can only use np0 ceramic capacitors!
I can only use np0 ceramic capacitors!
I connected the 2 rectifier diodes to L4, and 2x 10uF 630Vdc capacitors. (center ground)
And I am now able to tune again to a perfect sine on L2 L3 and L4.
input power is very low, with this unloaded test. 0.37A x2 4.5Vdc=3.33W
I expected the frequency to go up, but luckily it is tuned down.
So probably when I connect the load (resistive lamp), the frequency will again drop down.
The question now is... how much capacitor do I need to add to L1 to get the impulse duration right?
It now is around 400nS
L3 parallel capacitor: 11nF
L2 series capacitor:134nF
L1, no cap
input power 3,33W
theoretical output power 500W (cyan), this is peak to peak, so around 180W rms or so... but still theory
resonant frequency 133.1 kc/s
yellow= L2 voltage (impulse is absorbed at perfect resonant tuning)
green L4 current, around -13 degrees phase shifted
orange L4 voltage right before it enters the rectifier diodes.
purple, mosfet switch signal
And I am now able to tune again to a perfect sine on L2 L3 and L4.
input power is very low, with this unloaded test. 0.37A x2 4.5Vdc=3.33W
I expected the frequency to go up, but luckily it is tuned down.
So probably when I connect the load (resistive lamp), the frequency will again drop down.
The question now is... how much capacitor do I need to add to L1 to get the impulse duration right?
It now is around 400nS
L3 parallel capacitor: 11nF
L2 series capacitor:134nF
L1, no cap
input power 3,33W
theoretical output power 500W (cyan), this is peak to peak, so around 180W rms or so... but still theory
resonant frequency 133.1 kc/s
yellow= L2 voltage (impulse is absorbed at perfect resonant tuning)
green L4 current, around -13 degrees phase shifted
orange L4 voltage right before it enters the rectifier diodes.
purple, mosfet switch signal
Sounds like a promising development. I'm very interested to find out what happens when you connect a load.
I've been doing a lot of research into parts lately, and I came across a new SiC 1700V MOSFET from Microchip, part number MSC035SMA170B4. The price is good ($30) and the specs looks great - 1700V, 35 milliohms RDS(on). Manufacturer page link: https://www.microsemi.com/existing-parts/parts/152458
I also found a nice gate driver, the Infineon 1ED3124MU12HXUMA1. Voffset 2300V, Viso 5700V, Rise/Fall time 15ns. Manufacturer page link: https://www.infineon.com/cms/en/product/power/gate-driver-ics/1ed3124mu12h/
I've been doing a lot of research into parts lately, and I came across a new SiC 1700V MOSFET from Microchip, part number MSC035SMA170B4. The price is good ($30) and the specs looks great - 1700V, 35 milliohms RDS(on). Manufacturer page link: https://www.microsemi.com/existing-parts/parts/152458
I also found a nice gate driver, the Infineon 1ED3124MU12HXUMA1. Voffset 2300V, Viso 5700V, Rise/Fall time 15ns. Manufacturer page link: https://www.infineon.com/cms/en/product/power/gate-driver-ics/1ed3124mu12h/
evostars
3kv 68pF caps for tuning L1
« Reply #349, on January 6th, 2022, 12:20 PM »Last edited on January 6th, 2022, 12:25 PM
ordered a bunch of these caps, 68pF 3kV 5% to tune the impulse on L1.
one tuning trick is, to do an FFT on L2 with the impulse on top of the resonant sine.
If it is a proper harmonic, it should show on the fft.
https://nl.mouser.com/ProductDetail/75-564RC0GAA302EL680
before I tuned L1 with polypropylene wima caps, but they dont work with those fast impulses, the speed is to high for that dielectric. thats why I turned to ceramic C0G (NP0)
if the 68pF is still to much, I can put 2 in series, and get 34pF.
adding capacity parallel to L1 lowers the resonant frequency, the impulse is an resonant halfwave, so it is prolonged (and lowerd in voltage, so the power supply can be pushed higher)
the cap should be placed close to the source/L1 keeping the leads short to L1.
a proper ground for L1 is also essential, keeping the voltage steady at ground level creates a proper impulse.
one tuning trick is, to do an FFT on L2 with the impulse on top of the resonant sine.
If it is a proper harmonic, it should show on the fft.
https://nl.mouser.com/ProductDetail/75-564RC0GAA302EL680
before I tuned L1 with polypropylene wima caps, but they dont work with those fast impulses, the speed is to high for that dielectric. thats why I turned to ceramic C0G (NP0)
if the 68pF is still to much, I can put 2 in series, and get 34pF.
adding capacity parallel to L1 lowers the resonant frequency, the impulse is an resonant halfwave, so it is prolonged (and lowerd in voltage, so the power supply can be pushed higher)
the cap should be placed close to the source/L1 keeping the leads short to L1.
a proper ground for L1 is also essential, keeping the voltage steady at ground level creates a proper impulse.