After further analysis, I have finally accepted that once you are past
the break-even point, it doesn't matter what happens. I use to think
that having alternatives in different phases of their life would affect
the analysis.

I talked a little more with the service department, and it is $350 to
replace the secondary heat exchanger, or $800 for primary and
secondary. I did not think there was that much more labor to replace
both. In any case, for my breakeven analysis I will assume $800 since
the primary is likely to fail before the break-even date. So given the
options of $800, $2,600, 5% interest rate, and $200/yr savings, the
break-even is 11 years. Less if the old furnace needs more parts
replaced.

If the old furnace has to be replaced after 1 year, then it is
self-evident that I should not have replaced the heat exchanger. If the
old furnace lasts to 1 year before break-even, then I would have to
spend $2,600 in year 10 to get a new furnace vs. $0 for a 10 year old
furnace. Seems like I am still better off with the new furnace today.
What about year 5? $800 today and $2,600 in year 5, or $2,600 today.
I'm still better off with a new furnace because I have saved $200 per
year in fuel costs.

The only time reparing the old furnace seems to make sense is if the
new furnace has to be replaced before the break-even date. With a 10
year parts and labor warranty, that scenario would have a $0 cost.

I plan on living in the house for at least 8 years. Then the youngest
will be in college, and we will think about downsizing. I'm guessing at
least 10 years before we move, if ever.

The only remaining variable that could work against replacing the
furnace is whether the fuel cost savings are realized. Assuming a
marginal 2% increase in efficiency (vs. a possible 4%) and the DC
blower, the $200/yr should be achievable.

Thanks for helping me to clarify my thinking.