GDT Discussions
July 17, 04: I'm zeroing in on a bi or tri-filar and screened, circumferential, overlaid design. The strands will be twisted together to two turns per inch. Screened by using tinned cu braid and pushing the twisted wires through it prior to winding onto the core. John C.
July 12, 04: Iron powder cores are definitely out. The permeability is low and the high AC current levels will heat them like crazy. The best recoverable cores are either the actual transformer cores *as long as they aren't gapped* or the toroidal ferrite cores one typically sees threaded over some of the leads inside CRT monitors. Best results for the winding should be obtained if the windings are not only n-filar, but also twisted together as it is more important to get high coupling between windings than it is to get high coupling between windings and the core. Malcolm.
July 12, 04: A good way to
do it is use a small ferrite toroid, wind with 5 wires at the same time. When
you are finished connect windings in series to get the windings with more turns.
Turns need to be calculated from
N = VOLTAGE/4 x FLUX DENSITY X FREQUENCY X CORE AREA (IN MM2), flux density
would be about 0.2. Filter toroids are not good, you need a power ferrite
like Seimens n27 although n38 will also work. Phillips 3C8 etc. A smaller toroid
is actually better, about 14mm works well. Use grade 3 PI magnet wire
unless you need to be able to connect the drive side to the user interface, then
use triple insulated wire for the primary wire. Are you sure that is what
you need though, most of mine end up 1 : 1 +1? maybe increasing the primary side
supply voltage is easier. George T.
July 14, 04: Any simple
way to distinguish a powdered iron from a ferrite toroid core? Dave.
Answer 1. One way to distinguish powdered
iron from ferrite is the Curie-temperature. This is the temperature, the
permeability drops to 1. For RF-ferrite this temperature usually is in the
range from 160 deg C/320 deg F to 200 deg C/392 deg F. Powdered iron changes its
permeability at much higher (e.g. 750 deg C/1382 deg F) temperatures.
So you could wind several turns of teflon (PTFE) insulated wire onto the core
and measure the inductance. Then place it in an oven and heat it up as
high as possible. Remove the core and measure the inductance again. If it is in
the same order, you probably have a powdered iron core. Cheers, Herwig.
Answer 2. Another way is to simply wind a few
turns of wire, measure the inductance and calculate the induction factor (Al)
by: Al = L/N^2, you will typically get >1000 nH/turn^2 for ferrites and
just a few 10's for iron powder. The permeability of iron powder toroids
is inherently low - they have a distributed airgap and are specifically designed
to be used for inductors which carry high levels of DC without saturating.
Malcolm.
Answer 3. All of the
powdered iron cores I've seen are relatively soft (iron powder in a plastic
binder) and can be scratched with an awl or a scribe without hurting the
magnetic properties. Ferrite materials are glass-like, very hard, and easy to
distinguish for that reason. I'm pretty sure ferrite cores will be more dense
that powdered iron and that
measuring the density would be another method of ID. I'd start by trying the
scratch test. Either core may have a coating on the outside so you have to make
sure you're through that. Of course, if you have some to spare you can
break one with a hammer and look at the texture of the cross section. Ed.
Answer 4. For my 1:2:2
transformers, I do it the following way: For example, if I want a 10:20:20
turn GDT, i simply wind three strands of 20 turns and then cut the primary in
half and then parallel both 10 turn primaries. I can maintain a measure
coupling of greather than 0.99 when I do this. Just one thing. Once I
have 10 turns on the core, I'll pull some slack on the primary winding and
twisted maybe 4 inches of that piece outward. Then I continue winding the
additional 10 turns. It also helps to put tape on the primary to make
paralleling easy. Once you wind it, it may be difficult to determine which end
is which when you start paralleling. Dan.
July 7, 2004: Sorry, a
typo in the previous post, the mutual inductance between the filaments should be
M = pi * mu * Xa * sin(alpha) * sin(beta)
* sum{ for n=1 to infinity;
(Xa/Xb)^n / (n*(n+1))
* Pn( cos( alpha)) * Pn( cos( beta)) * Pn( cos( gamma))}
Funny how you only see these after hitting the send key. Paul Nicholson.