Chris,
Note that the 9mm is a higher-pressure cartridge than the two that are behaving for you. If you avoid pressures higher than 1400×BHN, you won't have the pressure distorting the bullet base. For BHN of 10, that's 14,000 psi, which will run the 45 and 38 just fine but is pretty low for a 9mm that hasn't been set up for light target loads.
The problem with wheel weight metal, even when it was still plentiful, is there wasn't a set formulation, so it varied over time and with which metals company was supplying most of it for a given region of the country. Lyman measured its wheel weights as 95.5:4:0.5 lead:antimony:tin. But Jim Calhoun analyzed some in a lab in the mid-'90s and got 97.4:2:0.6, IIRC. That's enough difference to lose a couple of BHN numbers pretty easily.
So we don't really know what you had that was working well for you before. What we can do, though, is guess that if your pressure is in the 20,000 psi range, then when you divide that by 1400, you get a BHN of 14. Teracorp Magnum Alloy (Hardball Alloy at Rotometals) is 92:6:2, IIRC, and that should get you up to about BHN 16 and would be a fairly safe choice, I expect.
In the case of the recovered backstop lead, a lot of that will be 22 rimfire lead, which, like a lot of commercial swaged bullets, isn't very hard. Probably one or two percent Antimony. the
Excel sheet at the Castboolits forum suggests 1.5% antimony, and the rest lead would give you about BHN 10. Lead used to fill jacketed bullets usually has a little more antimony than that, but people shoot so many 22 LR rounds that it may be swamped out by the little slugs.
If the Antimony Man were still in business, I'd send you to his site to order. Meanwhile, both Midway and Rotometals have 70% lead and 30% antimony for about $25 a lb. You could buy one pound and melt with six pounds of range lead and add about 2 ounces of 95-5 lead-free-zinc-free plumbing solder and come pretty close to that Tercorp Magnum/Hardball Alloy. Increase the solder to 4 ounces for better ease of casting if you are willing to spend that much.
Note that 1400, above, is an easy-to-remember number, but it is close to exact. Richard Lee preferred to use 1280 to provide some wiggle room for shot-to-shot pressure variation.