From Webshed
Jump to: navigation, search

Testing low frequency losses in broadband transformers materials

In the cause of converting a Softrock Lite II receiver to work on the new 472kHz (660m) amateur band, I found I needed a broadband transformer with low loss characteristics at these frequencies. The recommended core material to construct broadband transformers at low radio frequencies is 75-mix ferrite, I had none to hand, so I've decided to measure up what I do have handy.

Methods

Construct a tri-filiar winding of 5 turns on both FT50-43 and FT50-26 cores, keeping the windings as close to identical as possible. Insertion loss (S21) between two pairs of the windings will be measured with a DG8SAQ VNWA. The third winding will be left unconnected.

Loss will be measured from 10 kHz to 100 MHz and plotted on a log frequency scale. A spot measurement of S21 at 473kHz will also be made.


Results

The results are not what I expected to see, I'd been lead to believe that 43-mix material wasn't much use below around 1MHz. and that 26-mix material would be significantly better. It should be noted that errors in S21 measurement are estimated at approximately +/- 0.25 dB.

FT50-43

Ft50-43.png

This material is just about usable at 473 kHz, there is essentially zero loss between pairs of windings. The frequency response of the material rolls of very rapidly below 300 kHz, so it's useless for the 136kHz band.

FT50-26

Ft50-26.png
I checked and rechecked this measurement because I didn't expect it to be so bad. 16dB loss at 473kHz is hopeless for my application.

Conclusions

I'll make my transformer from 43-mix material unless I can get some 75-mix soon. I'll post up measurements of the transformer I end up using. It may also be instructive to try other common cores powdered iron and see how they compare.