Introduction: Cast Aluminum Intake Manifold
Several years ago I started a father/son project to build him a 1936 Pontiac 'gangster car' street rod. The photo shows the partially finished car which was drivable, had a finished drive train, all it needed was final paint and upholstery. I decided to replace the original flathead straight 6 cylinder engine with an overhead Chevy 230cuin straight 6 to honor the car design. One feature I didn't like about the modern engine design was the Siamese intake runners where cylinders 1&2, 3&4, and 5&6 shared ports in the head and intake manifold runners. In the stock design not only were these adjacent cylinders which were not sequenced evenly in the firing order, the intake runners were very long for the two sets of end cylinders and very short for the two inner ones. In short, the intake system design was about as bad as humanly possible.
I decided to try and make a custom head/intake manifold designed for maximum performance and efficiency. Specifications of the desired manifold were;
- cast aluminum for bling and robustness.
- 6 individual intake runners to 6 isolated ports in the cylinder head.
- '180degree' design using a small Holley 390 4-bbl carburetor.
- equal length intake runners.
- runner length designed for maximum torque in the 1800 to 2000rpm street cruising speed.
Tools required for this project were many since it required so many disciplines to complete. There were drawing tools, carving tools, wood working tools, fiberglass composite construction tools, welding tools among others. rather than listing all the tools in my workshop, I will leave it to the interested and probably skilled reader to know what tool was used for what step in the project.
Step 1: Designing the Manifold
The first obvious step was to design the manifold on paper. I wanted the manifold for street driving, not competition which determined the final specifications of the design. One of the main principles of good intake and for that matter exhaust manifold design on an internal combustion engine is a 'free' boost in torque that is obtainable by simply designing the proper length tube between the carburetor and intake valve in a cylinder. This principle has a couple of common names; 'organ pipe', or 'Hemlholtz resenator'. The way it functions is this; When an intake valve in the cylinder head opens there is a negative (vacuum) shock wave generated at the valve seat that is sent down the intake runner toward the carburetor as the pistons is descending in the cylinder, trying to suck in air and gasoline for the next power stroke. This shock wave travels at the speed of sound in air. When that negative wave reaches the plenum which is effectively open space under the carburetor, a second shock wave, this one a positive pressure wave, is sent back down the pipe toward the intake valve in the cylinder head. It too travels at the speed of sound and if the runner is the proper length for the time it takes those two shock waves to travel the length of the pipe, the positive pressure wave reaches the valve seat just as it is closing and 'supercharges the cylinder' with extra air/fuel charge, yielding a more powerful power stroke.
As mentioned the same principle works on the exhaust side of the engine but there the original pressure pulse is positive since it is a result of the exhaust valve opening and the piston expelling hot, high pressure gas. That pressure pulse travels the length of the exhaust pipe, reaches the header collector, expands and sends a negative pressure pulse back up the tube. The header is sized for that vacuum pulse to reach the exhaust valve just as it is closing. Camshafts are designed to have the exhaust valve just about to close when the intake valve just opens. The vacuum pressure pulse actually sucks the residue of the exhaust from the cylinder and sucks a bit of intake charge form the intake port, effectively supercharging the cylinder with more gas. Header tubes are longer than intake tubes since the exhaust gas is hotter and the speed of sound is much higher. The proper exhaust tube length for my 2000rpm street engine is 34" to go with my 11" intake tube length.
This sounds like black magic but it is easily demonstrated to work on an engine dynamometer. Since the speed of sound in a intake runner at typical air temperature is 1100ft/sec, it is easy to calculate a runner for the 1800 - 2000rpm optimum speed I wanted is 13". This length is the sum of the intake manifold runner plus the length of the intake port in the head between the manifold flange and the intake valve. In my case that dimension is 2". Thus my manifold runners needed to be 11".
If all this sound too good to be true, it partially is. The good news is we get extra power at the design RPM and another complementary boost at 2X the design RPM. The bad news is power is a bit less than expected at speeds between these positive numbers because a negative pressure pulse can arrive at the valve just before closing, causing a bit less torque. Detroit and their pals overseas take the approach of designing manifolds with purposefully un-designed runner lengths of different sizes to all the cylinders, thus the power strokes are all over the board for the cylinders which in effect smooths out the power curve at a lower level for a 'jet smooth ride'.
With the runner length designed, I set about putting my design on paper. One thing to take into consideration when designing a casting pattern is the pattern must be slightly bigger than the finished product since the metal used for the part will shrink as it cools making the part too small if not compensated. I added a bit of length (about 1/4" as I recall) to the manifold between the far port sets but otherwise didn't concern myself with this shrinkage since it would have little impact on this project.
I sized the runner passages in the manifold with the same height and width as the ports in the head. I consulted an intake manifold gasket for this dimension. The spacing of the ports and mounting bolt provisions were copied off of a factory intake manifold. The carburetor plenum (the volume of space under the carburetor necessary to allow the organ pipe shock wave to generate) were designed by referring to the Holley carburetor.
The pictures of my working drawing show how I achieve the equal length runners to all 6 cylinders. The 4 outside cylinders were easy. Just running the pipes from the plenum to the ports on the head generated the necessary 11". The big problem was how to get the requisite 11" length in the two inner cylinder runners. I took advantage of another characteristic of organ pipe theory. The generated pressure waves are not affected by bends in the pipe as long as the bends are smooth and not kinked. Thus by studying my drawings you see how I bent the two inner runners around the intake plenum for a compact structure while still achieving the 11" runner length.
Step 2: Carving the Pattern
The next BIG step was to carve the pattern for making the sand casting mold. I decided to use mahogany for the pattern since that is what the pros use. It has a very uniform grain with hardly any defects making shaping easy and predictable. It is also a fairly hard wood for durability. I cut the various pieces very roughly on a band saw and glued them together with no fasteners. I was pretty generous with margins since I was still a little fuzzy on the final shape of everything. I am much more comfortable working in 3D than on paper so this step was actually easier than the drawing step. The photos of the resulting pattern show I created fillets and built up parts of the pattern using polyester body putty. Note the protrusions sticking out of the ends of the intake runners and out of the top of the carburetor mounting flange. You would expect these to be hollow and in fact in the final product they are hollow. On the underside of the pattern there is a hollow chamber carved there which in the finished manifold will have an aluminum plate bolted on with a gasket and with inlet and outlet for hot water lines or tubes carrying hot exhaust gasses to heat up the manifold. All street driven manifolds have this heating provision to insure the gasoline that comes from the carburetor is vaporized and does not puddle as a liquid in the bottom of the manifold. The process of vaporizing liquid gasoline in a carburetor is an endothermic process meaning as the liquid vaporizes it absorbs heat causing the air/fuel charge to cool which in turn can recondense the liquid. Solids and Liquids do not burn, only gasses burn. Car intake systems need to insure the gasoline is in a gaseous state inside the engine. Those protrusions are there to accept the ends of runner cores used on the casting process. The sand casting mold made of the outside of the mahogany patterns is a void exactly that shape. Without sand cores introduced into that cavity to define the runners and plenum. The resulting casting would be a giant solid paperweight.
I mentioned that this manifold was to be a '180' degree' design. There are a couple of options the manifold designer has in deciding which barrels of a multiple passage carburetor feed which cylinders. If the manifold is intended for all out competition, that usually means the carburetor will be at wide open throttle most of the time. The engine will be at extreme speeds and all cylinders are fighting to get their fair share of fuel. In this situation the plenum under the manifold is wide open or '360degree' so every cylinder is fed by every barrel of the carburetor.
On the street, this setup is not very efficient. The engine is running way under max speed and if every carburetor barrel was open to every cylinder, the flow of the air/fuel mixture would be very lazy and neither the carburetor nor the cylinders would be running efficiently. In this scenario the designers place a barrier wall that divides the plenum under the carburetor so one half of the cylinders are fed by one half of the carburetor. This vastly improves the efficiency of both ends of the system, improving low to mid-range torque and power resulting in a much more pleasant and fuel efficient driving experience. Another desired feature of a 180degree manifold is to group the cylinders so now tow cylinders in a group fire sequentially. The cylinders ideally fire alternatively between the two plenums all through the firing order. This way the mixture has time to switch between runners at a leisurely pace. It worked out for this engine that runners 1,2,& 3 and 4, 5, & 6 are optimum.
Step 3: Making the Cores for the Passages Insid the Manifold
The next step was to make the patterns for the sand cores for the runners and plenum inside the intake manifold. Since those parts must fit precisely inside the casting mold leaving a uniform 1/4" space for the walls of the passages in the final casting, I used the big mahogany manifold pattern to help create these pieces. I made a plaster of Paris casting both halves of the pattern. Then as the picture shows I systematically lined each runner with a 1/4" thick layer of modeling clay, which in turn received a pouring of more plaster of Paris. Once hardened, each of the runner pieces was trimmed and massaged until it was a perfect negative of the desired runner passage. The next two photos show the resulting runner core patterns mocked up on a pattern board showing how the runners look inside the manifold. Imagine those lengths of plaster are actually hollow tubes inside the enclosure which is the intake manifold pattern. Note the extensions on the ends of each runner that extend past the upright board on the stand. Those extensions fit into the extensions on the manifold pattern discussed in the prior step. Note that runners 1,. 2, & 3 and 4, 5, & 6 are grouped together.
Obviously I couldn't use these plaster patterns for casting the manifold. The plaster is solid and would be very difficult to get out of the casting and it is likely I would want to make more than one manifold so I needed to convert these plaster models into patterns to make multiple sets of sand cores for casting. I turned to glass reinforced fiberglass technology for these parts. The last two photos in this section show these resulting fiberglass patterns. Each pattern was made in two halves so they could be split to take them off the plaster models and to take them off the sand cores when they were made. With these tools I can make many manifolds. The sand cores look exactly like the plaster models.
Step 4: Other Casting Considerations
Once I had the main pattern and the molds for the cores, I was about ready to take the tools to a local foundry to have them cast my manifold. This Instructable is not one on casting, it is on making a pattern for a casting so I will not go into the process of that skill. I didn't do the casting so we will leave that to another Instructable author to describe.
Just having the free mahogany pattern was not much help to my foundry. The way the make the sand mold for the casting is to make two halves of the mold form the wood pattern. These two mold haves are then pulled off the pattern and laid open. In this mode, it is easy to lay the sand cores made in the previous step into the extension indentations made by the wood pattern, the other half of the main sand mold laid on top of that, yielding a perfect cavity of a manifold into which to pour the molten aluminum.
The two pictures in this step show a fiberglass half-mold I made of the manifold. The foundry can place the manifold pattern in this half mold, ram the sand for the final casting mold half over this half of the manifold, then turn this over, remove the fiberglass mold and ram the sand over the other half of the wood pattern giving the two halves of the sand mold ready for cores and casting.
I apologize for the really bad pictures of the manifold in this step. I took them on the fly back then. If doing it today I would have taken better ones. I can't take new shots since my son sold the car to finance a down payment on his first home. It was a good investment.
Note the intake runner passages in the end of the manifold, the carburetor mount on top, just like a real manifold! I drilled the couple of mounting holes using an intake manifold gasket as a guide. Most of the mounts for this manifold use lugs that clamp down on edges of the manifold flanges with indirect bolts. The foundry flattened the mounting face of the manifold and carburetor surface on a huge belt sander they have. The photos show a commercial carburetor adaptor flange I bolted to the top of my plenum for the Holley carburetor. Using these commercial adaptors, this manifold can use any carburetor.
The last photo shows the manifold in place on the engine. Really came out nice and the car has so much low end torque you would think it had a V8 engine. I sanded and polished the casting for bling.
Step 5: Head Modifications
If this was all I did on this project, it would be all for naught. Even though the manifold yields a perfect design, the heads still have wide open cavities between the adjacent runners. Unless these are isolated, the organ pipe signals cannot do their job.
The attached photo shows how I isolated the Siamese ports. The identified ports are intake ports, the ones to the right in the photo are exhaust ports. The factory isolated the exhaust ports but Siamesed the intake ones. Go figure. Note that the ports are partially divided by a boss through which passes a head mounting bolt. I started the separation by brass-brazing a steel rod, bringing the boss structure to the surface of the head for all three sets of ports. I guess if you are a good welder you could weld up this surface to the level of the head. I had good luck with brazing. These buildups were slightly proud of the head surface and were milled flat by my local engine machine shop.
That still leaves a void behind the bolt boss between intake valve ports. To isolate these I first cut disks of perforated sheet metal that had a grid of 3/16" holes. These disks fit tight in that perimeter. I then stood the head on end and built up a clay dam in the port on the bottom of the pair about level with the edge of the bolt boss. This isn't critical it will be fine tuned later. With the clay dam in place I poured Devcon Liquid Steel on the top of the void which soaked through the perforated steel and was poured to the level of the top side of the bolt boss. Once the Liquid Steel set (it is an ultra premium epoxy), it was an easy task to remove the clay dam and using my die grinder and a carbide burr to shape the two ports to a perfect shape for the soon-to-come organ pipe waves.
NOTE: only use the Devcon Liquid Steel epoxy. It costs near the national debt but this stuff is well worth the cost. It is recommended for repairing cracked engine blocks, making machine parts, etc. It is the most impressive epoxy product I have ever seen. I have no financial interest in Devcon but I love this product.
Participated in the
Metal Contest 2017
2 years ago
Wow. Impressive. I've never heard of a foundry service like that. Most siamesed ports tend to be low performance and considerably smaller than 2 properly sized individual ports. Cutting it in half would sacrifice peak power, but favor low end response. How does the car run? Of course, you could've kept your car all Pontiac with Pontiac's OHC 6 Sprint motor, which does not have the siamesed ports (if you could find one), but it sounds like you like the challenge.
3 years ago
Great article and nice ride. I would love more information/pics on the getting rid of the Siamese ports issue. I know what you did, just having trouble picturing how you did it. lol
Reply 3 years ago
Sorry, I didn’t take a lot of pics of the project. Done before all these how-to web sites! It is really straight forward though. First task is to grind the bolt tubes cast into the intake runners to clean metal. Then braze strips of steel to fill the void and seal with the intake gasket. Need to add little strips of gasket material to the stock Siamese gasket. Don’t worry about being too proud of the surface, will be milled off when machine shop cleans up the mounting flange surface.
Then sand the rest of the port clean with sanding drums so the epoxy will stick. Using Devcon plastic steel is important. It is formulated specifically to repair engine iron castings. I have never seen anything like it. Next, find some thin perforated steel and cut and bend the edges to make insert baffles that ‘snap’ in place behind the bolt tube, between the valve ports to give support to the epoxy.
Now stand the head on end and pack the bottom Siamese port with modeling clay but well below the perforated separator. Don’t worry about shape of the plastic casting, it will be ground to shape after it sets. Mix and pour the epoxy into the open port. It will fill the space below and above the perforated baffle.
When the epoxy is set, dig out the clay then use your die grinder to shape perfect individual ports. It is pretty much as simple as it sounds!
Reply 3 years ago
Yeah, still lost. lol I currently have a lump port head, so the bosses are gone. I'd still like to divide the port some how. Was thinking of making a flanged piece that would sit between the intake and the head that protruded into the head. The part protruding into the head would bump up directly against the back of the port divider (where you poured the Devcon) to create a semi-seal. I know what I have to do, but I like your execution of how you used clay and perforated steel, just not completely understanding it. lol.
5 years ago
That casting is stunning! You have my vote. The pattern & the cores are excellent and would have taken a LONG time. A logical & well presented and documented Instructable.
Reply 5 years ago
Very kind words. It did take some time but I never look at that aspect of my projects. I so much love being out in my shop making stuff, I really don't care how long the projects take.
5 years ago
For certain applications it's hard to beat castings. It looks really good & I can definitely feel you regarding multiple carbs. If you're going for looks over reliability & drivability, sure, multi-carbs look badass, but they are a pain to tune. Did you, by any chance, consider a sheet metal intake?
Reply 5 years ago
Yes, in fact I have built a couple of those. One was a tunnel ram manifold I built for my hemi when in college in the 60s. In the 90s I built a 360degree one for the hemi. I wanted to make this one look factory so went the extra mile.
5 years ago
good project . Is ther a video of the car running?
Reply 5 years ago
No didn't video the car. It ran just like any car!! We had a beefed up PowerGlide tranny, 3:54 rear gears, Mustang II front end. Ran really nicely.
5 years ago
Interesting read -Thanks
5 years ago
Sounds like you did your homework. I have to ask, did you flow test the manifold after casting to verify your results? And the most obvious setup would have been individual runners with single barrel carbs mechanically synced together. Did you consider that option? Either way, nice project.
Reply 5 years ago
No we did not do any dyno work on the manifold. We did have seat-of-the-pants verification that told us it worked. Multiple carbs gets into the performance aspect with the attendant difficulties tuning those things. Our goal was a reliable, easy to maintain performance improvement which is what we got. I have a history with this technology including designing and building an exhaust and intake manifold for a single cylinder engine in the mechanical engineering lab at Fresno State for my Senior Project in earning my BSME degree. Did all the dynamometer work and even built a pressure monitoring device that recorded port pressures every 15degrees of crank rotation to confirm the power curve is a function of port pressure. That was 45 years ago but as I recall I got a 15% improvement in power at the design RPM and an improved torque curve over the entire range. This stuff really works!!