Testing Unun losses

In his recent feedback, Richard, VK3TXD, suggested I measure the loss in the Unun built with Jaycar LO128 core by making a second one and wiring them back to back.

On the spectrum analyser, I set the range from 5Mhz to 30Mhz and the plot shows a very slow loss over the range. Note that the tracking generator is at -10dBm.

In the middle of the frequency range the loss, going through two transformers is 3.26dB. 1.63dB for one Unun. Richard commented that 2 turns of primary winding is not good so I added a turn. Here's the plot.

The loss in the middle of the range, going through two transformers is slightly less at 3.24dB. In my test, hardly measurable. Of course with an extra turn on the primary the impedance transformation is much less so more turns on secondary will be needed.

Perhaps I'm not measuring this correctly but I'm not seeing the dramatic problems that were predicted.