Jack's homebuilt flowbench
(click on thumbnails below for larger picture)
This bench was built to a plan published quite a few years ago by someone called Hyperflow. I have been unable to find their address or contact information. The booklet that contains the plans and calculation tables is about 50 pages long. This bench was designed before current data acquisition tools were available, so it uses two manometers and then a calculation is necessary to get cubic feet per minute. While the book is copyrighted, it appears that the publishing company no longer exists. As such, links to a scan of this book is at the end of this page.
The bench cost about $300 to build. All parts came from our local hardware store. I take readings and then do the calculations on a computer using Excel tables. I only have to enter one number for each valve opening and it calculates cfm.
I have calibrated this bench using my machine shop’s Superflow as the standard. I have a head flowed by them, then put the head on my bench and adjust factors in the calculating tables to get the same readings. I regularly have the shop check the flow on one port of a customer’s head just to make sure nothing has changed.
To flow a head, you
It takes abut five or ten minutes to take this set of readings on one port.
Front view: You can see the cabinet we built, the riser on which the head sets, the valve opening gadget, the vacuum bleed. Not in the picture is the vacuum source, which is a Craftsman Shopvac.
Manometers: These are plastic tubes recessed into a groove on the display board. I use manometer fluid which works better than colored water. The manometers are connected to two reservoirs located on the back side of the display board. Lines are easy to make and numbers are the stick-on type. The upper manometer reads the test vacuum so you can set it the same for every valve opening, and the lower one reads flow across the sensor.
Bench inside: You can see the airflow plumbing and the sensor, which has a bunch of small plastic tubes attached (more later).
Sensor: This is the measuring device for the bench. I made four sizes but have need for only one size for TR4 intakes and a separate size for exhaust. The other sizes would be used on other heads.
Riser: The main reason for this is to put the head above a clear plastic tube so you can make sure the combustion chamber is lined up with the cylinder walls, look at thread you may feed through the port, etc. I covered the top of the riser with a soft material that seals very well. I have to be careful to have all the water passages in the head taped closed.
Vacuum Bleed: All readings are taken at the same vacuum. On my bench I use 7”, commonly used by commercial benches, and convenient because of the capacity of the vacuum source. This is simply a vacuum bleed used to adjust the vacuum for each valve opening, taking all readings at the same vacuum.
Computer: Generic computer with calculation tables.
Valve opener: Some means must be devised to open the valve the proper distance. I use a light spring and stock keepers to hole the valve in position. The gadget uses a $10 base and a $15 dial indicator and some common hardware pieces.
Speed control: Although other sources recommend air tools, I prefer a Makita Die Grinder, about $90. A means of controlling the speed is nice, so I used a common light dimmer, electrical box and two receptacles. Works great.
Grinding: You can grind away at things with stones, but it takes forever. I use carbide tools with long shanks. They are rather expensive at about $50 apiece but they remove metal quickly. Figure on spending a couple hundred dollars for the assortment you’ll need and be prepared to ruin a couple when you start. I then use stones and abrasive drums for final smoothing and polishing. When working on aluminum, stones and carbide tools load up in just a few seconds. This can be eliminated by using a magic elixir named “Tap Magic”. It is expressly designed for a lubricant for cutting tool used in aluminum.
Some general comments:
This bench is very repeatable day to day. I can measure a head, put it away, and get the same readings a week later.
I find that this bench does not yield exactly the same readings over the entire range as a Superflow does. It is commonly accepted that you cannot compare readings exactly between two benches. There are a couple of differences between this bench and the Superflow. First, the Superflow measures pressure drop across an orifice rather than a venturi. The designers of this bench recommended the venturi as being more accurate. This probably causes some differences. Also I use a six inch high plastic tube into which the head dumps and the commercial guys don’t necessarily do that. I find that the length and diameter of this part of the air path significantly affects readings. If I set a TR4 liner on the bench and put the head on top of it, I get a completely different reading.
The main benefit of a device like this is to measure the flow changes that take place when you alter the port, valve, or combustion chamber. I’ve used a number of scrap heads for development, and have had them sawed up to make sure of the casting thickness. Sometimes I get fooled, though, because some castings are different and it is not good to grind into a water passage. I’ve made ports larger and smaller, changed the curve, changed the shape to square, round, and D shaped, tried about 20 different valve seat angles and probably ten different valve shapes. Smaller ports can be mocked up with modeling clay inside and then replicated with epoxy, a pretty standard practice in the motorcycle and V8 guys. I work on the head and intake manifold as a system. The challenge with the manifolds are that to get the right diameter and taper in the manifold, you run out of aluminum. I have some aluminum material welded on the outside to give me enough material to work with, and then sandblast it to give a stock surface appearance to satisfy the purists.
Now a few philosophical comments:
Porting heads is a lesson in humility. Air in a port does not go where it is supposed to go according to common sense. In some areas sharp corners are better than round ones. Some combustion chamber mods that you “know” will improve flow simply do not.
Increasing size of a port will increase flow, up to a point, but it may actually kill horsepower because of the decrease in velocity.
A person needs to read as much as possible on the subject, but even then you will find that the written material is not such that you can apply it directly to our dumb old ports. Virtually all head porters do not divulge how they get their result, because we all feel that we should be compensated for those hundreds of hours we’ve spent experimenting.
You do learn some interesting things, though. Examples:
When you install a hard seat, it seldom matches the bore exactly under the valve. It looks like it should be blended, rounded, filled in, or something to make it match. Actually, this makes a surprisingly small difference.
Straightening out the curve in an intake port, something that we just “know” will help, actually decreases flow.
I have done engines where the former porter made the ports really large. Big flow numbers on the flow bench. Then to match the manifold, they put a big chamfer on the outlet side of the manifold to match it to the head. The end result, when the head and manifold were measured as an assembly, was that the whole system flowed less than a stock system!
Finally, I shudder when I read that someone is “matching” the ports. Unless done properly, this can decrease flow and horsepower. At the head face, if you match by enlarging the manifold a short difference, flow suffers. For intake ports, the square are of the port should either be constant or it should converge slightly. Any divergence or increase in area will harm flow and velocity. On the exhaust side, it is actually good to have the port slightly smaller than the header pipe and leave that shoulder, because it helps reduce reversion, that cloud of raw gas you see standing outside the SU carb inlet on a cold and damp morning.
All that said, have at it!
Links to Hyperflow booklet PDF file scans (right click, save as):