Powered by a mild 350 small-block shifted by a TH350 automatic trans, Bill Ancona's '37 Chevy two-door sedan had delivered decades of reliable service. But after 22 years and 30,000 miles, Bill started noticing small fuel drops on the garage floor below the car's old inline electric minipump. Figuring it was about to give up the ghost, he replaced the pump with an apparently similar-size and similar-capacity universal unit from the local auto-parts store. Then things went south.
After changing out the pump, the engine died whenever it reached operating temperature, not restarting until it cooled back down. Bill explains, "When the engine dies, you can jump out, pull the air cleaner off the engine, and stroke the throttle; there's no accelerator-pump squirt. I know what you're thinking: vapor-lock!"
Trying to solve the problem, Bill began a seemingly endless parts-swapping odyssey that ultimately proved futile. "I changed out the fuel pump again," Bill relates. "I rerouted the fuel lines to avoid possible heat from the exhaust. I replaced the fuel filter twice, including moving up to a fuel cooler/filter placed outside the frame in the wheelwell to assist in cooling the fuel. I replaced the fuel tank with a new poly tank that is properly vented and has a rollover valve. I replaced the wiring and the relay for the fuel pump. I ran independent/additional ground wires for the fuel pump. I rebuilt the carburetor. I placed a phenolic plate under the carburetor. Did I mention I still have the problem if I run it without the gas cap in place? You tell me what's wrong, because I sure don't know!"
Not only did these fixes not work, the car--which for years never ran higher than 180 degrees F, even in balmy, 95-degree Fort Worth, Texas, summers--began running hot at more than 200 degrees F, pushing coolant out the overflow tank when shut off. That only increased the frequency of the shutdown problems.
By this point, Bill had replaced or upgraded just about every fuel-system component. It was time to seek professional help, so HOT ROD sent him over to Speedtek. Located just northeast of Fort Worth in Haltom City, Texas, Speedtek is one of a dwindling breed of friendly and reliable local neighborhood speed shops with a sales counter/showroom up front and service bays out back that service both early- and late-model vehicles. As Speedtek owner Keith Sullivan puts it, "We still service what we sell, and sell what we service."
Speedtek is about a half hour drive from Bill's home. The car made it there under its own power, but just barely--stuttering and stumbling into the shop. It shut down once on the way over, forcing Ancona to stop on the side of the road and wait until it cooled down. "It surprises me that the thing would run at all," marveled Sullivan.
Speedtek technician Billy Ranck quickly determined that the ignition system was producing plenty of spark. Although the car idled OK, more or less, the stumble was noticeably worse when the motor was placed under load when putting the trans into gear and attempting to accelerate. The electric fuel pump itself kept running, even when the engine acted up, and inspection showed that it was wired correctly, yet when the problem was most severe (total engine stall), there was no fuel at all in the carb.
There was also no fuel coming out of the still-running pump or, for that matter, going into the pump. "We unhooked the line at the pump to see if fuel was available," Sullivan explains. "There wasn't." Speedtek had seen this problem before. The minipump--rated at just 20 to 30 free-flow gph (gallons per hour)--wasn't strong enough to generate enough suction to get fuel out of the tank in the first place. "It was not able to pull vacuum under load," Sullivan says. A probable contributing factor that put things irrevocably over the edge with the car fully warmed up is that electrical resistance increases as a factor of heat; as resistance increases, voltage drops; when voltage drops, pump output decreases; when pump output decreases, the carb runs lean; when the carb runs lean, the engine runs hotter. There's that pesky negative feedback loop rearing its head once again, as we've seen in many of these rescues. The problem starts feeding on itself.
Speedtek's solution was simple and direct: Install a new fuel pump with proven real-world hot-rod reliability that has more than ample capacity to supply the engine's fuel needs under any imaginable condition. Its choice: a classic Holley "Red" pump (PN 12-801-1, $139.99 at Speedtek), which more than fits the bill. This pump's official advertised rating is 97 gph, but that's under free-flow conditions (no restriction on the outlet side). More relevant is that under real-world conditions, the pump still flows 71 gph at 13.5 volts and 4 psi of fuel pressure, all while drawing only 3 amps of current. For this mild engine, that's way more than needed, so there is a plentiful safety reserve, even if there's some voltage drop when the system gets hot.
Speedtek also noticed the distributor vacuum-advance hose was improperly connected to a manifold vacuum source (a vacuum source below the carburetor's throttle blades that's highest at idle and decreases at off-idle). This was a contributing factor to the car's occasional off-idle stumble and post-shutoff dieseling problems. Stock GM HEI distributors in particular often have weird advance curves because in a stock application, the vacuum-advance system is modulated by TCS (Transmission Controlled Spark, a primitive advance-limiting smog-control system). Designed as they are to work with TCS, a typical factory HEI distributor often has an overly generous amount of vacuum advance with insufficient total centrifugal advance. A production-style HEI should either be connected to a spark-ported vacuum source (one that is referenced to a port above the carburetor's throttle blades that's zero or near-zero at idle and increases off-idle as the throttle blades open, or, as Speedtek did in this case, blocked off entirely, with more initial advance dialed in to increase the total advance (initial plus centrifugal) from 27 degrees to 35 degrees at 3,000 rpm.
Installing the red Holley fuel pump in addition to disconnecting the vacuum advance successfully eliminated all signs of stumbling. The '37 Chevy no longer stalls out under cruise on a hot Texas day. Post-shutdown dieseling has stopped. The engine operating temperature never exceeds 170 degrees, even with the A/C on. More than satisfied, Ancona reports, "I've tried everything I could to get it hot, even running with the air conditioning on full blast. Nothing phases it." Despite no functional vacuum advance, Ancona adds that "gas mileage is still A-OK."
This problem could have been prevented in the first place. As Speedtek puts it, "Buy the correct parts to start with. You are better off buying your performance parts from a performance shop where we can give you some technical advice." Failing that, basic diagnostic tools--fuel-pressure and vacuum gauges and an electric multimeter--combined with careful physical inspection could have quickly identified the basic problem.
On the '37 Chevy, a weak fuel pump merely caused stalling. But a stouter motor run hard on top might suffer a catastrophic engine failure. A typical Holley carb needs between 4 and 8 psi of fuel pressure at all times, so the simple solution is to run a gauge and make sure your pressure remains within that range.
If spec'ing a new pump, remember that most pumps' real-world flow varies from line restrictions and voltage drop, so published free-flow ratings do not accurately portray real-world performance. Still, the numbers can be used to compare one pump with another, and applying a 100 percent safety factor (double the pump you think you need) should provide a more than adequate cushion in a real car.
A typical, normally aspirated (NA) engine needs about 0.5 pounds of fuel per hour per hp (if that sounds like the Brake Specific Fuel Consumption [BSFC] number, you're right). But for a 100-percent safety margin, figure 1 pound of fuel per hp/hr instead of 0.5. Fuel pumps are rated in gallons per hour (gph), not lbs/hr. One gallon of gas weighs 6.1 pounds. The equation with the safety factor is:
GPH = Engine hp ÷ 6.1 lbs/gallon
With a mild 350 making 1 hp/ci (350 hp), you should be safe with a 57-gph-rated pump.
GPH = 350 hp ÷ 6.1 = 57.4
If you don't know the flywheel horsepower, you'll have to jump through a few more hoops:
1. Determine your four-stroke engine's airflow in cfm at max engine speed (rpm), factoring in displacement (cid) and VE (Volumetric Efficiency); use 0.9 (90 percent) for an NA engine.
CFM = (CID × RPM × VE) / 3,456
2. Convert airflow in cfm into lbs/hr: Use “4.38” as the factor for standard atmosphere and pressure--60 degrees F at one atmosphere (14.7 psi).
Airflow (lb/hr) = CFM × 4.38
3. Find required fuel-flow. A normal WOT throttle air/fuel ratio (A/F) is about “13:1.” The inverse is the fuel/air ratio --about 0.077. Multiply the airflow in lb/hr by 0.077 to get required fuel flow in lb/hr:
Fuel-flow (lb/hr) = Airflow (lb/hr) × 0.077
4. Find GPH: Double the airflow (for your 100-percent safety margin) and divide by 6.1, as above.
GPH = [Fuel - flow (lb/hr) x 2] / (6.1 lbs/gallon)
Assume a 350ci engine with 90-percent VE at 6,000 rpm:
(1) CFM = (350 x 6,000 x 0.9) / 3,456 = 546.9
(2) Airflow (lb/hr) = 546.9 × 4.38 = 2,395.4
(3) Fuel-flow (lb/hr) = 2,395.4 × 0.077 = 184.4
(4) GPH = (184.4 x 2) / 6.1 = 60.5
By this method, the engine needs a 61-gph pump.
