There has to be a sweet spot where not enough voters are worse off to scuttle it, and enough voters get better that they show up to support it.

There doesn't have to be any such thing. And there's a good chance that there isn't one, given how this is a policy that has resisted adoption for so many decades.

Again, you have three policy impacts that you have to balance. You have to set the level of coverage, figure out what you're going to do with provider compensation, and determine how you're going to pay for it. It's not possible to set those three things in a way that doesn't make an awful lot of voters unhappy. Set the level of coverage lower than "above average," and half your voters end up seeing their coverage get worse [I know, it's not a normal distribution, should use the mean, but we're just chatting here]. Slash provider reimbursement rates, and a huge chunk of voters lose their jobs or get pay cuts. Do neither, and the amount of money you have to raise through taxes becomes politically unpalatable.

Is there a sweet spot in there? I kind of doubt it. The problem - not unique to this issue - is that humans generally get vastly more unhappy when they suffer a loss of what they have than they get happy when they get a gain of something they don't have. Something as massive and consequential as switching the health care system affects so many people both positively and negatively that you can't really avoid creating large pockets of people who will come out losing in the switch. And those people will be vastly animated against your project, much more than the people who will gain will support it. Our political system is geared towards weighing intensity of position, not just breadth of support or opposition for a proposal. So even if you're making more voters better off, it still becomes a political impossibility.