As 84Turbo indicates, there is actually more to carb selection than the formulas provided by some of the carb manufacturers.
The carb manufacturers provide those formulas as a guideline in order to limit their warranty claims. A small carb is generally much more forgiving to tuning errors, and will run reasonably well even if it's not set up right. A larger carb is more sensitive to tuning errors. For people expecting a carb to be bolted on "out-of-the-box", a smaller carb will provide the least amount of problems.
For those of us who have spent a considerable amount of time on engine dynos, doing tuning and parts comparisons back-to-back with different components, it is clear that an engine does not follow the formula if you're wanting to produce the best possible torque and power curves - there are a lot more variables involved. As correctly stated earlier in this thread, once the carb size has been "maxed out" for a given engine, the use of a larger carb will not produce more power. But an engine will produce better power with a much larger carb than that indicated by the standard formula provided by the carb manufacturers. These larger carbs, to run well, do need to be tuned correctly, though. Once correctly tuned, a large carb will produce outstanding throttle response and drivability, even in the lower rpm ranges.
As an example, my last dyno session was on a 407 street engine I had just completed. I ran a 650, 750 and an 850 on this engine: back-to-back testing. Each carb was optimized for air/fuel ratio so the A/F was identical for each carb. The 650 ran OK - probably accceptable for most people. The 750 raised the torque and power curves through the entire rpm range, with average power between 2500 and 6000 rpm up 24 hp over the 650. The 850 duplicated the 750's curves within 2 hp and ft/lbs at any point on the curve, indicating that the engine was no longer cfm limited. But the 850 did not destroy power or detract from any point in the performance. The smaller carb produced a significant power limitation, even though the "formula" will show that it should have been more than adequate. For most people, it probably is. This engine is now running the 750, and the car is severely traction limited at 2500 rpm...
I have also run back-to-back testing of Q-Jets, Holleys and BGs of similar cfm ratings (Q-Jets are in the 750 cfm range). Once the carbs are properly tuned to produce the same A/F ratio, there is little or no power difference between a Q-Jet, 750 Holley, and a 750 BG ("Demon"), except the Q-Jet will produce quite a bit more torque below 3000 rpm than a similar sized square-bore carb. We've run bracket cars where we've swapped between the Q-Jet and a Holley with no detectable difference in ET or MPH. Don't expect to make more power by simply removing your Q-Jet and installing a 650-750 Holley in its place. If doing so produces a power increase, your Q-Jet was not set up right to begin with, and you could have saved yourself $450 by just fixing the Q-Jet.
Bottom line: Use those "formulas" as a guide. Run the carb brand that you are most comfortable with tuning. For a stock engine in a daily driver where mileage is a concern, run the Q-Jet and set it up right (use my paper as a guide). If mileage is not a concern and you don't mind changing intake manifolds, run a Holley, BG, QuickFuel or whatever. If you're not good at tuning and setup, run a smaller carb per the formula. If you can tune and set up a carb, and you're after the best performance possible, I wouldn't run anything smaller than a 750 on a 350-400, and I run 820-850 carbs and anything larger than 420 ci on the street.
Also, just as a little sidenote, it is interesting to note that a properly built street engine with good heads and a well-matched cam will actually produce over 100% volumetric efficiency in certain rpm ranges, even if normally aspirated. An engine's capability to run at over 100% VE should be considered in carb sizing if you're trying to make peak power and torque numbers.
Lars
