The accepted answer states drag does not scale with mass. Which is true. But drag does scale with frontal area. It is fair to assume the two frames are made of the same material therefore the larger frame has a larger frontal area.
Rather than compare balloon and soccer ball of the same size a more appropriate comparison is two rocks of the same density but different sizes.
Terminal velocity is when the force of gravity equals the drag
See this link for the calculation of terminal velocity
Terminal velocity
In the equation once you take out the constants
vterminal is proportional to square root (mass / area)
vterminal is proportional to square root (r cubed / r squared)
vterminal is proportional to square root (r)
So at constant density if you double r the terminal velocity increases by 1.414
Double r is eight times the mass for only 1.414 the terminal velocity.
Drag proportional to v squared is the real drag (pun intended)
Now lets pretend you could double your mass and keep the same area
vterminal is proportional to square root (mass / area)
vterminal is proportional to square root (mass / constant)
vterminal is proportional to square root (mass)
If everything was constant (including your area and rolling resistance)
If you increased mass by 2 you would increase terminal velocity by 1.414
If you increased mass by 4 you increase terminal velocity by 2
Rolling resistance is not constant so it would be less than that 1.414 and 2
Let say 180 lb rider and add 20 lbs of lead to the frame - that is only 5% straight down hill
On a 10 grade that is only 0.846% - 40 mph versus 40.43 mph (without accounting for rolling resistance).
Even climbing bikes are designed to be light.
Basically up the hill you pay for all the weight and down the hill you only get credit for the square root of the weight.
Drag proportional to v squared is the real drag