RE: Toroid cores for GDTs

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Mccauley, Daniel H" <daniel.h.mccauley@xxxxxxxx>

Skip,

We have evaluated these for use commercially in our projects and found
that the leakage inductance is much too high for driving devices with
large gate charge (capacitance) such as 40N60 IGBTs or similar.  They
work fine for smaller devices, but for larger devices commonly used in
SSTCs and DRSSTCs, they just don't work well.

Better off to wind your own.

Dan





Steve, I know EXACTLY what you are referring to.  The transformers
that I am talking to are EXACTLY what you describe.  These are
trifilar wound with VERY close windings on a ferrite toroid
core.  See my other reply.  For the switching power supplies I
designed, I originally used a GDT from Pulse Engineering.  These were
of the EE core design.  These had occasional problems.  I eventually
changed to a toroid trifilar wound toroid.  The leakage inductance
was about 1/10th of the EE core design.  All problems were
solved.  the ones in my link
http://www.lse.com.tw/upfiles/spec01125457782.pdf are EXACTLY what
you and others are describing as the better transformer.

Skip

At 08:03 PM 1/25/2006, you wrote:
>We need to drive several large MOSFETs at high frequency, so we need
>absolute minimum leakage inductance. We usually get that by
>forgetting we ever knew about "code" and making the transformer with
>a multi-filar winding of ordinary PVC insulated wire. If you're a
>radio ham you might know this as a "Transmission line transformer".
>You are limited to integer ratios (1:1, 3:2, 4:3, 2:1 etc.) but that
>is hardly a problem. Finally, using a large toroid core of high
>permeability ferrite lets you reduce the number of turns, reducing
>the leakage even further, while keeping the magnetizing current
acceptable.