This all made me curious so I started to do some analysis. Using the book I mentioned (NOTE that one of the formulas he quotes from another book is wrong. He forgot the square root notation) and Performance Trends Software. If you run a series of plots using a validated 340 engine and the stock manifolds that have an average 5" length the length is not long enough to do any scavenging since one cylinders pulse has reached the collector before the next one has time to fire so bigger is better in the diameter. There is a point where it makes little difference but 2.0" is better than the 1.5" I stated above. This is swept from 1.00" to 2.25"
Then I used the formula to calculate the ideal diameter of 1.5" and then swept collector length. This is where things got more interesting. If you use the formula for the length to diameter it says I should use a 35.25" for a peak torque 4000 RPM (which is validated in the run). Here are the sweeps from 15" to 60" (max the tool will allow) in 5" steps. What is interesting is that the peak HP
is made at the 35" length at about 5600 RPM (the blue one) but if you make the length just a little longer the peak comes down a bit at 5600 RPM but there is a fair amount of increase in HP in the 2500-4500 RPM range and the torque gets higher and higher until about 55" then the length stops having an effect.
The design of the collector makes a difference but there is not much theory that can help design it. The books say its design is best dialed at the track.
So if we take a header and hold the length constant and use a 1.5" and 2.0" tube on the 340 we get this. The 2.0 pushes the peak HP point up in RPM (as expected) and does make a bit more peak HP there
but the rest of the band where most of use want the power for the street takes a hit. Also notice that the factory manifolds really make more torque at very low RPM where us granny drivers typically are. Thus my decision to keep the stock manifolds and not use headers because of the way I drive my Coronet appears to be the right choice. I rarely am above 3000 RPM.
So now lets use the formula the 2.0" pipes... The first thing is for a 2" pipe the peak torque will need to be shifted all the way out to 7000 RPM with a collector length of 24". Having peak torque at 7000 RPM seems like a pipe dream as it is likely falling by that point so I don't have high expectations for this combination. And I was right. Worse performance everywhere other than keeping the torque up with the factory manifolds under 2000 RPM.
Note that your milage may vary as this is a model and it assumed everything was held at a constant other than the pipes. Optimizing an engine is a very interesting and complicated problem to say the least.