Consider for a moment what it might take for an engine to make 150 horsepower per liter of displacement. It's an impressive ratio, but not necessarily implausible. It'd be nice to have, but hasn't the tuner community already been there, done that? Well, it depends a lot on context.
In the era of 10-cylinder, 3-liter powerplants, Formula 1 engines were said to produce twice this level, over 300hp per liter. Put a snail on just about any OE Honda mill and instantly 150 per 1,000cc doesn't seem so inaccessible. A stock block 1.8 GSR at 10psi will typically deliver about 270, which equates to exactly 150 per liter. But a normally-aspirated street motor that must try to make efficient explosions on crap 91 octane West Coast gas? Now that's a reality TV-style challenge.
Obviously it will take revs to meet this power goal. It will also take just the right combination of parts. Finally, it'll take the right person who has the experience and is up to the challenge. We found that person in Larry Widmer, owner of Energy Dynamics, or Endyn, in Fort Worth, Texas.
Widmer is one of those people you would have to label a character. He is old-school V8 from the '70s and '80s era of Vic Edelbrock, Richard Petty, and Roger Penske (yes, Penske is that old), among others. As a drag racer, he competed in what is today called the Pro Stock class. In road racing, he made engines for nearly every successful Ford-based NASCAR team of the '70s and '80s, plus a few Indy Car engines. As a tuner, his motors-specifically the headwork-have won an Indy 500, a NASCAR title, and multiple Pro Stock titles. He's also worked for Toyota, General Dynamics, MSD, and most importantly, himself.
He is also a cancer survivor who greets such grimness with a defiant middle finger every morning. Today, he is doing what he loves, Honda engines, and "doesn't give a you-know-what if anyone thinks I haven't done what I've said I've done, or if I'm full of shit." Quotes like that just scratch the surface with Widmer, so instead of the usual tech story routine, the real substance of what makes this project engine tick will be described by the man himself.
Honda Tuning: Larry, we want to build a normally aspirated engine that can produce 150hp per liter [on an engine dyno]. It has to run long term on 91 octane gas, idle well enough to drive air conditioning, get decent mileage, and deliver the kids to school, as well as traverse the drag strip. Can it be done?
Larry Widmer: I can do it. All my street engines are just that, street engines. They have to start in the morning and run all day. I stand behind what I make.
HT: You sound like you've done this before.
LW: I've been building engines since I was a kid racing karts. I built a '65 Mustang small block into a big block stomper, and by the time I was 20, I had a Ford factory-sponsored drag team. When Ford pulled out of racing, I started to do NASCAR heads for Penske and the Elliot family. I've done heads on the 4-valve Cosworth, 2-valve Fords in NASCAR and Pro Stock, and of course these lawnmower engines. Today it's nearly all Honda. Those engines are the best factory engineered things I've seen in 30 plus years of doing this.
HT: Where should we start?
LW: I think a Dart block is a great foundation. They have a B20 version with a tall deck, so I can put longer rods in, and it's strong as hell.
HT: Why the Dart block? We're not going turbo (yet - wink, wink).
LW: Like I said, Honda builds an incredible engine. Only recently have high-end race engines started using machining tolerances as tight as Honda's factory production pieces. But to get 300hp, we're going to need more displacement than any stock B-series block is capable of. We'll have to buzz it higher than a stocker, too. That leaves us with 2 block options: we can either sleeve an existing Honda block, or use a purpose-built Dart.
HT: Why a sleeved stock block and not just a stock block?
LW: A sleeved open-deck block isn't as rigid as we'd like for the horsepower and RPM we're going to generate. But stock B-series cranks are incredibly strong; I've seen them on 800hp turbo motors. What actually happens is the crank keeps things together while the block flexes, so we need a rigid block and that's what the Dart provides.
HT: Still, why not sleeve a stock block? Isn't it a lot cheaper?
LW: Yes and no. The cost of a Dart block is less than most people think, and there are problems with sleeving.
LW: First is cylinder roundness. If you look at a sleeved block, you can see where the sleeves butt together; they have flat areas on them where they meet. When you have material of varying thicknesses, it expands at different rates. Consequently, as temps go up, the cylinders don't tend to stay round. Anyone who's ever built engines for serious competition will tell you ring seal is everything when it comes to making power.
The Dart block has sleeves that are the same thickness all the way around, so when the block heats up, even in machining, the cylinders stay round. Straight bores are something that Dart has specialized in for years. The Dart casting also has heavy main bearing webbing, so it supports the bearings and crank instead of the other way around. When you sleeve a Honda block and you mill out the original cylinders, it causes the whole bottom end of the Honda block to relax and the mains go out of alignment. When the sleeves are pressed in, it forces them even further out of alignment. In order to fix this you would have to align-bore, or align-hone the mains, and there aren't many shops doing that correctly, keeping the crank centerline precisely where it needs to be. This is important, because any crank offset will dramatically shorten the life of the oil pump. This process also needs to be done with a torque plate in place.
HT: Ok, we get that taking out material from the Honda block makes it less stiff on the bottom, but explain align-hone and the torque plate.
LW: An align-hone, or line-hone, is an extremely rigid [and round] bar with an abrasive coating that's slipped though the mains of the block. It's expanded and rotated simultaneously, boring all the main journals to the same size, as well as making them perfectly straight. You can imagine how a block might sag a little on the ends so everything's not in a straight line any more. When the mains aren't straight, the crank won't rotate freely. Even on a properly machined, sleeved block, at high stress levels the only thing that really keeps the mains aligned is the crankshaft's rigidity. Fortunately, Honda's factory crankshafts are more than up to the task, but in an ideal world where we're looking for both power and longevity, the block should support the crank.
A torque plate is a thick [flat] plate that bolts to the top of the block and simulates the head being torqued down. Regardless of block used, when the head, with head gasket, is torqued into place, the cylinder bores will distort. If you want your cylinders to be straight and round, it's essential that you simulate the installed head by using a torque plate during machining. We also have the torque plate on during the line-honing to properly stress the block. If you don't do these things, you can wind up with a machined block that is straight all by itself, but then different once the head is bolted in place.
HT: That seems like a lot of work. You don't like sleeved blocks at all?
LW: Open deck blocks provide the best upper cylinder cooling. That's why the factories use them. But like I said, the stock B-series block doesn't have the strength capacity we need for the target RPM power level. We build a lot of sleeved-block combinations, so we have experience with every sleeved block out there. At this point, the ERL sleeved-blocks seem to be the best in the marketplace, but the Dart block with its strength advantage is available with a higher deck and, at about $500 more than a sleeved block, it's well worth that extra money. And if I'm puttin' my name on this 300hp engine, we'll use the Dart block.
HT: The bore from Dart is 84.5mm. If it's true that there's no replacement for displacement, how big should we bore this thing?
LW: It'll stay where it is. They're really 84mm, but they need finishing work so Dart says 84.5. We could go as large as 85.0mm, the limit on a conventional Dart block, but I always prefer to leave a space for clean up in the future. One of the advantages of the 84-plus millimeter diameter is the better breathing it affords by unshrouding the valves. I have yet to see any breathing advantage from increasing the bore even more, so going larger than this with any block will net less head gasket clamping space between the cylinders, increasing the likelihood of failure. If you're doing a race engine and you're ready for rebuilds then OK, but you said you wanted a daily driver and I expect a daily driver to last about 5 years, so the bore is 84.5mm. Plus if you f*ck up, it can be re-bored to 85 and put back together.
HT: Stroke is another issue. A B20 has an 89mm stroke; do we want bigger than that?
LW: It depends what you want to do. In my opinion, the best stroke for a turbo motor is the stock GSR stroke, just over 87mm. But since we need more displacement for an N/A motor, you can go bigger, assuming the rod ratio doesn't get too short for reliability. I've done several 95mm strokes, but I don't really like the cylinder and bearing wear I've seen. You also start to get into some harmonics issues. I have a saying from the old Junior Johnson NASCAR days: "You can't fit 10 pounds of shit into a 5-pound bag." That's the way I look at this.
HT: What's the magic stroke number?
LW: I've built a few different combinations at 92mm and I think that is the right number. With a custom length Crower rod and matching crank, we still get a good rod ratio that'll keep side loading down and should spin past 9,000. Using a tall-deck block allows us to use a longer rod, so friction and associated wear will be low for a Honda engine of this size.
HT: That's a lot of RPM, more than an S2000 can do. Is that alright? Will it actually make power up there? Plus, what did you say RPM stood for?
LW: "Ruins People's Motors," and it is a lot, but you'll be amazed at how much low end this engine will have. It'll stay together and make power, and depending on the manifold I make for it, make peak power past 9,000.
HT: In the B16 versus GSR head debate, which will you opt for?
LW: This month, I like the GSR. One of the big things about the GSR head versus the B16 is the quench pad. On a B16, the pads are recessed about 0.038-inch in the head, but the GSR runs flat quench pads, or the head is the quench. I like using quench this way.
In the combustion chamber, the B16 has bulges on each side between the intake and exhaust, and this can be used to influence airflow. The GSR head is not designed this way. The intake port geometry is also different. When you look at them what you first see is the steep angle of the injector on the GSR, and that's because of the 2-stage manifold. The B16 ports are actually lower in the head and lower as they enter the combustion chamber; they're not as steep as the GSR. Airflow studies suggest that the GSR has more of a random turbulence to it, where the B16 is more controlled.
We can make a bit more torque with the B16 head, and I use them on most road racing engines we build. The exhaust is kind of the same way. The GSR exhaust ports have a fairly short bottom side radius, so the exhaust gas has an abrupt direction change as it exits. The B16 has a higher floor to make that radius less, so what we do is weld in the floors on the GSR and then CNC our B16 exhaust port to get the best of both.
HT: It sounds like the B16 is the better head, though you're saying you like the GSR one. Since Honda already had a good head on the B16, why would they create the GSR head?
LW: I think with the greater displacement and the dual runner manifold, another head design had to be made. But if you do the same aggressive valve job on both stock port heads, we can get a bit more high RPM power with the GSR. The port volume and angle lends itself to high revs. But it's inevitable I'll find something new the next time I work on a radical B16 head and it'll become my favorite again. Then in 6 months the reverse could happen. That's how I am on these things; the heads I do are continuously evolving, based on flow bench studies and analysis on our dyno. Each and every head I do is a custom piece, configured for an exact application.
HT: Then we're all set. You'll build a Dart B20 with an 84.5mm bore, drop in a Crower 92mm crank with custom length rods to take advantage of the taller block, and then top it off with a custom GSR head.
LW: Don't forget, we'll be using Endyn's pistons, cams, valvetrain and intake manifold as well.
HT: We'll save that for next time!