What are the tricks for winding a tight air-core inductor?

[T3sl4co1l]

Sounds about right.  And the above part, probably isn't rated anywhere near enough pulse to be considered reliable, but yeah, same idea, who cares.  In this case, they're oversized and under-volts, so, series combinations will do.

Inductor type doesn't matter based on given constraints (i.e., not enough of them!).  Toroid has lower external field.  Solenoid is easier to make, and, I don't know how Q compares.  But Q isn't very interesting for pulsed operation anyway.  DC current ratings are irrelevant (in contrast to arc starter application!).  The surge network needs a lot of resistance already, so the relatively thin wire isn't a big deal (as long as it doesn't fuse, lol).

I was going to do a EFT generator, eventually.  Figure C0G chip capacitors, SiC MOSFETs, and enough in parallel (about 8x) plus a transformer to match back to 50 ohms at enough voltage.  Making a good enough quality TLT isn't going to be easy, but feels like the better option compared to a cascode switch.

Or just get a hydrogen thyratron and slam-dunk it that way.

Tim

[tsat]

Hi, I have recently purchased a multitester for EFT/Surge tests thanks to a national call that support the development of small companies by funding with the 50% of equipment cost.
Even so was not easy to buy it so I didn't purchased any coupling/decoupling networks. I purchased a used EFT capacitive coupling clamp for EFT and I am thinking to build some CDNs for surges.
To start I would like to build a decoupling network for I/O signals (figure 9, EN 61000-4-5) so does anybody knows what kind of chokes I need to use ? Are there any COTS available or shall I make one by myself ?
Unfortunately even if I have access to CDNs these are completely potted so there is no way to check the components.
Then I need to built the coupling part so a 0.5uF suitable to be stressed up to 5 kV. Any ideas?
I am not very familiar with high voltage circuits that is why I am seeking help, I am an EMC/Wireless test engineer.
BR,
Stelios - Greece   

[jonpaul]

In 1980s for arc lamp series inj ignitors We got HPTZ square or rect wire from Belden, Phelps Dodge etc.

Mandrill wound on a LARGE winding machine like Micafil or Adama Maxwell.

For the fastest high current pulse switch, The DNA, AEC, LLAN used triggered spark gaps , like the EGG ceramic types.

Best pulse caps were from Maxwell,

Jon

[T3sl4co1l]

Hi, I have recently purchased a multitester for EFT/Surge tests thanks to a national call that support the development of small companies by funding with the 50% of equipment cost.
Even so was not easy to buy it so I didn't purchased any coupling/decoupling networks. I purchased a used EFT capacitive coupling clamp for EFT and I am thinking to build some CDNs for surges.
To start I would like to build a decoupling network for I/O signals (figure 9, EN 61000-4-5) so does anybody knows what kind of chokes I need to use ? Are there any COTS available or shall I make one by myself ?
Unfortunately even if I have access to CDNs these are completely potted so there is no way to check the components.
Then I need to built the coupling part so a 0.5uF suitable to be stressed up to 5 kV. Any ideas?
I am not very familiar with high voltage circuits that is why I am seeking help, I am an EMC/Wireless test engineer.
BR,
Stelios - Greece   

Just has to pass the calibration steps (following section).

Typical values are going to be in the 100s uH, and rated... more than a few amperes peak, of course.  Recommend running a simulation to get a feel for the waveforms, current and voltage magnitudes, and what ratings you'll need.  Likely need either air-cored or some big powder cores (or ferrite if you like, but it's probably going to be rather big, or maybe even laminated iron).  Inductors will most likely be larger for the higher impedance tests (smaller C and added resistance).

Tim

[tsat]

Thank very much Tim & Jon. I will give a chance to myself.
Stelios
 

[jonpaul]

CopperCone!

>>why is the rod inductor better then a torroid for pulse generation?>>

EXCELLET question.

In HV, pulse and other apps the insulation and volrag egradient are critical.

A toroid is wosrt as every turn is clost to the core, so S><F voltage must be withstood by S turn and F turn to core insulation.

A solenoid air cosr is best  for capacitance, SFR and insulation

next best is rd core solenpoid, SINGLE layer perfect lay.

See Nikola TESLA nes 1889..1900, ALL of his HV coils were air core, perfect lay solenoids.

With a rod core, a heavy insulating tube is easuly slid on before the cois is applied.

Finally a solenoid is MUCH easier to eind than an toroid.

We designed and manufactured may HMI, Xenon and other arc lamp ignitors in 1970s..1900 like this

 HV lamp series injection ignitors use that design  to this day.

Cinema lights to 12/18/14 KW with ignitors 35..75 KV are by  IREM, BAG Turgi, Ianairo, ARRI!


An ABSOULTELY FANTASTIC DAY to all!

Jon

[coppercone2]

I don't get the S><F thing you wrote.

Do you mean that the wire on the inside of the doughnut is basically disfavorably positioned compared to wire on a cylinder that is uniform ? Like I get that on a rod core, the end is far from the start, so it is highly isolated, but in a torroid the end is near the start where there is dielectric problems?

And that the wire wrapped around the interior of the doughnut is closer to each other then the wire on the other side since its like folded. But I assume with a proper core - spacing - gauge ratio the problem is mitigated, but I guess every slice is offset kind of from the ideal and seems imbalanced. The bending of the solenoid makes me think of a cusp related problem. Gradient on the insulation or gradient of insulation materials might help optimize to make it stronger to breakdown in the middle but not have so much capacitors on the edges. Like a semi air core torroid built around something like a chirstmas star acrylic ornament to have dielectric resistance in the middle but not the edges.

[T3sl4co1l]

I think: Start, Finish, and everywhere inbetween.  Cores are notably conductive, so when you need say 10kV isolation, it's a big deal.

Toroids probably aren't much better than solenoids even when air core, where pulse applications are concerned; unless you need a compact construction with shielding to nearby stuff, in which case the self-shielding nature of the toroid is of course helpful.  But you still can't close it very tightly (start to finish distance) because of insulation still, which leaves the gap open, emitting a field equivalent to a solenoid the size of the gap.

It's also possible to make a solenoid with a "shingled" arrangement, i.e. wind 10 turns forward, 7 turns backwards, repeat ad nauseum; this gets the inductance and density of a multilayer inductor, but with ~full length voltage handling, and a fraction of the capacitance of a multilayer winding.  Alternately, "pie winding", making blocks or banks at a time, and stacking those up to the same effect.

Tim

[jonpaul]

Solenoid .....even voltage gradient over the legnth

eg 10 kV pulse 10T

START (S) = 0  turns/V

1/1kV
2/2kV....
10 10kV

Toriod: S and F are adjacent so highes tstress and EVERY turn is clost to core.

j