David Vizard
September 11, 2006
Contributers: David Vizard, David Vizard Photos By: David Vizard
The success of Ford Nextel Cup teams in the '05 Nextel Cup season was plain to see. One of the prime attributable factors was the D3 head. Seen here is a specially lightened Dirt Late Model derivative.

Finding the power to win big-time races is sometimes no more difficult than looking in the right place. When top teams find a source of power, they attempt, at almost all costs, to keep such a source secret. So far, a major factor in most Ford Nextel Cup, Busch, and Craftsman Truck series wins, namely the cylinder heads, has seemingly been kept below the radar for more than 18 months.

If you think we're about to reveal some front-running Cup car teams' key technologies, you're wrong.

That's the bad news. The good news is we can tell you about the subsequent generation of heads in question, those based on the D3 casting, that are too advanced in specification for NASCAR to approve. How good are they? Well, consider this. The Nextel Cup heads are good enough to make over 840 hp from a 355-inch, single four-barrel endurance engine-and the derivative heads we are about to look at are better yet.

The Cup car UPM D3 derivative that was tested had the intake port angled about 5 degrees at the manifold face. This improves overall intake flow over that seen on a Nextel-legal intake port.

No quantum step forward in racing ever happens by itself. There always has to be somebody who is prepared to push the envelope. In the case of the Cup car D3 castings, the guy who went out on the limb to make Fords significantly more competitive is Don Losito. Losito is the boss of Ultra Pro Machining. Although it's a sizable company, it is little known outside the ranks of top professional racers, especially in NASCAR classes.

The D3 head came into being as a substantial evolution of Ford's SC1 casting. Although it's something of a simplification, we can say it is essentially a high-port (intake and exhaust) version of the C3 head. Prior to the introduction of the D3 head, the C3 head was, from 1992, the approved Nextel Cup head. The intent of the C3 head was to service the needs of ARCA, Sprint cars, Midgets, Dirt Late Models, and drag racers, as well as the NASCAR classes. Although good for its day, by 1998 the C3 head was beginning to suffer from aging technology, and we all know what that means in a competitive sport such as racing.

In 1999, Losito set about developing some-thing better with future NASCAR approval in mind. Starting with the SC1 Ford casting, he juggled valve angles, port locations, and chamber forms in an attempt to get a more optimal overall design. But outlining what was done in one sentence really trivializes the effort involved. The reality is that every one of the many relevant aspects considered was first tried on the com-puter. If it passed muster there, it went on to the flow bench, and finally the dyno.

In the D3 casting, the exhaust port is long enough for it to morph into a round shape at the manifold face without loss of flow. This helps to make header production easier.

Most of the computer time went into analyzing situational geometry delivered by various valve angles, sizes, and positions, and the effects and interactions produced with the piston crown and cylinder wall. An issue of some importance was the effect valve pockets could have on the ring package. The better the resulting low lift flow, the wider the optimum cam lobe centerline could be. This would allow the valve to be open less at and around TDC, thus requiring shallower valve pockets, meaning the ring pack could have a higher and more optimal placement.

Computational Fluid Dynamics (CFD) also crept into the picture here, though there is reluctance to say to what extent as it is a cutting-edge F1-type technology. What is CFD? It's a method of computing airflow through and around objects. The beauty of this is it allows you to not only see where the air is flowing, but also local velocities and pressures. We have seen some of this applied, and trust us, CFD is the head development tool of the future. The only problem at present preventing its widespread use is its complexity and need for huge computing power.

Step By Step

View Photo Gallery
0609_MMFP_04Z D3_type_head Port_cross_section
This computer-generated cross section of the ports (blue intake, red exhaust) shows the generous short-side turn radii used.
0609_MMFP_02Z D3_type_head Saenz_flow_bench
UNC Charlotte Motorsports student Nathan Bornitz runs the numbers on a computer-controlled Saenz (www.saenzdynos.com.ar) flow bench. The long-term averaging capability of this bench is an asset to accuracy.
0609_MMFP_05Z D3_type_head Cg_port_flows
The port forms around the guide bosses are critical to flow. In the case of the intake (green), the guide boss also is used to influence swirl in a positive manner.
0609_MMFP_07Z D3_type_head Intake_velocity_map
The smaller-than-usual intake port proved highly active over its entire area. This gives the column of air in the intake port/runner greater ramming capability from the momentum involved.

Although CFD could do much to show pictorially where and how air was traveling through the cylinder head, Losito still chose to validate any such information by thoroughly velocity-probing all the areas of concern on the flow bench.

OK, so much for wandering off into the land of high-tech computational flow testing. Getting back to terra firma and moving on from valve angles, Losito addressed the intake port positions at the intake flange face. Using what has been developed for Cup car racing as a starting point, the port shape and position were reevaluated. Since this head was not constrained by NASCAR rules, Losito had greater freedom to be more innovative with the intake port. What he found was needed could not be had from the SC1 casting. On top of that, the relocation of the valves also meant the rocker locations had to be revised. That in itself was cause for some serious casting changes to become necessary.

At this stage, a pair of dyno-proven prototype heads were carrying an indecent amount of weld, so the next problem to address was to get a new, redesigned casting into production. Here, Losito worked closely with Ford engineers. Externally, the most obvious change, if you were to check with a caliper, was the extra 200 thousandths on the intake manifold face and the modified rocker pads.

What cannot be readily seen, however, were the changes made in the water jacket-among these changes were some to make the castings more porter friendly. The principal ones here were the thickening up of the material under the short side turn of both the intake and exhaust. Another such move was increasing the thickness around the intake bowl. Once the new castings were done, it would seem that Losito had a winner on his hands. Not quite. He is as fussy over his projects as a mother hen over her chicks. With a new head casting to work with, a substantial number of refinements were made and tested. None of these were of major significance in themselves, but when the sum total was considered, it all added up to something worthwhile.

On the dyno, the heads produced satisfying numbers, even for the most power-hungry racers. In fact, Losito selected some of these racers to do the field testing. These included multiple Dirt Late Model champ Scott Bloomquist and ARCA champ Frank Kimmel. As for drag racing, this was considered to be of prime importance as it was a venue to test the valvetrain to some stratospheric rpm in the 10,000-plus region. Though it might seem somewhat removed, the drag contingency was actually testing in order to quell the reliability fears that NASCAR Ford teams historically had with previous heads and valvetrains.

In 2001, Ultra Pro Machining canted valve heads went to Penske, Yates, and Roush, but the then-current race pressures and politics (just keeping up with the busy race schedule was more than enough to deal with) precluded any serious testing. This, in part, was because NASCAR was comfortable with the balance of power between the various brands of cars, and was reluctant to approve the new headIn 2003, both Chevy and Dodge made some serious improvements in output, thus upsetting the competitive balance between the three manufacturers competing. This caused a reevaluation of the situation that led to Ford ordering UPM's then-latest version of the Ford SC1 canted valve heads. Working closely with Roush, a test engine was built and dyno tested. Just about everyone present was impressed with the results. Ford then submitted the head design to NASCAR, and in the fall of 2003 gave the head design approval.

At this, UPM had to consolidate the head design to be sure it met all of NASCAR's template requirements for port centerlines and heights. From all this emerged the casting and port/ chamber configurations that helped Kurt Busch win his first Nextel cup championship in 2004. It also was instrumental in placing five Ford-powered cars into the chase, with 2005 seeing five more Fords added.

Since its introduction, the D3 casting-based UPM canted valve heads have won literally hundreds of races in the top and most competitive classes in the country as well as an enviable number of championships. These classes include drag racing, Dirt Late Model, ARCA, Nextel Cup, Busch GN, Craftsman Truck, Silver Crown, Sprint cars, Midget, and Off-Road. Depending on the application, there are a number of port configurations that UPM can supply. It can be very flexible in this area because it starts with a raw casting. If you want to talk versatility and big power, try 2,400 hp from a turbo motor.

Our Evaluation
So much for the UPM canted valve heads pedigree and success to date. Now it's time to take out our magnifying glass to see what makes this head such a powerhouse. Dyno testing on anything less than an all-out spec engine won't prove much, and we can't afford to build one just to show you such results. But we do know that a good Nextel Cup engine (remember, that's a flat-tappet deal) makes over 840 hp with these heads. What we can show you are the results of our flow testing on UNC Charlottes flow bench. Since big and possible controversial numbers were expected, the bench had its calibration checked immediately prior and subsequent to our tests. As a result, the confidence level of these numbers is to about the 1 percent accuracy mark.

Let's start by perusing the flow numbers in the flow chart. On the intake side, the number that most of you are going to be impressed with is the 434 cfm of peak flow. There are a lot of performance big-block heads out there with intake valves bigger than the 2.180 inch ones this head is equipped with that don't match that number.

As good as this may be, the number that really impressed us was the flow at 0.250 lift. The best head we have done with that size valve flowed 186 cfm at 0.250. On our list of all-time bests, that ranks in third place. Second was a GM head, which as it happens was a UPM head that flowed 188 cfm. At this stage of the game, a couple of cfm improvement represents a hard-fought deal.

Step By Step

View Photo Gallery
0609_MMFP_06Z D3_type_head Flow_graph
As this chart shows, the flow capability of the UPM D3 head is impressive. Peak figures of 434 and 283 for the intake and exhaust respectively are likely to set new standards for SB heads of any variety. Note the swirl curve which starts at about 0.450 lift. This may be a key factor toward faster burn rates.
0609_MMFP_11Z D3_type_head Exhaust_velocity_map
Like the intake port, the whole area of the exhaust is highly active. The only way to show what relatively small differences existed was to exaggerate such differences with large color changes as seen here. The slowest area is on the bottom of the port which nonetheless is still flowing fast.
0609_MMFP_08Z D3_type_head Closed_ports
The search for intake flow ends about 1/2 inch into the combustion chamber. In just the same way, the search for exhaust flow starts in the region around the exhaust valve.
0609_MMFP_09Z D3_type_head Open_ports
Because more of the exiting exhaust gases come from the region nearer the center of the cylinder, the exhaust port is shaped with a small but discernable bias in that direction.

We expected the 0.250 figures of this UPM D3 head to be good because of the time put in to both position and angle the valves to minimize the effect of intake shrouding, but the 195 cfm seen was truly impressive. This is the kind of flow normally seen with a valve of some 231/48 diameter (2.375). At this level of valve lift, flow is all about the seat and immediate area up and downstream of the seat. The UPM D3 head uses a 50-degree seat which favors high lift flow at the expense of some low lift flow. This makes the 195 cfm at 0.250 all the more impressive. However, the main reason for the 50-degree seat is not so much any possible flow advantage, but as a means to help suppress the valve bouncing off the seat at closure. The increased angle helps out here by acting as a damper.

The flow this port is capable of is not the sole contribution to the high output this head has shown. The port velocity is also a factor. This is in the required range because the port is highly efficient in that it does not have any significant redundant area. What that means is the whole area of the port is utilized to near maximum potential. As a result, the port looks a lot smaller than you would expect when compared to other high-performance heads.

This is part of the ingredients needed to generate horsepower through a boost in torque rather than being able to hold the same torque to higher rpm. Indeed, the small but highly efficient port is probably the primary reason why this head also produces more torque than its predecessor. The computer-generated velocity map of the intake shows how little velocity variation there is over the area of the port. These velocity test results put this head in a top-of-class category.

Exhaust Flow
If the cylinder has the potential to fill well, then it should also empty well if all the potential power is to be realized. Again, the UPM D3 did not disappoint. By a small margin, it proved to be the best-flowing 1.625-inch-equipped exhaust port ever to go across the test bench.

At the 0.750 reference point so often used by head porters, this head, without the flow-enhancing aid of a flow pipe, delivered a might over 274 cfm. That's good, but the flow curve had not by any means gone horizontal at this point. Taking the valve lift up to 1 inch, which is about where a really serious drag race engine would be, showed this port was capable of no less than 283 cfm. With a flow pipe (this simulates the exhaust pipe), the peak numbers went over the 300-cfm mark.

Like the intake, the exhaust port showed a uniform velocity map. Having a small, efficient port is good for improved high-speed chamber scavenging and widening of the torque curve. That's just what an oval track racer needs to get off the turn faster, especially on tracks that have tight turns and long straights.

Swirl
The value of swirl is always in debate with race engine builders. It's one of those things that interacts so much with other factors that it is hard to pin down exactly what is wanted. A few years back, the big deal was not to have any swirl for fear of centrifuging the fuel out of the intake charge in the cylinder. However, this proved to be a problem only if the effect of lower intake charge temperatures and fuel droplets a little too big for the job were not properly addressed. Cooler intake charges from thermal-barrier manifold coatings meant the fuel leaving the carb boosters needed to be more finely atomized. When that was done, swirl started to get back into favor.

But a lot of swirl at low lift might mean the opportunity for some of the intake charge to go out of the exhaust port during the overlap period is increased. Part of the reason why the UPM head works well could be that at the lower lifts (when both valves would be open) the swirl is nonexistent. As lift progresses, the swirl builds to an appreciable level and is retained throughout the compression stroke and into the ignition phase. This, in theory, promotes a faster burn, which at the 10,000-rpm mark is definitely a positive factor.

An indicator here that our theoretical faster burn actually takes place is that this head typically needs about 2 maybe 3 degrees less timing than its contemporaries. The ability to use less advance means less pressure rise while the piston is on the way up the bore, without any sacrifice of pressure rise when it is on the actual power stroke.

Supply
It's pretty evident from the forgoing that this UPM D3 Ford head is a hot commodity. Because it is available in a number of versions to suit a specific application, Losito deals only with professional engine builders. Now this may seem somewhat offhanded, but when you know why, it will all make sense. To stay on top of his game, Losito must spend just about all his time trying to improve on whatever is the best today, otherwise tomorrow any one of a number of smart guys out there may have found what it takes to snag the No. 1 spot. It is only by virtue of a huge amount of effort and time on the flow bench and dyno that remaining totally competitive is possible.

If you believe these heads are for you, then get them through an engine builder who is recognized as being successful and competent in the type of racing you are into. But before you call, be aware that big horsepower usually costs big money, and things are no different here as a set of these heads with external CNC lightening, less valves, springs, and so on will set you back about $6,400.

Sources

Engine Builder Only Contact:
Ultra Pro Machining
Dept. MMFF
6350 Brookshire Blvd
Charlotte, NC 28216
(704) 392-9955
www.ultrapromachining.com

Engine Builders To Contact:
Drag Racing
Bennett Racing Inc
Dept. MMFF
1640 11th Ave
Haleyville, AL 35565
(800) 240-7223
www.bennettracing.com

Bob Panella Motorsports
Dept. MMFF
5000 E. Fremont St.
Stockton, CA 95215
(209) 547-1778
www.panella.com

Dirt & Asphalt Late Model & Road Race
Custom Race
Dept. MMFF
W. 2010 Governor John Sevier Hwy
Knoxville, TN 37920
(865) 573-1449

Cornett Racing Engines
Dept. MMFF
1647 S. Hwy 27
Somerset, KY 42501
(606) 678-2226
www.cornettthunder.com

Truck and Off-Road
Rancho Performance Machine
Dept. MMFF
28073 Diaz Rd., Ste.
Temecula, CA 92590
(951) 676-9292