Agreed, and I also have yet to see an instance where productivity scaled 1:1 with time put in. So, it works out. :)
Suppose you start at $80k and work at the company for 5 years. After five years the strict 8-5 eng in your example is making $87k--barely keeping up with inflation--and the hard worker is making $101k. The total compensation difference over 5 years is $35k.
Unfortunately, real world compensation schemes often don't even work out as well as your example...
Working an extra 20% so that you are paid appropriately larger salary 19 years down the line is not a terrible choice if you really need the money. It is also possible you will earn true grade improvement along the way and get a one time additional bump of 5% for your effort.
Of course, if we reverse the equation, things might look differently. I have about 3 potential free hour per workday and 12 free hours per weekend day, for a net 39 hours of "my own time", earned by showing up to work. Giving 10 hours of those to the employer is a "pay cut" of 25%.
I have yet to see an instance where salary scaled 1:1 with time put in. Where I used to work, the people that busted their asses for 50+ hours week in and week out, showing they were "team players", might get a 6% raise at review time instead of a 2% raise for a strict 8-5 engineer. They might also get a few thousand in extra bonuses, if their department has funding for them. After a while they might get to the point where they were making enough extra to justify the effort, but "a while" could be 10 years or more.
Note that I am specifically avoiding any discussion of non-monetary considerations and that my analysis sounds rather mercenary. I am specifically trying to address the issue of engineers making more money for being expected to put in more effort. Discussion of the non-monetary considerations could be a lengthy tangent on its own.