Head porting

[2seater]

 

[2seater]

Porting the heads is perhaps a bit of a misnomer, since it is primarily consists cleans up casting marks, smoothing core shift lines and generally removing sharp edges and blending. There are only two areas that are actually enlarged; the throat right below the valve seats and where the pushrod intrudes on one side of the intake port right near the entrance. 

 

       The pushrod area doesn't actually shrink the width of the port but it does cause a curve in the wall before it can straighten and widen out. The pushrod intrudes by differing amounts so while the ports for #1 & #6 could be straightened completely, #2 & #5 must have some of the curve remain. All were made to match as much as possible. The intake port entrance was left at the original size since it already matches the gasket closely.

 

        The throat below the intake seat does contain quite a few abrupt changes, sharpish edges and is not completely round. This area was opened up to be round. The intake looks to have a standard three angle opening for the valve and I ground out the opening to within approx. .050" of the 60* cut made just below the 45* seat. The rest just consisted of blending and transitioning the round valve opening to the more rectangular port runner. The valve guide casting was smoothed but was left pretty much the original size except the top edge toward the valve opening was chamfered. The Buick head has a 10* valve angle so the valve guide is located more in the bowl below the valve so there is less intrusion on the actual runner.  The exhaust port is already very good in stock form so just the basic smoothing and blending the edges below the valve seat was all that was done.

 

         The net result is the intake flow increased across the board, with small 2-4 cfm increases at lifts below .250" and then steadily increasing the gap to 23cfm @ .400", or 185cfm. The flow continues to rise up to 194cfm @ .480" lift and falls back to 189 @ .500".  The best way to visualize: the port now flows slightly more @ .300" lift (163) ported vs .400" lift (162) unported.

Average airspeed is 262fps @ 185cfm.

 

          The last item was back cutting the intake valves. I experimented with many combinations but there was little to be gained beyond expanding the small back cut already on the valves, both stock and aftermarket, which appears to be 22.5*. The goal isn't as much to widen the cut but to move it closer to the 45* seat contact face. Both stock and aftermarket valves have a .115" wide 45* face. This was reduced to .100" by increasing the width of the 22.5* back cut. To illustrate what I mentioned previously about the "flatter" back face on the stock valve, the width of the back cut on the stock valve increased from .050" wide to .085", but the aftermarket went from .085" to .110". This appears to improve low lift flow, below .200" lift by 1-3cfm but I hesitate to quantify as my bench is not configured for reading low flows below 60 cfm or so. I intend to examine this area with better resolution in the future. The pic is of the two valves with the back cut. The 45* was dyed red before grinding.

[2seater]

Reconfigured the bench with a smaller measuring orifice to better capture the low lift flow, and I did find a slight advantage to a small 30* cut at the edge of the 45* seat face followed by a 15* toward the stem. The gain is very small, and I did run back to back tests, switching the valves twice, to confirm. The gain was a consistent 1-2cfm gain through the entire range through .400" lift where it lost 1 cfm above that. The other positive I found was the bench again confirmed the flow gains from the porting work that was done. This is by no means a full on, maximum effort, porting, just a basic cleanup, smoothing, and correcting of obvious flaws. Something anyone with the time and interest could do with a die grinder, a couple of carbide burrs and sanding rolls.

[Ronnie]

I use to do some head porting on Honda motorcycle heads. I didn't have a flow bench to test my work. I usually followed instructions that others with a flow bench had written in magazines. As suggested in the instructions, there seemed to be more power gains by opening up the intake ports but leaving them a little rough instead of polishing them while making the exhaust ports a smooth as possible. The theory was that the rougher intake ports would promote some turbulence (swirl?) that would keep the air/fuel mixture from separating. Have you found that to be true on the low RPM 3800 engines or do you make both ports as smooth as possible?

[Ronnie]

I just realized that my comments about head porting might not be comparing apples to apples. The head work I did was feed with a carburetor and the 3800 is fuel injected,

[2seater]

I do leave the intake with a fairly rough surface, either from a carbide burr or cartridge roll. Even in a dry manifold, my understanding is the rougher surface breaks up the boundary layer that flows more slowly along the walls, but I have not tested that? I do polish the chamber and exhaust port as I have found there seems to be less carbon buildup and I think there may be less heat transfer with the smooth surface. I'm glad you mentioned this. I will see if I have a piece of stainless tubing with a smooth internal surface that I can rough up to see if there is a difference.

[Ronnie]

I look forward to hearing the results of your test.

[2seater]

Well, I tried three different diameters of pipe, and the general conclusion I found was there is little difference, but seemed to be a slight loss with the roughened inner surface. As mentioned, the difference was small, less than 1/2 percent but there was certainly no large gains. This is simply straight sections of tubing 8"-10" long, without bends, so, it looks like I should do a finer finish, without material removal, on one of the ports I have been testing.

[Ronnie]

Interesting. I think the advantage of the intake passage being slightly rough in the carburetor engines is to create turbulence that keep the mixture atomized properly all the way to the combustion chamber. It has been a lot of years since I was reading about head porting.

[2seater]

I agree for a wet manifold, the semi-rough finish has a different type of benefit. In this case, based on the possible benefit of a smooth surface when testing straight pipe, I sanded the bulk of the roughness from one of the test ports. I tried not to remove much metal or alter the contour but some is inevitable. It is not Cratex polished, but is smooth to the touch. I thought I would see some definite change, but that was not the case. I record the air flow reading every .050" up through .500". The gain/loss every .050": zero, zero, -1, zero, zero, zero, -1, -1, -1, +7. The small differences are probably within my margin of error, but the big gain at .500" (the lift it never gets to) is because the port kept increasing in flow as the lift went up rather than falling back @ .480" as it did before. It boils down to "nothing to see here".

[2seater]

My good friend has a white '87 Turbo Regal, aka albino Grand National. He has spare parts to make a couple of engines and he gave me one to test that he bought at the Buick meet in Bowling Green a few years ago. The history is unknown but it certainly doesn't appear to have been ported in any way. Our 3800 heads certainly look better in stock form, smoother and more gentle bends. The valve sizes are the same as ours although the combustion chamber appears flatter and has a bit different shape. In any case, I ran a quick flow test on it to get a baseline comparison. Stock for stock, the 3800 head flows better across the board, not by a ton, but consistently better. The flow pretty much peaks @ .400" lift and stays there, while ours does increase a little over that lift, so the gap widens. In order, from .100" to .500" lift, every .050", the results for the GN head were:

-5, -7, -9, -7, -8, -8, -5, -12 and -16cfm.

 

It looks like a 3800 head on a GN would yield a useful improvement, except the order of the valves is different so a cam would be a special and a 3800 intake manifold would need to be modified as the 3800 block is actually a little longer to center the cylinders over the crank throws (aka on-center).

[2seater]

This is loosely related to the porting project. My friend with the GN engine has built a rig to test valve springs fully installed in the head, which I have been trying on the heads I did. So far, the few springs I have checked have been within the book values, kudo's to the low mile heads I received from Daves89.

 

 

[Ronnie]

Tell us more about what we are seeing. Is that setup measuring the amount of pressure needed to compress the spring for a certain amount of valve opening? I can't tell by the photo. My friend with the engine shop has a setup that does that off the head but I've never seen it measured on the head. His fixture has a handle on it to compress the spring.

[2seater]

Maybe this view will help. It has end plates that bolt to accessory holes in the ends of the head and two round stainless steel cross bars that extend from end to end on the head. There is a bridge between the crossbars that can travel end to end which carries the pressure gauge for the spring pressure measurement. The dial indicator rides on top and bears directly on a tab that sits on the retainer to measure movement.

 

It is probably more complex than absolutely needed for simple checks where a lever type pressure arm might suffice, but this allows you to develop a pressure curve at any increment desired. It also works better on the GN style head which has a rocker shaft and no individual studs that line up with each valve for a lever action.

 

[Ronnie]

That setup might be better for the results you are trying to get with it. We used a tester similar to the one in the photo below to determine shims that needed to be added to get the proper spring pressure without the spring binding. The tester was basically a lever operated arbor press with gauges to monitor pressure and travel.

 

I don't remember doing any plotting with the tester we used. I think we assumed valve train geometry had been determined by Lunati Cams with the exception of how much we needed to cut the spring seat in the head or shim as needed to get the initial spring pressure correct. I don't remember all the details. It's been over 25 years ago when I was helping build drag racing engines.