On Tue, Jul 31, 2018 at 06:38 AM, Warren Allgyer wrote:

In actual operation these losses would be the best case and would be experienced only when the antenna is a resonant multiple of a half wave. The losses when not terminated in a 2470 ohm resistive load, which would be the case if the antenna were used on any band other than the single resonant band, are far higher. On the order of 8 - 10 dB.

Thank you Warren.  Your VNA and 'scope plots illustrate very clearly why in antenna literature, "all band" doesn't mean what most folks would have it mean - it almost always means "lossy".  Even commercial manufacturers have a difficult time producing a high transformation-ratio UNUN that operates across multiple octaves of frequency with acceptable losses - it's just not an easy set of design objectives to satisfy...  so for casual construction projects it a reasonable design constraint to limit (focus) yourself to making a transformer with minimal losses across ONE octave of frequency.  I.E 1.8~3.6Mhz; 6~12Mhz, and 14~28Mhz.  Even minimizing losses in the 14~28Mhz range can be challenging, due to the very high transformation ratio (1:49) - look very closely at your use case before choosing core materials -

Material #43 isn't  "the best" choice in many cases; in fact I try to avoid it whenever possible below 5Mhz and above 20Mhz because of it's high(er) losses. 
Material #61 yields lower losses than #43 for nearly every application in the 10~30Mhz; and Material #31 has lower losses than #43 in most applications at or below 5Mhz. 
In practical applications, the loss curves of materials #43 and #31 'cross each other' in the ~8Mhz range, so an UNUN | BALUN made for use on 40 Meters can utilize either core material and achieve a respectable loss figure, when all other design limitations are observed.

This is why I strongly recommend designing and building separate matching units for (80)/40/30 meters, and 20/17/15 meters - that way, each can be optimized for minimum losses in the specific band(s) it will be used on.

I would use matching techniques for 12/10/6 Meters which don't employ ferites or iron powdered toroids at all.  Air core coils and/or transmission-line sections are compact enough on these bands; yield very generous matched bandwidth figures; and have much lower losses when compared to the toroidal transformers [tuned and un-tuned] being discussed throughout this thread.

Cheers,
Chuck