If we're going to make improvements, we have to know where we started. The first step is to measure the combustion chamber and port volumes on the cylinder head. Conventional wisdom says you want the most flow you can get with the smallest port volume possible to keep flow velocities high, and we'll try and abide by that rule of thumb until we see otherwise. Chamber volume is important because the last thing we want to do is lose compression. If the chamber needs to be opened up to improve flow, then we may need to mill the head a bit, or weld in some material in the chamber to keep the compression ratio up. Additionally, we want each chamber to be as close to identical in volume as possible. While we won't be making final port and chamber volume measurements until the end of the project, the baselines are as follows:

These numbers tell us a couple of things. First, the consistency from cylinder to cylinder on the stock head is very good. Combustion chamber volume varies by a maximum of 0.5 cc, or less than 1%. Standard deviation is less than 0.2 cc from an average of 53.6 cc. Obviously, Honda's manufacturing processes are pretty good. The intake and exhaust ports show similar consistency as well. One thing that Joe pointed out is that the intake ports on the F20C head are significantly larger in volume than the B16A and B18C1 cylinder heads - on the order of 25 cc bigger stock to stock. Of course, the head has to support a lot more HP than those engines, so maybe that shouldn't be surprising.

Other observations made during the measurement process include some inconsistencies in valve seat width from chamber to chamber. This is most likely due to some casting variations. With almost any casting there will be some core shift (movement of casting cores within the mold) and the variations we see from chamber to chamber are probably an effort to minimize the effects of these imperfections on performance. This will be important to flow maximization, as you'll see in a future installment. Joe also noted that the stock ports show signs of slight cleanup work with a grinder. The cleanup work is slightly different from port to port indicating careful attention to detail. Uh-oh, with clean up work from the factory, will there be anything left for us to find? Overall, the casting is of very high quality, and Joe said it's the best he's ever seen on a production head.

The next step was to get the cylinder head onto the flow bench and see how it performed. This required fabricating a mount with an 87mm bore to simulate the F20C cylinder. The mount was made out of plate steel and tapped with holes to allow the use of studs to mount the head securely. Joe also fabricated a micrometer mount for the head to allow the precision opening of the valves.

Before getting into the numbers, it would be useful to cover what is involved in flowbench testing. Essentially, a flowbench forces air through the port and valve. To measure this flow, a reference pressure level is established (usually in inches of water) and then a percentage of maximum flow is read off a meter. For example, at 10" of vacuum under a maximum flow setting on the Superflo 110 (the model used at ACHT), 100% flow would equate to 185 CFM. For our purposes, we'll be listing all flow in CFM at 10", so don't worry if you don't quite understand. You should also be aware that many shops and magazines like to list flow at 28" of pressure. This is pretty hard to achieve on most flow benches, so a conversion factor has to be used. To convert reading at one pressure to another, simply divide the new pressure by the old pressure and take the square root. Multiply the flow reading by this factor and voila, you've got a corrected flow number.

Flow @ new pressure = Flow @ old pressure*(new pressure/old pressure)^0.5

Here's a good look at the readout from the bench. The angled tube is filled with fluid that will move corresponding with flow rates. The reading from the tube can then be referenced against the flow chart seen below. Simple, but highly effective. To examine the performance of the head at varying valve lifts (corresponding to available lift from the camshaft) we take flow measurements at a variety of lift points. These are usually marked off in increments of 0.050". For this test we went as high as 0.550" since the F20C head uses a large intake valve which can support high cam lifts. So, what were the numbers? Take a look at the charts below, and then we'll explain what they mean. The first thing you'll notice is that from cylinder to cylinder, flow is pretty consistent. Each cylinder is within 1% of the average, which is pretty sweet for a production cylinder head. The #3 intake port was a bit higher than the rest while the #4 exhaust outflowed the rest. Interestingly, there was a trend for the outer intake ports to flow a bit less than the inners (#2 and #3) while the outer exhaust ports tended to flow better. This may have to do with the port bias (orientation of the port relative to the valves and chamber). The outer intake ports take a sharper lateral angle while the situation is reversed on the exhaust side (the inner ports take a sharper angle). In the end though, the numbers are too close to complain about - most porters are happy to get within 1%. The second thing you'll notice is that while the intake flow tends to flatten above 0.450", it still continues to climb through 0.550. The stock intake cam only lifts about 0.460", which matches up quite well with the stock ports, but a higher lift cam could show noticeable performance gains. The same can be said for the exhaust side. The flow begins to flatten after 0.350" and the exhaust cam runs about 0.400" of lift. But there is more flow to be had at higher lifts. The real kicker will be whether or not the valvetrain can support big lifts, especially with a more aggressive ramp on the cam. It might be interesting to note how well the F20C head flows compared to previous DOHC VTEC heads. Because Joe specializes in Honda heads, he just happened to have all the relevant data, which can be summed up in the chart below. Copyright 2002 Temple of VTEC [Previous Page] reader's comments Viewing Threshold (What is this?) [fancy] [flat] [simple] Subject Thread Originator Replies Last Post   similar research for the K-series? Tuan 0 04-30-2002 10:41 by Tuan