GDT on an EE or EI core / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

GDT on an EE or EI core

1 2 next

Move Thread

LAN_403

Dr. Dark Current

Tue Sept 15 2009, 07:47AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

What's the problem with a GDT on an EE or EI core?

I'm gonna try one for my upcoming SSTC and post the results.

HV Enthusiast

Tue Sept 15 2009, 11:26AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Not necessarily a problem, but you won't get as good leakage inductance as you would with a toroid core. It all depends on your particular application, switching frequency, etc...

Actually, the best cores for lowest leakage inductance would be a Pot core or multi-aperture core.

Multi-Aperture cores work great when highest performance is required, especially if you use a coax. as the winding with the shield as the primary and the internal conductor as your secondary. However, you won't get many turns with a multi-aperture unless its huge.

Which is probably why a toroid is still practically the best.

Dr. Dark Current

Tue Sept 15 2009, 01:22PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Dr. GigaVolt, I thought the leakage inductance is a function of physical placement of the coils in relation to each other. So when an EE core has the same cross-section as a toroidal core, wouldn't the leakage be the same with the same turns and winding technique?

BTW, I'm now "testing" a core from a PC SMPS, with 3 turns on it and two UCCs have no problems driving it at 250-500kHz and 15V (no FETs connected so far).

HV Enthusiast

Tue Sept 15 2009, 03:24PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Dr. Kilovolt wrote ...

Dr. GigaVolt, I thought the leakage inductance is a function of physical placement of the coils in relation to each other. So when an EE core has the same cross-section as a toroidal core, wouldn't the leakage be the same with the same turns and winding technique?

BTW, I'm now "testing" a core from a PC SMPS, with 3 turns on it and two UCCs have no problems driving it at 250-500kHz and 15V (no FETs connected so far).

Leakage is a function of the geometry of the windings with respect to the core.

GeordieBoy

Tue Sept 15 2009, 04:18PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

There is nothing particularly wrong with making GDT's or pulse transformers using E-core sets. It's just that you usually get a more compact and better performing design with a suitably sized toroid.

If you wind your secondary turns over the full circumference of the toroid, and then wind your primary over the top of this, you get very high coupling coefficient as it leaves very litte room for the magnetic field to leak out of the core.

With an E-core set, you can typically only wind on the centre limb, which means that the cross-pieces of the core set and the outer limbs are usually not covered with windings. The core also usually comes in two halves. There is always some non-magnetic gap at the mating surfaces due to machining tolerances, improper clamping, poor alignment, glue etc! Some of the magnetic field will always stray from the core as a leakage field in air around the centre limb. That is why you often see flux-bands on high power ferrite transformers.

A Mn/Zn toroid is one continuous loop of high permeability material, that doesn't need a bobbin, and you can cover it's entire surface with windings. In that respect it's hard to beat!

-Richie,

Dr. Dark Current

Tue Sept 15 2009, 04:52PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Thanks for the replies.

I wanted to use E-core because you can solder the wires to the pins at the bottom which is more convenient, and also easier to just solder into a PCB without any clamping "equipment". It's also easier to wind...

HV Enthusiast

Tue Sept 15 2009, 05:00PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Dr. Kilovolt wrote ...

Thanks for the replies.

I wanted to use E-core because you can solder the wires to the pins at the bottom which is more convenient, and also easier to just solder into a PCB without any clamping "equipment". It's also easier to wind...

You can do that with a toroid too. Just get a carrier, or make one using a piece of FR-4 or something and epoxy the core / windings to the FR-4 with pins going through the board.

Sulaiman

Tue Sept 15 2009, 07:28PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Industrially I see quite a few toroid cores used as gGDTs
the minus point is winding difficulty.
More common on older equipment are pot cores
but EI and EE are the most common in modern equipment;
the ferrite cores join together with very small airgaps
and the cores are easier/cheaper to mass produce.

Even a thin layer of adhesive in the gaps should be avoided
clamp the cores together.

Where safety/insulation allows multi-filar windings are best due to low leakage,
but industrially most GDTs are wound on separate coil formers with primary to secondary insulation in mind and suffer a little leakage inductance.

Dr. Dark Current

Thu Sept 17 2009, 08:16PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Ok, here are the promised results.
The GDT is wound on a little transformer core from a 200W PC PSU, 4 turns of screened stereo audio cable.
Both FETs connected, each with 2.5nf gate capacitance, NO gate resistors.
Driver 2x UCC push-pull

Gates at 250kHz, 1us/div:

Gates at 500kHz, 500ns/div:

(the trace is fat because the supply transformer for my gate driver can not cope with the power consumption

)

Is this good?

brtaman

Thu Sept 17 2009, 08:30PM

Registered Member #2161 Joined: Fri Jun 05 2009, 03:36PM
Location:
Posts: 247

Very nice waveforms! No resistors!? :D

Only thing is that the 250kHz waveform shows signs of slooping (not a great deal though), meaning inductance isn't going to be optimal under this frequency, what frequency will you drive the coil at? If lower then maybe a turn or two extra. Otherwise, perfect waveforms at 500kHz!

1 2 next

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