Don't be fooled into believing that the "double hump" stampings on the ends of these two h
When it comes to old-school stock car racers and stock cylinder heads, Chevrolet's long-running "double hump" cylinder heads have always been the gold standard. They are among the best stock heads ever allowed in racing, they are still relatively plentiful in junkyards and machine shops and they don't require a lot of machine work to make ready for racing service.
These heads got their nickname because of the raised identification shape on the each end. It's two half-ovals connected by a straight bar-most easily described as "two humps." Some of these are also nicknamed "Fuelie" heads because a version appeared on fuel-injected 327-powered Chevy Corvettes. These heads appeared on 327 and 350 cubic inch engines across most of the 1960s (beginning in 1962). Although there are some variations, they generally have small chambers for the high-octane unleaded fuel readily available at the time, smaller intake ports and valves that measured either 1.94 for the intake and 1.50 for the exhaust or 2.02 and 1.60 respectively.
UNCC engineering student Dallas Dallman sets up the flow bench for extensive testing on th
Because of these traits, these heads are very popular in Street Stock and Pure Stock classes that require many stock components-including cylinder heads-with minimal modifications. The ports may be a bit small for big street motors, but they are well suited for a class that usually requires a 500-cfm two-barrel. Plus, the combustion chambers are a closed-chamber design and already optimized for a smaller chamber size. You can cut any head down to the minimum 62cc size that most rulebooks require, but an open-chamber design just isn't going to respond as well. Finally, even if you have the heads with the smaller valves, they are easily modified to the larger 2.020/1.600 valve sizes by cutting larger seats and opening up the bowl area slightly on a seat-and-guide machine (and it's all legal).
Because of this, many racers will assume that just because they've found or purchased a set of Chevy heads with the infamous raised double hump on the ends, they've got the best thing going. In general terms, yes. But did you know that not all double-hump heads perform equally? Chevy actually made several running changes to the heads over the years, changing the part numbers as it went. So even though two pairs of heads may both have the double-hump branding, checking part numbers is the only way to know exactly what you've got.
The difference may look slight, but the small reduction in the diameter of a portion of th
Circle Track's University Partnership
Finding out how much different part numbers can vary within the double-hump family was actually a project undertaken by the Motorsports and Automotive Research Center at the University of North Carolina at Charlotte (UNCC). UNCC's mechanical engineering program offers a specialty allowing students to concentrate their studies on motorsports-related topics. The University also offers a very well-stocked motorsports laboratory that supports student-led racing teams (including a Legends car, a drag racer, a Formula SAE open-wheel racer, and others) with the intention of helping them get real-world racing experience. UNCC Engineering graduates can be found working on several NASCAR Sprint Cup teams as well as other racing organizations and even working for the OEs.
The good news here is that the university has recently opened its doors to Circle Track to work together on upcoming projects. But while discussing the possibilities with Steve Johnson and Luke Woroniecki, who lead the student's activities in the Motorsports Lab, the two showed us the work being done by engineering student Dallas Dallman to compare different cylinder heads within the double-hump family.
Dallman took the two most popular part numbers in the double-hump lineup-3890462 and 3917291-measured the critical dimensions of each and then subjected them both to several tests to determine their respective strengths and weaknesses. There are other part numbers within the double-hump family, but these two (called 462 and 291 for the last three digits in their part numbers from here on out) are generally considered the best of the bunch.
All the tests were performed back-to-back, and every precaution was taken to eliminate as many variables as possible. For example, Dallman put both heads on the flow bench and tested them with standard as well as undercut valves. But instead of using two sets of valves, he used the same physical valves in each head to ensure that any variances between two different valves could not cause a false analysis. Tests were also only performed to a maximum of 0.500-inch of valve lift because in the real world these heads are only raced in Street-Stock level classes, and the rules almost always limit maximum valve lift to half an inch or less. The result of testing at higher valve lift levels could only skew the results away from a racer's real-world needs.
So, without any further ado, here are the results.
Dallman uses a pitot probe to map out how flow varies in different portions of the ports.
Although they may look alike to the naked eye, careful measurements show that there are small differences between the two heads. We're only talking about differences in a couple hundredths of an inch when it comes to port width and height, but these are both areas-unlike the seat angle or depth-that you usually aren't allowed to touch, and when you are looking for the absolute maximum in performance it may make a real difference.
Ignore measurements here that are usually allowed within the rulebook to be changed during the course of a valve job-things like seat angle, valve diameter, and even valve stem diameter-instead, look at the height of the ports, the average port width, and even the port length. Port height is important not only because it creates a larger cross-sectional area, but also because the straighter a line you can draw from the entrance of the port to the valve seat, the better air should flow through it. So a higher average port height should be better.
On the Flow Bench
Dallman ran several different tests with both heads on the flow bench. We won't bore you with all the details, but we will pass along his more interesting findings. Dallman tested both heads alone, and with a dual-plane intake, a two-barrel carburetor, and an air cleaner in place. He also tested two different types of valves (an intake valve with a standard, straight stem, and an intake with an undercut stem) with both setups. Keeping with the real-world aspects of these tests, Dallman used a Holley 500 cfm two-barrel carburetor and an Edelbrock 2101 dual-plane intake with a two-barrel adapter plate. It's a typical Street Stock setup.
Many racers assume that the higher-priced valves with undercut stems are for improving flow. After all, they reduce the volume of the valve stem in the port, which should mean more room for the air/fuel charge to get by. In fact, they usually do the opposite. As you will see in the flow bench results (grey chart, page 70), undercut valve stems actually are a small detriment to airflow. Their true worth, according to Woroniecki, is to cut the overall weight of the valve. So even though the flow numbers don't show an advantage, you may be able to gain better valve control at higher rpm levels by using the lighter undercut valves.
Interestingly, the undercut valves do two different things with the different heads. With the 291 heads, the straight-stem valves perform better. But when the carb and intake are added the flow numbers are virtually the same. This shows that the two-barrel carburetor becomes the choke point and the undercut valves no longer have any detriment to the flow. In this situation the undercut valves would definitely be a better option (if cost isn't a factor) because the reduced weight will allow better high-rpm valve control with no tradeoff on flow efficiency.
But, for some reason the undercut valves actually flow better on the bare 462 heads than the straight-stem valves. However, when the carburetor and intake are added to complete the intake tract, the straight-stem valves perform better. We aren't sure exactly why this happens, but it does prove that it's always better to look at entire engine systems as a whole versus several small components individually. If you tested only the cylinder head and the undercut valves on a flow bench without looking at how adding the carburetor and intake manifold affected the intake tract, you would mistakenly be led to believe that the undercut valves are the better option.
The wet-flow setup works off of the flow bench and uses a liquid flowing through the head to show how fuel reacts as it moves past the face of the valve. This is only a visual test, there are no numbers or measurements involved. Also, although the tube is several inches long, during wet-flow testing you should really only look at how the liquid behaves right around the face of the valve and the combustion chamber. UNCC
A comparison of how fuel (in green) flows through the 462 cylinder head as the intake valve opens. The first shot is with the valve at 0.150 lift, the second shot is at 0.300 lift and the final shot is as 0.450 lift. Woroniecki points out that as lift increases, accumulation of the "fuel" on the surface of the combustion chamber around the valve decreases. This should result in a more efficient burn. UNCC
|Port Volume (cc's)
|Avg. Port Width
|Avg. Port Height
Port mapping is also done on the flow bench. This time Dallman placed a pitot tube in different areas of the port and measured the amount of flow. Then he entered those numbers into a program that created a color-coded map showing the rate of flow through the port. A map like this can show you the areas of the port that move air most efficiently.
As you can see with the maps, the intake ports of both heads flow air most efficiently across the floor. This is consistent with most heads of this type. But the 291 head is capable of flowing air across the floor at a faster rate than the 462 head both at medium and high lift.