On Fri, 4 May 2007 09:49:48 +0000 (UTC), Tehrasha Darkon
wrote:

First I wound the core in a bifilar manner with the leads being labeled
A1,B1 and A2,B2. Then I measured the input impedance across A1 and A2
leaving B1 and B2 unconnected. Then placed a 200 ohm resistor across A1
and B2 while tying B1 and A2 together, and measured across A1 and A2
looking for 50 ohms (4:1).

I have also seen a reference to a K value for an indication of
broadbandness. K = 1 - L(short)/L(open)

Here are my results...

8 bifilar turns of teflon coated AWG 16, on Type 43 core.

Hi Teh,

Your numbers suggest Type 75 instead, OR some other mechanism at work.
Given your lead description above is fine, but could have been poorly
implemented (in text, it is difficult to sort those things out).

From what I have read, the open reactance should be at least 4 to 5 times
as high as the normal drive impedence. In this case, it is over 12
times the 50 ohm input at 30Mhz and over 6 times at 54Mhz.

So did I do good? Am I way off?
How can I tell if/when core saturation will be a problem?

The ratio is as good as any. More important is how much power the
BalUn can withstand. Your last question gets into the common
misconception of "core saturation" which is typically a consideration
for the winding linked magnetics not found in this application. BalUns
are not classical power transformers that are so common to 60Hz AC.

The choke BalUn (which is what you are building) employs the lossy
resistance of Ferrite to establish the isolation you want in a BalUn.
There may be some reactance achieved along the way, but it is not
usually a design goal and these class of Ferrites would be a poor
choice in achieving that. If you note that your data describes a
very, very broad curve; it demonstrates a very poor Q which is its
principle attraction (broadbandedness).

As a simple test, pass one very short wire through one bead and
measure its Z (give us a complete description of resistance AND
reactance) over frequency. This will confirm what type material you
are using and will remove unnecessary complexity. Type 43 should show
a frequency peak roughly ten times above your current data's frequency
peak.

73's
Richard Clark, KB7QHC