On 2/24/2018 6:22 PM, Riki, K7NJ wrote:
Rick,
According to Sevick's "Transmission Line Transformers" (latest 5th edition), there is a technique for winding transformer cores with parallel lines in a way that doesn't create magnetic flux in
the core(s) thereby reducing magnetic core losses (theoretically) to zero. Sevick shows how such windings are interconnected using integral multiples of them such as 2, 3 or 4 parallel conductor
I don't follow what you are saying here. Can you give us
an exact quote from Sevick that you are referring to?
This sounds like it is referring to Guanella transformers.
But they are not limited to 1:2, 1:4, 1:9, 1:16 ratios,
and they do not have zero magnetic core loss. Their
claim to fame is that they theoretically have
infinite bandwidth, which you don't need.
I have demonstrated 1.8 to 54 MHz with a Ruthroff
design, so there is no need to resort to a Guanella
design, especially if you only need 1.8 to 7.3 MHz.
Granted, without the limitation imposed by the above technique, then many
different impedance step-up ratios may be used, some of which are discussed in
Sevick's book. I would welcome information on tested designs that includes
insertion loss (as well as frequency coverage and power handling capability).
Do you have data like this for any of the designs that you've mentioned?
73,
Riki, K7NJ
If you are trying to step up to 450 or 600 ohms, one
important thing to do is to use LOW permeability FERRITE,
not powdered iron, cores, such as Fair-Rite Products
61 with a permeability of 125. I've already mentioned
the other important thing to do which is to separate
the impedance transformation function from the balun function.
If you follow these to steps and use a sufficiently
large core stack, you should be fine in terms of
core losses.
It is very easily to tell if a transformer is lossy;
just run a kW of power through it for 5 minutes and
see if gets hot. My transformers don't get hot.
73
Rick N6RK
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|