Calculating self-resonant frequency of HV Transformer? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Calculating self-resonant frequency of HV Transformer?

first previous 4 5 6 7 next

Move Thread

LAN_403

Ash Small

Wed Feb 20 2013, 10:10PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Sulaiman wrote ...

If your insulation is (say) one and one eighth of a turn
the return wire can run between the overlap.

Stagger each return wire by say 3/8 turn each layer and the nett effect will be
9 layers of insulation per 8 layers of windings.

Some insulation overlap per layer is required anyway so you loose very little space.

I keep mentioning this because it's simple and it works !

I see, you only have one turn of insulation. I was planning on using multiple layers of kraft paper, giving a combined thickness of 1mm.

0.5mm of oil impregnated kraft paper should be plenty though for 1.25kV (If 1.5mm air is the minimum for 1.25kV, and, as Steve suggested, oil has 6 times the insulation of air, 1.5/6=0.25, so 0.5 has a safety margin of 2. Or will I need more at 25kHz?).

EDIT: I could stack the kraft paper before winding, I suppose, which would amount to what you are suggesting.

Newton Brawn

Thu Feb 21 2013, 04:10AM

Registered Member #3343 Joined: Thu Oct 21 2010, 04:06PM
Location: Toronto
Posts: 311

I think that wind a 0.2mm wire over a 1mm or 0.5mm insulation is someting really unstabble.
Not wire tension enough to old the paper....
The winding after 4 or 5 layers the things becames wild.
The transposing of the return wire make the winding more critical.
My sugestion:
1-Make a new drawing with more info, add the dimensions.

2- Try to make one pancake, paper insulation, just to feel the stability of the winding.

Polyester dielectric constant =2.1, oil impregnated kraft could be 5 ~ 6

Paper margin more than 4mm is required.

Considerations about interlayer insulation:
If you consider a small air bubble traped between the wire and insulation, a 0.5mm Mylar insulation may effort a test voltage of 3200V peak, Air or oil immersion . (corona starting voltage =3200V peak)

However,

If you consider a small air bubble traped between the wire and insulation, a 1.5mm oil impregnated Kraft insulation may effort a test voltage of 3200V peak........(corona starting voltage =3200V peak)

The Tony drawing appears that are based on the above parameters

Ash Small

Thu Feb 21 2013, 10:20AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Those are some very good points you've raised, Newton.

The transformer will be vacuum de-gassed, so I should avoid the air bubble issues.

The points you raise regarding mechanical integrity have got me thinking about using polystyrene again, and although capacitor grade polystyrene is no longer manufactured (Bayer has ceased production), it does have advantages over polypropylene, as long as I can keep the temperature down to ~85 degrees C. (ODAF cooling)

It appears from my research that polystyrene doesn't suffer dielectric degradation at high frequencies.

I found this paper, produced ten years ago for the USAF by QinetiQ, which verifies this (page 19).
]gettrdoc.pdf[/file]
It also states that boiled coconut oil is better than standard transformer oil (page 6).

It would appear that polystyrene is the dielectric of choice, and that it has only been replaced by polypropylene, which is not as good, because polypropylene can be used up to 105 degrees C.

I think I can form the polystyrene insulators into the form suggested by Sulaiman above it's glass transition temperature, then, when it cools, it should provide the structural integrity required, thus avoiding the issues raised by Newton.

I also found this:

1361442012 3414 FT150013 Dielectric Strength

Newton Brawn

Thu Feb 21 2013, 09:42PM

Registered Member #3343 Joined: Thu Oct 21 2010, 04:06PM
Location: Toronto
Posts: 311

Ash:

Big difference between Mylar (polyester of Dupont)
and
polystyrene.

Mylar is THERMOSET material for use up to 150C and polystyrene is THERMOPLASTIC material for use up to 65C.....

Mylar can work without impregnation or impregnated with oil. One time I made a oil impregnated capacitor for 4000V peak using a 0.1mm thich Mylar as dielectric. It works for more than a year.

Such capacitor could be made easely made in a DRY construction, without oil. Then the dielectic thicknes should be 1.2mm (6 x 0.2 mylar sheets).

I have made several bobin forms just coiling the Mylar film over a mandril. Of course the sheets are painted before with very thin layer of epoxi (one day - long time setting epoxy) The forms are circular or rectangular shape.

Ash Small

Thu Feb 21 2013, 10:12PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Not much difference in dielectric constant @1mHz, Newton. 2.5-2.7 for polystyrene, against 3 for mylar. Polystyrene should be adequate for this application. Dielectric constant may even be slightly better @25kHz.

The thermoplastic quality is advantageous in this application. How would I use mylar to form the insulating spacers that Sulaiman has suggested?

Polystyrene capacitors are rated for 85 degrees C (the sheet you provided says 60-65, the glass transition temp is 100 degrees C). I think I only need to dissipate ~60 Watts. If I use ODAF (oil directed forced air) cooling I don't expect it to get very hot at all.

I'm still condidering alternatives, though.

EDIT: Also, how do I eliminate air bubbles/voids using mylar?

Newton Brawn

Thu Feb 21 2013, 10:58PM

Registered Member #3343 Joined: Thu Oct 21 2010, 04:06PM
Location: Toronto
Posts: 311

please delete this reply

Ash Small

Fri Feb 22 2013, 12:42AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Newton Brawn wrote ...

sorry, my server is too sloow to upload the pics

They were *.jpg and not *.bmp?

Newton Brawn

Fri Feb 22 2013, 02:07AM

Registered Member #3343 Joined: Thu Oct 21 2010, 04:06PM
Location: Toronto
Posts: 311

jpg

Newton Brawn

Fri Feb 22 2013, 02:26AM

Registered Member #3343 Joined: Thu Oct 21 2010, 04:06PM
Location: Toronto
Posts: 311

Hi !
Here is a pulse trafo with 2 pancakes,
21 volt per turn,
The primary is 420Vpeak, 20 turns 2 wires #24AWG, linear coil, just one layer.
Secondary 6300 -0- 6300Vpeak in two pancakes. (6300Vpeak each pancake)
Each pancake has 300 turns of #32AWG wire. 30 turns per layer, total 10 layers.
Each wire layer lenght is 6.6mm, the interlayer insulation is 15mm wide, keeping 2 margens of 4.2mm.
The pancake form is 15mm wide and the wall of the form is 1mm thick .
The pancakes can be assembled and disassembled any time .
Dry construction.
Bmax = 0.3T, 500kHz or more.
The insulation is all polyester, glyptal clear vernish as finishments.
The bobin form is Mylar 0,1mm coiled to build up 1mm thick wall, glue with epoxi.
Please note that the begining of the pancake winding (brown wire) is the ct, that is grounded. I this way the insulation between the primary and pankake is only one turn of 0.1 + 1 mm polyester, that is good because the max potential is limited to 420V....

EDITED: 27/02 fixing some grammar and organizing the data.

Ash Small

Fri Feb 22 2013, 10:36AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

That's pretty impressive, Newton.

The main reason I want to use ~1mm between layers isn't primarily for insulation, it is mainly to reduce inter-layer capacitance.

I've not used any maths for this, though. I'm just guessing.

Maybe using more layers, with less turns per layer will give less capacitance, I don't know.

I'm guessing that the distance between plates and capacitance is not a linear relationship, and that capacitance falls off at the inverse square of distance, or a similar non-linear function. (See EDIT 1 below)

If I centr-tap and ground the secondary to the core, the maximum voltage anywhere will be ~25kV (except between the two ends of the secondary), and 2mm of polystyrene or 3mm of oil will be plenty of insulation. (as far as I can tell)

I wish to utilise the whole 'window', in order to minimise capacitance, and use 'pie windings' to reduce leakage inductance. Insulation isn't really an issue here, even for a 50kV transformer. Also, I wish to keep losses to a minimum by keeping Bmax to a minimum.

The concerns I have at the moment are these:

reducing inter-layer capacitance tends to increase leakage inductance.

As this transformer will be driving a CW multiplier, it will 'see' a capacitative load, which will resonate with the leakage inductance/stray capacitance.

This concerns me greatly.

This thread is primarily about self-resonance. The general consensus here on the forum is that I should reduce inter-layer capacitance as much as possible, and use pie windings to minimize leakage inductance.

I assume that centre-tapping and grounding to the core will reduce capacitance between the secondary and core also?

EDIT 1: It seems it is just a linear function of area divided by distance

, however, it would also appear that the dielectric constant also plays a part. It looks like using an insulator with a lower dielectric constant will also reduce the capacitance, so yet another reason for using polystyrene over mylar. Maybe I should consider using an insulator with an even lower dielectric constant?

first previous 4 5 6 7 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob