I had tried a small (3/4" diameter) yellow core that I thought was suitable, but it hardly registered anything on the ammeter and the core itself got quite warm. Apparently it's not the yellow -6 mix.
Anyhow, I found a (marked) T140-61 core in the toroid drawer and it works great.. except that to get a 2:1 current ratio, I ended up with a 2 turn primary and a 3 turn secondary... each pass through the center of the core is counted as one turn. I would have thought it would require 4 turns.
I rigged up a toggle switch with short leads for x1 (transformer not in circuit) or x2. So 100W is 1.4 amps on x1 and 1KW is 4.5 amps, 2.2 on the x2 range... covers most of my 50 ohm power measurement needs
The dummy load itself is ten 470 ohm 10 watt wirewound resistors, fan cooled, with an air variable in parallel to tune out the inductance on the higher bands. DCR is 48.7 ohms. It's nearly a perfect match up to 10m when the variable is set for minimum SWR (barely moves the needle on my HM-102 bridge). Who needs noninductive Globars anyway
Edit: even better match than I remembered