This month we wanted to test electronic fuel injection (EFI) again. But this time we wanted to use a throttle body injection (TBI) system like the ones that came stock on GM trucks for many years, instead of a multi-port injection (MPI) system like the F.A.S.T. unit we tested early this year (see "Danger Mouse, Part 11," August 2003). So we dusted off a Holley Commander TBI kit that had been lying around the office for quite a while and took it and DM back to Vrbancic Brother's Racing DTS dyno with high expectations.
Then our hopes were kinda shattered. Once before in Danger Mouse's history did we test the wrong combination and lose power because of it. This is now the second time. However, the mismatch did have its benefits, mainly at low rpm where the TBI added power over the HP750 carburetor we baselined the engine with. What we learned from this test is that even if you have a great running engine, you can't just slap any old EFI system on and go. There are several key components that must be addressed in order to make the EFI work best. You must also keep in mind that EFI will rarely make more peak power than a well-tuned carburetor. The benefits from EFI, as we've again proven this month, are better low-end power and drivability, easy starting, and all-climate running.
WHAT WE DID RIGHT AND WRONG
We didn't pay as much attention to what was underneath the TBI as we should have. As we've said in the past, EFI needs the proper cam to really shine. The engine will still run with the wrong cam, but power might suffer. And the healthy COMP Cams Xtreme Energy XR282HR hydraulic roller camshaft was not the best choice for this EFI set up due mostly to its longer duration and narrow 110-degree lobe separation. What we should have done was reinstall the COMP Cams prototype EFI hydraulic roller cam that we used with the F.A.S.T. EFI system. But, we thought you might want to see the results of just swapping the carb to EFI, since that's the easiest thing to do. With a cam that wants to keep breathing well above 5,000 rpm and a throttle body that basically runs out of air at 5,000, there was little chance this engine would make any more peak power. The power increases we did find came in below 3,000 rpm, where the precise metering of the TBI allowed the engine to build more torque, which is actually way more beneficial to almost any car than an extra 10 or 15 peak hp would be.
Another area we've learned an EFI engine can benefit greatly from, but still ignored in this test, is the method used to trigger the Engine Control Unit or ECU. In this case, we simply hooked a wire to the dyno's MSD 7 ingition's tach output, which in turn triggered the ECU to fire the injectors. This method is crude at best and not very conducive to accurately firing the injectors at the precise time. The best way to get the signal to the ECU is with a crank trigger, just like the ones that all OEM's now run. But, there's another choice in between that we've used with much success in the past. It's a simple electronic distributor. It's not your typical HEI however, instead the distributor's pickup triggers the ECU as well as firing the spark plugs. Another benefit of running an electronic distributor, which we will include the next time we test, is that the Holley ECU can control ignition advance with it. That means that instead of loosening the distributor hold down and turning it with a timing light pointed at the dampener, when we want to make timing changes we just tap a few keys on our laptop while the engine's running. Ignition advance changes can even be made under power in this fashion. There's really no better way to trigger things.
WHAT HOLLEY TOLD US
Naturally, we wanted to get Holley's expert opinion on our mixed results. Woddy in Holley's tech department told us that we were trying to control an outdated throttle body with a fairly new ECU. It turns out that since we got the Commander TBI kit more than a year ago, Holley has redesigned the four-barrel throttle bodies to work progressively, meaning the front two injectors run all the time, while the back two injectors are brought in progressively as demand requires. Our older throttle body fired all four injectors at once, all the time. That wouldn't have been too bad of a problem except that we were firing the injectors using a fuel map in the ECU that was programmed for the progressive throttle bodies. So we basically had to rewrite the entire fuel map on the dyno otherwise the engine ran so rich it wouldn't even idle. Once Woddy told us about a few changes we could make in the software's Engine Parameters section to help things out, we were able to fully rewrite the map and make the engine run. But it took several phone calls and about 4 hours and over 60 pulls on the dyno.
HOW CAN WE LEARN FROM THIS?
The best advice we can give from our test results is to research your application before buying any EFI system. There are a lot of really good systems out there, but each one has a specific area that was engineered to work best in. And if you're thinking about finding a great deal on a used EFI system at the swap meet, think twice before you buy. The technology in these things advances so fast that a 1-year-old system might take more time and money to make work than a brand new system would.
Like the ending of a good movie, the end results we got were not what we'd hoped for. So we're writing a sequel to this test. We're going to get a new smaller cam, the correct progressive throttle body, and even a new computer-controlled distributor and re-test DM again to see if there's any more power to be found. Of course, we'll still compare results to the HP750 carburetor only this time the cam will be smaller and better suited to work with EFI.
DYNO TESTING PART 14
This month's dyno test did not involve any direct comparison to previous months tests. We tested DM as if it were an entirely new motor this month. We did, however, reuse several of our original key pieces, like the same Motown block and Lunati rotating assembly as in all the other tests. The results shown here are listed chronologically as tests 27 and 28, just in case we want to refer back to them at a later date. This month was also the second time we bolted DM onto the Vrbancic Brothers Racing DTS dyno in Ontario, California. Shop owners George and Bob Vrbancic both supervised the test sessions with George at the dyno controls. Faithful DM followers may note that peak power numbers for this test were down by a few points compared to last month's test (Part 13) with the same engine combination at Westech. That discrepancy is most likely just the difference between how a DTS dyno and a Superflow dyno collects and interprets data. It's kind of like stepping on the scale at your house vs. your friend's house, either scale may be correct, but you'd still weigh the same regardless of the difference in readings so don't look too deeply into the power differences between different these two dynos.
Danger Mouse specs for Part 14:
355 cid, 9.4:1 cr, 4.030-bore four-bolt Motown block, 3.48-stroke Lunati crank, 5.7-inch Lunati rods
Test 27: GMPP Aluminum Fast Burn heads (62cc chambers, 210cc runners, 2.00-inch hollow stem intake valves, 1.55-inch sodium filled exhaust valves), GMPP single-plane intake manifold, Holley HP750 carb with 73/84 jets, COMP Cams Xtreme Energy hydraulic roller camshaft XR-282-HR (230/236 at .050, 282/288 adv, .510/.520 lift, 110 LS). GMPP 1.5:1 aluminum roller rockers, 36 degrees total advance.
Test 28: Installed Holley Commander TBI900cfm four-barrel electronic throttle body fuel injection.
|DANGER MOUSE PARTS LIST FOR TEST 14|
|EFI||HOLLEY COMMANDER TBI PN 950-21S (correct at press time)|
|CAM||COMP CAMS PN 12-432-8|
|HEADS||GMPP PN 12464298|
|LIFTERS||COMP CAMS PN 885-16|
|MANIFOLD||GMPP PN 12496822|
|PUSHRODS||COMP CAMS PN 7809-16|
|ROCKERS||GMPP PN 12370838|
|GASKETS||MR. GASKET (head gaskets PN 5727), GMPP (intake gasket PN 12497760)|
For this month we used a regular MSD billet distributor with locked-out timing. The sequel for this story (coming soon) will include a new, computer-controlled distributor and ignition package from Holley.