1996 Honda Accord LTFT and TPS

The Engine Computer was programmed to provide the correct amount of fuel for any combination of throttle position, throttle direction and rate of change, barometric pressure, intake manifold vacuum, intake air volume, intake air temperature, coolant temperature, and other similar factors. Those are the default values it will always start out with after the battery has been disconnected or run dead.

Once the coolant temperature reaches a certain point and the oxygen sensors reach 600 degrees F, the computer modifies the numbers in those "look-up tables" based on the readings from the oxygen sensors. Those are the "short-term fuel trim" (STFT) numbers and they will be constantly changing while you drive. They apply to the current conditions.

When the computer sees that it is always making the same corrections for a given set of conditions, it will adjust the "long-term fuel trim", (LTFT) numbers to reflect those changes. From then on, it will use those numbers as the starting point instead of what was programmed in at the factory. Those numbers will change very slowly over a long period of time.

If you see the short-term fuel trim numbers are close to 0.00%, that tells you there is little need to adjust the mixture beyond what is expected to be needed. If you find the long-term numbers close to 0.00%, the desired mixture is very close to what the engineers determined should be needed for best performance, fuel mileage, and emissions.

If you find a number is high positive, the computer is adding fuel beyond what was programmed in to be needed. A negative number means fuel is being subtracted from the calculation.

Nope. Some cars and some scanners read throttle position as a percentage, but it's much more common to read it as a voltage. On 99 percent of car brands and models the TPS, along with most other sensors, is fed with 5.0 volts. The ground wire would have 0.0 volts but since it goes to ground through monitoring circuitry inside the computer, you'll actually find close to 0.2 volts on it. For purposes of this story, we'll consider that 0.0 volts. If this were a volume control on a radio, you could turn it to vary the signal voltage from 0.0 to 5.0 volts. However, on a throttle position sensor, there are mechanical stops, either inside the sensor or on the throttle body, that limit its travel to roughly 0.5 volts to 4.5 volts. (Your 9.4% would equate to 0.5 volts).

The only way the TPS signal voltage could go all the way down to 0.0 volts is if there was a break in the 5.0 volt feed wire or the connection inside the sensor. The only way it could go to 5.0 volts is if there was a break in the ground wire. Either condition will trigger a diagnostic fault code. It's by those voltages outside the acceptable range of 0.5 to 4.5 volts that the computer identifies when there's a defect.

There's one other condition that will set a fault code. That's a break in the signal wire or that connection inside the sensor. That can allow the signal voltage seen by the computer to "float" to some random value due to being interconnected to all the other circuitry inside the computer. The computer could use those erroneous readings to try to run on, and that would cause all kinds of weird and intermittent symptoms. To prevent that, they always use a "pull-up" resistor or "pull-down" resistor. A pull-up resistor is more common. It is so big electrically that it has no effect on the circuit or its operation, ... Until there's a break in the signal wire. Then, since it is connected to the 5.0 volt internal power supply, it puts 5.0 volts on the signal wire which forces the computer to see the defective condition and set a fault code. If a pull-down resistor was used, it would apply 0.0 volts to the signal wire which would also trigger a fault code.

Every car and every sensor is different. The acceptable range of 0.5 to 4.5 volts is used for training, but you might find one car has a range from 0.42 to 4.45 volts, and an identical model car or a different sensor produces a different range. The computer has to learn when those minimum voltages change, as in when a sensor is replaced. In the case of a TPS, when the computer sees what it knows to be the lowest TPS voltage, it knows your foot is off the accelerator pedal and it has to be in control of idle speed. If you install a new sensor with a minimum voltage even slightly higher than the old one, the engine could stall from the idle speed being too low until the new "minimum throttle" is learned. All manufacturers have a different strategy to cause that relearn to take place. On some you'll never notice a problem. Others simply require certain conditions to be met during a test-drive.

No. That's perfect. The only way it can go to 0 percent is if there's a defect in the sensor or the wiring.

There's no indication of a problem. The Engine Computer appears to be responding appropriately to readings from the oxygen sensor. It's adjusting the fuel needs relative to what was programmed in at the factory.

What IS normal is for adjustments to be constantly taking place. I don't know what it takes for the various monitors to run, but it usually requires meeting a specific set of conditions while driving. Those could include things like a brief burst at wide-open-throttle, a certain number of miles above a specific speed or without touching the brake pedal, or a brief drive with the fuel tank almost full or empty.

If you get a fault code for "running rich too long", look for things like fuel pressure that's too high or a leaking injector. If there is no fault code, it's not running rich. If you're looking at the fuel trim numbers and interpreting that as running rich, it's doing what it's supposed to do. You have a problem when all the numbers stay on "0.0" and won't adjust.

There's a lot of factors that determine how fast it switches but in general it will be once or twice per second. Don't use any other car for comparison because they could be totally different and neither one has a problem.

Also be aware that '96 and newer cars will have a second oxygen sensor after the catalytic converter to monitor its efficiency. That one has nothing to do with fuel trim numbers or how the engine runs. The front sensor will switch rapidly between rich and lean. During the lean times, oxygen is stored in the catalyst. During the rich times, the excess fuel mixes with that stored oxygen and is burned. When the catalytic converter is working properly, the gases coming out of it will be slightly rich for a long time, then slightly lean for a long time, as in a minute or more. When the catalyst stops working, less and less change takes place in the composition of the gases as they pass through it. When there's no change at all, the "downstream" sensor will switch between rich and lean at the same rate as the "upstream" sensor. The Engine Computer watches the rate of change of those two sensors. When the downstream sensor switches very slowly, it's because the catalytic converter is doing its job. When it loses its efficiency and becomes less and less effective, the switching rate of the downstream sensor speeds up, and when it reaches a specific switching rate, that's when the computer sets a fault code for "catalytic converter efficiency".

Yup, but I have to qualify that. Typically they take the barometric pressure reading in the instant between when you turn on the ignition switch and the time you start cranking the engine. That's just a fraction of a second, then that number is stored in the computer until the next time you turn on the ignition switch. I can't say if all of them or any of them continue to update after that. They're designed to expect to be reading intake manifold vacuum once the engine is running, so if you connect a hand-operated vacuum pump to the sensor, it will show a vacuum reading on a scanner, not a change in barometric pressure.

To take the barometric pressure reading, the sensor doesn't know or care if it's connected to the intake manifold or not. The pressure will be the same either way. What CAN happen is if you start the engine, you may get a fault code for "no change in MAP between start and run". In that case the computer knows the sensor is working electrically but it knows there has to be intake manifold vacuum because the engine is running, but it didn't see the change in the vacuum reading. That is a pneumatic-related fault code related to that sensor, as opposed to an electrically-related fault code.

I wouldn't worry about it. You're looking at something we never pay any attention to. Even if the barometric pressure reading is incorrect, it's just the base line or starting point for fuel metering calculations. The actual barometric pressure is going to constantly change a lot just during normal driving as you go up and down hills. Imagine how much it changes when driving through the Rocky Mountains. The Engine Computer can accommodate those changes because the final word comes from the oxygen sensors. It's when you get to a real high altitude that changes in barometric pressure put the MAP sensor out of range and it can't measure intake manifold vacuum accurately. The fix for that is to stop the engine and restart it. Doing that updates the barometric pressure reading that is put in memory, and corrects the range of the MAP sensor's readings.

The number you should be interested in is the vacuum reading when the engine is running. That should be close to what you would read with a mechanical vacuum gauge.

The fact that the signal voltage changes when you reconnect the plug proves the circuit is working. The 4.6 volts, which in this case is close enough to 5.0 volts, is put there by a "pull-up" resistor inside the computer. It is so large in value that it has no effect on the circuit, ... Until you create that open circuit by unplugging the connector. Then, that resistor puts the 5.0 volts there to force a diagnostic fault code to be set. For this purpose, 4.6 volts is close enough.

When the service manual gives you a voltage to expect, with no tolerances, you're looking for something reasonably close, and 2.58 volts is close to 3.0 volts. What you don't want to see is close to 0.0 or 5.0 volts. I would have expected to see around 4.0 to 4.2 volts with the engine not running, but if the book says differently and that's what you found, I can't argue with that. What they're calling for would be typical of a turbocharged engine where the intake manifold vacuum can go into a pressurized state. Normally that MAP sensor is different and won't work properly in a non-turbo engine.