So I started looking at other options. Si570 looked nice, but here the lower limit was the problem: 10 MHz. Even with divide by 4, that knocks out 160 meters, a band I am very interested in lately, and that seems to sound especially good in a direct conversion receiver.
Once again, the controversial Si5351 was calling my name. It would go down to 8 kHz and up to 160 Mhz. Woo Hoo! If I could build a divide by 2 IQ VFO, I could cover 160-6 meters.
Here is the basic idea. From:
http://www.markimicrowave.com/blog/2015/04/top-7-ways-to-create-a-quadrature-90-phase-shift/
With lots of hardware help from Todd VE7BPO, and software help from Tom AK2B (wizards both), I got my Si5351 divide by 2 circuit working today. You can see the resulting I and Q in the picture at the top. But I am discovering that there may have been wisdom behind those divide by 4 circuits. My opposite sideband suppression isn't as good with this /2 scheme as it was with the AD9850 divide by 4. I'm still trying to figure out why. I may have to go back to divide by 4. Stay tuned.