Blueprinting AFR Heads

[Chris Duncan]

This is a continuation of another thread that was a victim of thread drift. THREAD HERE


After research it was determined that the AFR head had the best flow and they were subsequently purchased in Nov. of 2001. Just now getting around to setting them up and had a sort of rude awakening. They are the SBF 165cc emissions heads with all hardware.

Bolted the Crane energizer rocker arms and right away noticed the roller tips weren't centered on the valve tips in either direction. Having already heard some horror stories from the machinist about stuff breaking on an AFR head I researched further.

Not only do the tips not line up but they are inconsistent from one to the next, some are way off and some are dead center.

Degreeing the cam showed some other problems. Timing variances at the valve were off by as much as 4 degrees between valves and lift was inconsistent and way over the cams advertised specs. Measuring timing and lift at the cam showed it to be dead on so the problem is in the upper valve train.

First will be discussed inboard to outboard centering of the valve tip contact.

When the rocker arms are set up to ultimate geometry and correct cam spec lift at the valve, the roller tips are about .075" short of center. When the pushrod is lengthened to center the valve tip contact the geometry is past ultimate and the lift on the valve goes from an advertised spec of .542" to .572". This is because the rocker geometry goes from a designed ratio of 1.6 to something closer to 1.7.

This extra lift causes more problems because the cam was purchased to work with the valve springs which AFR says will handle .550" lift. It also brings up a valve to piston clearance issue. Again here it is inconsistent with the best being about .075" and the worst about .043". This is going to require some eyebrow cutting to bring it to the .090" to .100" range that is standard practice.

Next is the front to rear centering of the valve tip contact. This is adjusted by the push rod guide plates but you have to adjust each cylinders valve pair together in that each plate has two slots that are a fixed distance apart. 7 of the 8 cylinders would not center in that the tips of the rocker pairs were too far apart to line up with the valve pairs.

What's causing all these problems? It looks like production tolerance stackup. A straightedge shows that the valve tips are slightly inconsistent in all three planes and the rocker arm studs are also off in the two planes that matter (front to rear and side to side). Further investigation and machinst's input shows that the rocker studs are budget items that are inaccurate and soft. They aren't straight and are easily bent with a soft blow from a mallet. The threaded stud holes in the head are also inserted with heli-coils which aren't the most precision method.

The first thing that was tried was a rocker arm swap just to be sure. The Cranes had the same part number for small block ford and chevy and the Comp cams pro-magnum had a different number so it was thought that the stud hole might not be the same. The pro-magnum was just the same as the crane as far as valve tip contact, the only benefit was it's a more durable and lighter rocker. So the rockers weren't causing any geometry problem.

The next correction was to modify the guideplates to get front to rear valve tip contact centering. Isky is the only one that makes adjustable guideplates for adjustment between pairs. The AFR plates were cut down the center with a grinder and then the rockers centered individually and the plates welded back together in place. Heat dam paste was used to keep the weld heat from ruining the plates hardening and damaging the head. The weld area has to be ground down also in that the surface coating or hardening affects weld quality.

Then to help inboard/outboard centering a set of ARP rocker studs was installed. This improved considerably the centering consistency it's not perfect but it's acceptable. The pushrod length that yields correct geometry still doesn't center the tip. The compromise will be to center the valve tip and accept the increased valve lift/ rocker ratio. This necessitates a smaller cam to work with the existing springs. A call to AFR brings up another problem, the installed springs only work to .510" lift when a roller cam is used, now you tell me. This is caused by the ramp rate of the roller cam needing more spring. So now the cam and the springs will have to be swapped. It might be nice to stick with this same cam but the pockets and installed height of the existing springs won't allow a big enough spring. The pockets would have to be opened up and a longer valve installed.

The cam is brought down from a .542/.563 286/294 to a .499/.510 275/299. These specs are with a 1.6 rocker ratio but the increased ratio due to existing geometry will bring the lift to about .520/.540. The springs are swapped for a Comp cams 978-16 that has the same dimensions but a higher rate to handle the roller cam and increased lift from the higher ratio rocker geometry.

The only other problem will be piston to valve clearance, will have to remeasure with the smaller cam but some eyebrow cutting will probably be necessary, hopefully not so much as to require balancing.

Just a heads up, no pun intended, for anyone purchasing AFR heads. If you want to blueprint to a high standard maybe consider bare heads and do your own machining of valve guides, rocker stud holes, and an improved thread insert. Maybe they have improved since the time these were purchased I'm not sure. I still think these heads have the best flow out of the box they could just use some improvement in the valve train precision, although this may be standard for the price you are paying.

Will update this post and maybe some pics as time permits.

 

[Lynn Larsen]

Chris,

Thank you for the excellent description of your findings. I will print this out and give it to the machinist. I will ask him to very carefully check this out. You are familiar with what I have seen in a mock up of the Jesel lifters on AFR 205s, so I am quite sure I will have similar issues when it comes time to blue print the valve train.

I am glad to hear that the cam itself wasn't off as was my initial thought when you said the valve timing was off as much as 4 degrees from the documented event locations of the cam.

Thanks again,
Lynn

 

[Bill Musarra]

I have a set of 205s on order. AFR said the order time was several months. We located a set in Mississippi and are having them set up with the Comp springs that come with the cam. I don't have the cam number infront of me, but it is a hydraulic roller with the matching steel roller rockers just below the XR286R that was used in the Muscle Mustang head shootout. We will be checking the heads also, but I find it hard to believe the heads in 2004(or in 2001 for that matter) are manufactured like Kalun found. All that he found must be a quality control issue.(for 2001 at least) Have any others found this to be a usual finding in aftermarket heads. I hope not. In the article in Muscle Mustang, Ultimate Head Guide VI, all the heads were setup without any special reworking with the exception of the TFS heads which had an offset valve location. They need a different valvle relief than other heads. All the heads were measured and checked, especially for valve to piston clearance. The 205s sneaked by with a .080.
Quoting from the article:
"In addition to measuring piston to valve clearance, it was also necessary to adjust the rocker geometry. This was accomplished by altering the pushrod length. the majority of the heads were run with pusnrods measuring 8.250 inches, the exceptions being both the TFS heads. The TFS high port heads required an 8.40 inch pushrod while the TFS R heads were longer still at 8.70 inches. The pushrod lenght was determined by the contact patch achieved on the valve tip. We were looking for an even contact positioned in the center of the valve tip. Too long and the contact patch would move away from the pushrod, too short and it would moved toward the pushrod (on the valve tip). A centered contact patch ensured accurate geometry and helped maximized the effectiveness of the cam profile."
The lifts and other measurements were done with dial indicators and no mention of any problems. Now they may have measured only one valve for the proper pushrod lenght and one valve for the proper valve to piston clearance. I don't know. The article stated that they were going to do the VI article with a 418 radical engine, but in their haste, didn't measure the valve clearance and bent several valves after they noticed power production was off. This article was done with a Coast High Performance 393.(3.85 x 4.030). Maybe this is just a problem of solid rollers???? At any rate I plan to stick with the Comp equipment as it was designed to work together. Why the variances that Kalun found??? What do others think??? Better still, What does AFR think??? Have they changed their manufacturing processes since then??? Their quality control???? It sounds like a lot of problems for a company with a very high reputation, and a lot of heads out there. Anyone interested in the article, let me know and I will scan it and send it along.
Bill

 

[Chris Duncan]

"" All that he found must be a quality control issue.(for 2001 at least) Have any others found this to be a usual finding in aftermarket heads? ""

Three different people I've talked to locally had at least minor problems. Two of them had to split their guide plates, set and weld, the machinist related the story of the engine blow up on the dyno, (which could be the owners builders incorrect setup), and they all replaced the studs with ARP parts.

""This was accomplished by altering the pushrod length.""

Yes the adjustable pushrods are at 6.450" about .200" longer than stock. I haven't ordered the pushrods yet still waiting on the smaller cam.

""A centered contact patch ensured accurate geometry and helped maximized the effectiveness of the cam profile.""

This is not what happened in my case, once the tip was at optimum center the geometry was not accurate as indicated by the extra lift. This is exactly what the tech at AFR said also. "All you do is center the tip and the geometry will be right, I've never even had anybody ask about checking the geometry any other way" I beg to differ.

""The lifts and other measurements were done with dial indicators and no mention of any problems.""

two dial ind. one test ind. digital caliper, 30" straight edge.

"" Now they may have measured only one valve for the proper pushrod length and one valve for the proper valve to piston clearance.""

I did the same at first until it was noticed the tip centering had such wide variance with the same pushrod length, after that all measurements were taken on all cylinders.

""but in their haste, didn't measure the valve clearance and bent several valves after they noticed power production was off.""

Exactly what happened to the machinist's customer. Here's what really threw me. I always assumed that the principle with levers fulcrums and arcs says that as you go away from a 90 degree situation that a point on the end of the arc travels less in a linear direction. That's why I couldn't understand why I was getting MORE lift with less than ultimate geometry. The difference here is the roller tip, as you get way past the 90 degree point it contacts further around on the roller thus the distance between the rocker pivot point and the roller tip contact (NOT the roller axis) becomes longer, thus increasing the ratio.

I think where people may be getting into trouble is measuring the cam lift at the lifter, measuring the valve to piston clearance on ONE cylinder and not measuring the lift at the valve, just assuming that because the lifter is stamped 1.6 that means it's exact, which the rocker is but the geometry isn't. This extra lift overloads the spring if your not running any margin and the inconsistency could also cause problems if your not measuring ALL cylinders for clearance.

Here's the measurement differences I got.

Valve

height max difference .005"

side-side max diff .010"

front - rear max diff .004"

Studs

side-side max diff .007" (this was .015" before the ARP stud swap)

front-rear max diff .020"

Valve lift diff as measured at valve .015"

Valve piston clearance diff between cylinders .030"

I may be hyper critical and I do note that ARP now suggests a stronger spring for roller cams because this head in the basic configuration definitely doesn't work with roller cams.

 

[Mark Worthington]

Three words: Edlebrock Victor Jr.

Yeah, the AFRs have great flow numbers, but I've heard from several sources that for a performance build, the AFR heads should be ordered bare or the upper valve train should be thrown away and rebuilt with quality parts.

 

[Lynn Larsen]

I am not going to comment on the valve train geometry issues until my heads are worked on. But I can tell you this, when my heads were initially checked to determine the exact closed and open pressures of the stock springs and the install height we had to work with, neither I nor the engine builder who checked them was very impressed with the parts or their setup on the heads. The retainers were far to loose of a fit. There were no cups or locators on the bottoms and the springs were only located by the pocket cut into the aluminum. If the heads had been run for any time in the state we found them in, the spring pockets would have been beat to crap, there is no way one would have had consistent behavior out of the valve train and the harmonics would have been horrible. Eventually, something would have broken; either a spring or a retainer would have eventually let go and a swallowed valve would have been the most likely outcome.

If you buy complete heads from anyone, you are taking a real risk if you don't have them checked by a competent engine machinist. This much is clear.

Lynn

 

[Chris Duncan]

Update on the AFR head blueprint. The roller tips are centered within .020", good enough for this engine. I ended up going with a stronger spring but used the existing valve height/pocket dia. The pushrod length is 6.400". I ended up using a trick flow spring that is recommended for the cam.

Apparently the new AFR's come with ARP studs now. Don't know if they resolved the guide plate problem or not but at least they resolved the stud problem. These heads are 3 years old and had just been introduced at the time, so keep that in mind when considering a purchase.

If I was going to do this all over I would decide on a cam first and then you would know if the existing hardware would be adequate. The problem I had was I didn't know a roller cam needed more spring for a given lift, and when I purchased these heads all that was stated was that the supplied spring would handle .550" lift, but that's only for a flat tappet cam. Another problem is the rocker geometry creating an effective 1.68 ratio making the lift even greater.

I'm thinking also that these heads have a "one size fits all" position on the rocker studs. If you put taller valves, which you need for taller springs when you go over .550" lift, the taller valves would bring the geometry closer to optimum. This would bring the rocker ratio back to the proper 1.6.

 

[Jim Rosenthal]

Agree with the above regarding AFR upper valve train gear...I had my heads gone through by a fellow in Rhode Island who builds a lot fo AFR heads and he replaced the springs, retainers, etc. My engine is intended to make far less power than all of yours (385-400) but I want all the power to be usable and that means a motor that can take frequent trips all the way up the rev band. They do use ARP studs- it is one of the things we did not replace.

 

[Mark Worthington]

[ QUOTE ]
I had my heads gone through by a fellow in Rhode Island who builds a lot fo AFR heads and he replaced the springs, retainers, etc.

[/ QUOTE ]

Ed Curtis?