Fuel Pump Current

To get these waveforms, you will need a low amp probe. You can find several different ones available from AES and other suppliers. Somewhat expensive, in the $200-$400 range.

Briefly, I would like to tell you how I approach a no start situation, I begin by checking the fuel gauge to see if it has fuel in the tank. Even though it shows fuel, I don't assume its telling me the truth (Gauge's often lie a lot) If it shows an adequate amount, I move on to the next step in my quick diagnosis, I listen for a pump prime signal, (assuming its fuel injected) if it exists, and the way it sounds. If it is very loud, the tank may be empty. At this point, I would want to put a few gallons in the tank, to make sure that this isn't the case, If it is, I just saved a lot of time investigating further. Now I will probably be looking at a fuel gauge problem. If the pump prime signal exists and seems to sound normal, muffled buzz for about 2 seconds, I crank the engine for a few seconds. After Cranking, I will go to the tailpipe to see if a raw fuel smell exists, if so, my next step is a quick spark test, first the coil wire, if one exists and is easy to get at, or right to a easy plug wire with spark tester. If ignition seems ok, it may be flooded, some timing issue, or a mechanical problem. If no fuel seems to exist at the tailpipe, its surely a fuel supply problem, be it injector trouble, or pump trouble, now is the time to do some testing. In most cases, its easy to hook up a noid light and see if injector signals are there and, if so, now I will do a quick pump current test that can quickly point me in the right direction.

Where to get the signal? Probably the easiest place is the fuel pump relay, most often found under the hood on the firewall or in the underhood fuse-relay center, or the pump prime wire on those equipped (most often near the relay). I just bridge the socket from power terminal to pump terminal, (this verifys power to the relay) or directly from the battery to the prime wire with a heavy circuit breaker (10 amp) equipped jumper wire to hookup the amp probe. To get the signal with the engine running, extend the circuits with jumpers between the relay and socket (with this, you could see the draw including the points in the relay and any intermittent relay operation). If the breaker pops open, you know right away that there is a problem on the main power circuit to the pump, be it a shorted pump or wiring to the pump. If the pump is operating, you should hear fuel passing the regulator through the return line (hissing sound). Hooking up your vacuum pump to the pressure regulator, will allow you to check it for operation while viewing the draw of the pump. Now you can look at the signal. Here is the first example.

Fuel pump signal, 97 Honda Passport 3.2V6 56k miles. Signal acquired at pump relay socket, pump power and pump circuit bridged with jumper wire, engine off. Draw high.

If you get the signal at the relay, make sure nothing else is loading the circuit except the pump, there were four o2 heaters on this one, pulling the fuse to the o2's dropped the reading by almost 4.5 amps. Below is the diagram, check them first.

Lets take a close look at this waveform and see what it can tell us in the image below.

This pump is weak, it has a good pattern, but draw is low, Avg. 5A, speed is high, specs for this pump are 48-55psi. Pressure was about 42psi non regulated. Pump slightly noisy. Although this thing ran ok, this pump is on its way out at only 56k miles. Customer opted for new pump.

If you know the amount of segments on the armature's commutator, you can find the pump speed in rpm with some simple math. No one know's the amount of them on every pump ever produced, so how do you find out? If you can open up the vertical scale on the scope and the picture is clean enough, you might be able to see some pattern in the waveform that repeats itself. If so, now you can count the peaks in between the areas that repeat themselves and you have the segment count. Look at the image above, you can see the pattern (A-B-C). Once you know the segment count, look at the time over the count in the waveform, use your cursors on the scope if you have them, or just look at the divisions on your time scale. The math is easy, there are 1000ms in a second, 60 seconds in a minute, 60X1000= 60000ms in a minute. So just divide 60000 by the time it takes for one complete revolution, in this case, 8.62ms, and you have the speed of the pump in rpm. Normal draw varies with different pumps, a rule of thumb by pressure ratings, 2-4A at 10-15psi, 4-7A at 30-50psi, 8-12A at 50-65psi, there is no magic number, just watch speed and draw. Once you have viewed many applications, you will get a feel for whats normal.

Knowing the pumps speed along with its current draw will tell a lot about the pumps condition, and I will get into that later, I want to dissect the waveform itself, so lets focus on it a bit. Its interesting with its see-saw pattern, and if you give it thought, its easy to understand how its drawn. The current draw fluctuates as the brushes pass over the gaps between the commutator segments as the armature turns.  Each high peak is when the brush is directly over the center of the segment, and would be the point of highest load, when the brush moves away from center, current draw falls off slightly as it moves over the gap until the next segment begins its path beneath it. This is whats producing the see-saw effect. Each high peak is the total draw across the winding for each segment. What would happen to the waveform if one of these circuits goes open? It would result in a dropout in the waveform that would be quite obvious, and would show up as a loss of one or more segments.

Now, why is it important to know the pump speed in rpm? This is a good way of knowing what the pump is doing, it may have a fine pattern, but it may have other problems that the pattern will tell you. If you see low current amount with high rpm, the pump isn't producing pressure, this could be the result of an empty tank, worn out pump cell, plugged tank sock, cavitation, bad hose from pump to sender, If you see low current amount with low rpm, the pump commutator is worn out, coated, pitted, worn out brushes, weak brush springs. If you see high current amount with low rpm, the pump is producing pressure, but its probably working against some restriction, most often a very overlooked fuel filter, the pump may or may not be howling, with this indication, it would be best to get the pressure gauge hooked up for pressure and flow testing. Do a quick pressure and flow test now, pressure may be close, but flow will most likely be down.

To verify the filter it must be removed, then flow check the pump and recheck pump speed and draw. If its better now, its most likely the filter, If no change, its time to go into the tank. There isn't much in the tank between the pump, and the line to the inlet nipple on the filter, (depends on where the filter is) but there is always a chance the line could be plugged between them. This is rare, but not impossible, you could unhook the line at the sender, (maybe even without dropping tank) and check for flow once again, this may save wasting a new pump. If a new filter will not correct whatever the restriction is, you now know it is ahead of the filter. This will be a result of a restriction in the path between the filter outlet, and the return line at the tank, it could be plugged, dented, or kinked lines, plugged regulator, plugged injector rail, you will need to unhook each point along the path to find the source of restriction beginning at the point closest to the filter outlet.

89 Ford Tempo 65k miles, good pattern. How many segments? Hard to tell in this image.

If the vertical scale is opened up, now a repeater is seen. Segments are easily counted, how many? Looks like eight, now count the divisions starting with one that lines up with the top or bottom peak and count over eight peaks, and divide 60k by the number and you now have the rpm.

I use the software with my fluke to help look at the pattern, you may have this luxury. Eight segments, pump running at 4477 rpm, amps at about 5.5 avg. The regulator is at rest in these images, with vacuum applied, draw would increase slightly, rpm would drop.

More Fuel Pump Waveforms

See a Case study here