I ran this dude from a 50 roll and beat him by two car lengths, and last week he barely lost against this guy in a Camaro that runs low 11s and makes like 500 hp, so I figure I’m making 600 hp.
Right. Horsepower by proxy.
This kind of guesstimation is rampant in the hot rodding community and, truth be told, is the basis for most of the bench racing that goes on. Comparing an unknown or untested car against one with a real history of track or dyno time is the only way most of us can relate one combination to another. The question then becomes, how much power do you really have? How are you going to check it? Are you sure you trust the numbers?
We posed this question to several racers and their answers provided the impetus for this comparison. You see, some gave answers in terms of flywheel horsepower, some with wheel horsepower using an eddy current dyno, and some using an inertia dyno. So which one is right? Searching for the answer put us on a rocky path to enlightenment.
We began our journey by having a straight-up talk with chassis dyno owner/operator Erin Carpenter of Carma Performance Engineering in Nashville. In between blasts from a tunnel-rammed 362-cube LS-powered S-10, he explained that most dyno customers show up with a posse of their buddies wanting to puff their chests and get a sheet they can proudly display on the Internet. Forums and dyno sheets have become the new dragstrips, where e.t.’s have been replaced by peak horsepower numbers, and actual performance be damned. Unfortunately for some of Carma’s customers, their bubble gets burst when they see the true power their car makes. We almost never use a correction factor when we dyno, because the numbers that it puts out are actually what it is doing on that day, with that weather, on that dyno.
Before running the “real”...
Before running the “real” tests, we ran the engine with air running through the old intercooler and bonnet and discovered that it robbed 93 rear-wheel horsepower! We ended up running the rest of the tests with the bonnet removed and a fan blowing fresh air toward the front of the car.
So how is it that the car that is strapped to the dyno doesn’t make nearly the power as that other one on the forum with the same setup? Carpenter says a shop looking to boost their business could easily boost power numbers and egos by fudging correction factors. You could change the correction factor if you wanted to. You can also go in and reset barometric conditions or temperature. Another thing on some dynos is that they’ve got a place to set the drivetrain inertia. So on a motorcycle you’d put the inertia really low, but you could jack it way up, and it would correct and show huge numbers.
OK, so what are these correction factors? Basically, they just take the current weather conditions and modify your dyno curve so it reads as if you were at a barometric pressure and temperature standardized by the Society of Automotive Engineers (SAE). Ideally, any given engine in any weather condition on any dyno should give the same power curve, but is that really the case?
We’ve made tons of dyno pulls on engine and chassis dynos, but we’ve never really taken the same engine and run it on an eddy current chassis dyno, an inertia-style chassis dyno, and a water brake engine dyno to see how they all compare back to back to back. We also wanted to see what those numbers added up to in terms of real performance at the strip. Fortunately, there happened to be a big-blockpowered Nova ripe for the test.
Our test mule was a ’70 Chevy Nova that has proven reliable and repeatable as a race car, paramount in the comparison. Equipped with stock-style suspension, the car has been as quick as 8.67 at 158 mph using a twin-turbo setup. The hair dryers were pulled off and regular Hooker Super Comp headers were installed to bring the car into a more real-world status.
With a 9-inch rearend housing, 3.73 gears, a Powerglide trans, and a 3,500-stall converter spinning BFGoodrich drag radials set at 30 psi, the driveline was as common as anything else.
The tandem rollers on the...
The tandem rollers on the Dyno Dynamics dyno showed the tires in perfect shape when rolling with no load. Under load, however, it’s easy to see how the tires begin to wrap around the smallish 12-inch roller. Future testing at various tire pressures and compounds is in our plans.
The engine was a pretty big 555-cube big-block Chevy running Pro-Filer heads and a street-friendly COMP Cams solid roller. With 8.9:1 compression, our engine was happy to gulp down even the lowest rank of pump gas swill and run like a champ. Of course, tuning would be a major factor in making sure the engine lived, and fortunately the use of a Holley HP EFI system allowed us to dial in the engine on the first dyno in our lineup, then keep it completely unaltered throughout the remainder of tests.
Eddy Current Chassis Dyno
Test and tune began at Carma Performance Engineering where Carpenter and co-owner Brad Mayo strapped the deuce to their Dyno Dynamics chassis dyno. Their eddy current dyno, similar to those offered by Mustang or Land and Sea, operates by having the wheels of a car spin a pair of drums that has a metallic rotor attached to the end of one drum. The rotor, which looks like a giant brake rotor, spins within a field of electromagnetic coils. The operator basically adjusts power to the coils, which can slow the spinning rotor down or allow it to speed up, and measures the power required to do this. Then the dyno software does the hokeypokey to determine the horsepower produced.
It sounds a bit complicated, and it is, but the upshot is that with this type of dyno a load can be applied and held at any rpm or allowed to run a sweep through the rpm range, giving tuners a great chance to complete 95 percent of driveability and race tuning without the tires touching pavement.
On our little project, Carpenter began the process of converting the tune from a turbocharged race-fuel setup, to a naturally aspirated pump-gas tune by merely connecting his laptop to the HP EFI ECU. Using a wide-band oxygen sensor stuffed in the tailpipe and comparing against a second oxygen sensor hooked up to the HP EFI, he was able to simply enter the desired air-fuel ratio into his laptop, hit the Self-Learn button and let the Holley system do most of the grunt work getting the fuel map set up. Final tweaking was done manually, and it was time to see how the car baselined.
Running the engine to a harmoniously high rpm without any bad noises was music to our ears, though we were slightly surprised at the dismal power the computer screen displayed. We saw the raw power numbers and asked Carpenter to display the corrected numbers so we’d have a better comparison against the other dynos. Unfortunately, with the weather being good and cold, it actually corrected down to 487 horses at the peak point of 5,400 rpm with torque maxing out at 476 lb-ft at 100 less rpm. The power never went higher, so we stopped the pulls by 5,700 rpm.
We figured the converter and automatic tranny hurt us some, but still had hoped for more considering the cubes involved. But hey, maybe the next test would boost our spirits!
Test two brought us to Carthage and the Dynojet chassis dyno of C&C Motorsports. Owner Casey Lawson is a regular drag racer and has run his dyno long enough to see that actual dyno numbers versus track times can vary greatly by the torque converter and tire design. As we strapped the Nova down, he related that the inertia-style dynos, while lacking in steady state control, offer a more real-world version of power because they allow the engine to accelerate at its own speed, like at the track, rather than a predetermined speed.
Overtightening the ratchet...
Overtightening the ratchet straps to keep the car in place is something that Casey Lawson of C&C Motorsports tries to avoid. He’s seen that cause a significant loss in measured horsepower. Under no load, the tires look totally normal on the Dynojet’s big 24-inch single roller.
Unlike eddy current dynos, inertia dynos use some cheap, basic physics to determine horsepower. The car rotates a drum that has been precision machined and measured to determine the Polar Moment of Inertia (J) of the drum. That Polar Moment is then multiplied by however many rpm/second the drum is accelerating as measured by an encoder on the end of the drum and divided by a constant to give a torque number. Horsepower is then easily derived by multiplying torque by rpm and dividing by another constant, 5,252. No complicated electronics other than the encoder is required, which is what makes the price of inertia dynos so much more reasonable than their eddy current cousins.
No tuning changes were made and fortunately the environmental conditions, as with the remaining tests, were almost identical to the original test, so repeatability should be guaranteed. Surprisingly, the results were anything but repeated. The car made almost 100 more horses on the Dynojet! Peak power came at 6,200 rpm, this time at an SAE-corrected 564 rear-wheel horsepower with torque reaching 522 lb-ft at 5,400 rpm. We’d seen some small differences in dyno numbers in the past but nothing this big, so we were a little confused and wondered if there was a way they could have had a proverbial thumb on the scale. Lawson showed us the uncorrected numbers though and sure enough, it corrected down again. He assured us that with the Dynojet, the weather conditions were measured internally, and they didn’t have a way to fudge those numbers.
Shacklett Automotive Machine in Nashville is where we’ve done several dyno tests and felt very comfortable trusting the accuracy and reliability of their SuperFlow SF901 engine dyno.
Even at full-tilt boogie,...
Even at full-tilt boogie, there was no visible deflection of the tires on the Dynojet inertial dyno. Lawson pointed out that though these are technically radial street tires, they are sticky enough to rob some power. He’s seen soft-wall drag slicks drop the measured output over 50 rear-wheel horsepower and suggests a “normal” street radial for dyno testing.
Very similar to a torque converter, a water brake dyno like the SF901 uses a viscous coupling of water to the housing to measure work. A driveshaft hooked to the engine’s crank spins an impeller inside the brake housing as water is fed into it. The water couples with fins on the inside of the brake housing and tries to rotate the housing. The force of that rotation is measured by a strain gauge, which records the torque produced. Horsepower is, again, a simple math equation away.
Normally, an engine dyno test would be done at the beginning of an engine’s lifespan, but since we had plans to freshen up the old girl we decided to do this test last since we had to yank the engine anyhow. Making every effort to be sure test conditions were as close as possible to the chassis dyno tests, we installed the same full exhaust and mufflers as run on the car as well as the alternator and vacuum pump. Again the tune-up was unaltered and aside from a few hours of setup time, the engine fired right up.
The temp was a few degrees cooler, but the relative humidity and, more importantly, water grains were a little higher balancing out any advantage. The results were once again surprising. It made 675 horses at the same 6,200 rpm and 653 lb-ft at 5,000 using an SAE correction factor that bumped power up this time. The rpm range looked the same but the power numbers were far from close.
Regardless of whatever number the dynos spit out, none of them would be relevant if we didn’t have a timeslip to know how they relate. It wasn’t easy finding a dragstrip open in the middle of January, but after several calls we discovered that Montgomery Motorsports Park had an eighth-mile race the weekend after we finished all the chassis dyno pulls, so we got up early and headed south for warmer weather. It turns out that the track temp was hovering around 48 degrees. Lousy for traction-starved bracket racers, but good for us since that was right in line with the conditions we had on all of our dyno pulls, and it would show real-world equivalencies.
The Dynojet basically acts...
The Dynojet basically acts like a dragstrip in a building, using only acceleration to determine horsepower. Stick shift cars typically show more power on eddy current dynos, but often high-stall automatic cars will show more on an inertia dyno since it doesn’t keep the converter from doing its job properly.
The car scaled in at 3,501 pounds with owner Shannon Carnathan at the wheel and a tank full of Shell 91-octane in back. We cycled the car down the strip several times to get a good read on what it would do. Sixty-foot times ranged from 1.58 to 1.61, and the car was consistent enough that it ran 6.6s at 106 mph. Good enough for us, so we loaded it up and made the five-hour drive home, digesting all of the information at hand and trying to make sense of it all.
This exercise raised as many questions as it provided answers and required further digging to get to the root of them. It seems the biggest question we had was how the numbers were corrected from the raw data. All the dynos we used had an SAE correction factor but it turns out that they all used different factors. On the eddy current dyno, they used a correction factor SAE J1995, the Dynojet used SAE J1349, and the SF901 used the SAE J607. A lot of letters and numbers that didn’t mean anything yet.
We were fortunate that the...
We were fortunate that the Hangover Nationals, usually raced New Year’s weekend, were moved back three weeks. Montgomery Motorsports Park had their track prepped well, and consistent e.t.’s were the payoff. The car normally runs 5-second e.t.’s with twin turbos, so slowing it down over a second made the passes uneventful and almost boring.
Basically, J1349 corrects the atmospheric conditions to that of 77 degrees F, zero percent humidity, and a barometric pressure of 29.234 inches Hg, allowing for accessories, full exhaust, and emissions equipment, and gives the net horsepower of the engine. Standard J1995 uses those same conditions, but excludes accessories, giving us the gross output of the engine, which tends to be about 20 percent higher than net. Standard J607 provides gross output as well, but corrects to a more lenient 60 degrees F, zero percent humidity, and 29.92 inches Hg. In the end, all this means is that these should give respectively higher horsepower numbers. This would all make sense if the results from our eddy current and inertia dyno numbers were switched, but they weren’t. So there must be something else amiss here.
After looking through pics of our dyno sessions, we noticed a significant amount of tire deflection on the Dyno Dynamics dyno, which had two smaller drums instead of the one large drum like the Dynojet. While the tires are basically centered during the strapped-down and unloaded-yet-rolling state, once the engine goes to WOT, the tire climbs forward on the drum. With the one large drum, deflection didn’t really change but on the twin roller setup, the tires were majorly out of shape and the horsepower-robbing deflection got worse as rpm increased. It looks like the design of the dyno itself might have been the source of our low numbers on Carma’s dyno. Perhaps bumping up tire pressure to 45 or more would change our readings. But maybe not. Altering the atmospheric conditions entered into the computer would also show a different power reading, but typically that would be skewed upward if we were dealing with an unreputable dyno operator, which we weren’t. Maybe, like flow benches, some dynos are just stingier than others. What we do know is that plugging 675 horses and a 3,500-pound race weight into various e.t. calculators shows that those numbers are pretty accurate.
It wasn’t easy but we squeezed...
It wasn’t easy but we squeezed the mufflers and all into the dyno cell and ran the engine exactly as equipped in the car. As a final measure, we even used the fuel pump and regulator off the car and ran the engine in open-loop feedback to be sure no tuning changes would show up.
There is no doubt that the vast majority of dyno operators are as straight up as those we frequent, but you can bet that if someone is bragging about dyno numbers that are too good to be true, they probably are. Ask to see the uncorrected numbers and the weather conditions. You can correct your own numbers using calculators online. If they don’t add up, then you know the dyno shop is cheating bigger than a slacker in finals week. As for us, we’ll just brag that we have a 675 hp car that can beat some guy’s car who barely lost against a Camaro that makes like 500 hp.
On The Dyno
| ||Eddy Current ||Inertia ||Water Brake || || || |
|RPM ||TQ ||HP ||TQ ||HP ||TQ ||HP |
|5,000 || || || || ||653 ||622 |
|5,100 || || ||502 ||488 ||649 ||630 |
|5,200 ||465 ||460 ||513 ||507 ||649 ||642 |
|5,300 ||475 ||478 ||518 ||523 ||649 ||654 |
|5,400 ||473 ||487 ||522 ||536 ||642 ||660 |
|5,500 ||456 ||478 ||520 ||545 ||634 ||664 |
|5,600 ||435 ||465 ||518 ||552 ||621 ||662 |
|5,700 ||408 ||441 ||512 ||556 ||612 ||664 |
|5,800 || || ||508 ||561 ||608 ||672 |
|5,900 || || ||501 ||562 ||597 ||671 |
|6,000 || || ||493 ||563 ||590 ||674 |
|6,100 || || ||484 ||563 ||576 ||669 |
|6,200 || || ||478 ||564 ||572 ||675 |
|6,300 || || ||465 ||558 ||562 ||675 |
|6,400 || || ||459 ||559 ||552 ||673 |
|6,500 || || ||450 ||556 ||541 ||669 |
Car owner: Shannon Carnathan
Rear suspension: CalTracs monoleafs, CalTracs bars, Afco shocks
Front suspension: tubular control arms, Afco shocks
Rearend: 9-inch, 3.73 gears, Eaton TrueTrac
Tires: Mickey Thompson 315/60R15 Drag Radial
Fuel injection: Holley HP EFI
Cylinder heads: Pro Filer 320, ported by Revolutionary Performance
Camshaft specs: COMP solid-roller; 268/272 degrees at .050, .721/.721-inch lift
Headers: Hooker Super Comp
Mufflers: AutoZone 3.5-inch