1998 Honda Accord LX - P0336 error code

Have the crankshaft position sensor and wiring checked/tested back to the computer

I picked out a few useful clues and observations and it sounds like there's two things to consider. The first is a plugged pickup screen in the gas tank. That will prevent sufficient fuel from getting through. They often collapse after driving a little ways, then stretch back out when the engine is off. Connecting a fuel pressure gauge should show a severe loss of pressure when the problem occurs.

The next thing is the generator. You mentioned some voltage going down to 12 volts. The charging system voltage must remain between 13.75 and 14.75 volts with the engine running. Besides that, it must be able to deliver sufficient current to run all the electrical systems plus recharge the battery. When one of the six internal diodes is defective, you will lose exactly two thirds of the output capacity. The common 90 amps generator will only be able to supply about 30 amps which is not sufficient to run the electric fuel pump, fuel injection, ignition, and other computers and circuits. When the capacity can meet the demand, the system voltage will still be high enough, and sometimes a little too high, (above 14.75 volts). That is a clue but the definitive test is to have a load test performed on the generator.

Also, when one diode is bad, "ripple" will be very high. I can explain that further if you want me to. Basically it means the system voltage will be varying a lot and very rapidly. You might also hear a louder than normal whine on an AM radio station that changes in pitch with changes in engine speed. That ripple wreaks havoc on computers because they are not tolerant of voltage fluctuations or low system voltage. Bad diodes are often responsible for computers doing weird things that defy most troubleshooting procedures.

If I understand correctly, you measured the voltage at the battery and it dropped to 12.7 volts AFTER the engine died. If I have that right, that is normal and correct. 12.6 volts indicates a fully charged battery. Dropping to 14.2 volts when many accessories are turned on may be a clue. While that is within the acceptable range, it suggests the generator isn't keeping up with demand, however, all generators lose efficiency at lower speeds. If you find the voltage comes back up when you increase engine speed a little, that is normal.

That still doesn't address the load test though. That must be done with a professional load tester. Most of them also display ripple, not as a voltage, but as a relative bar graph. High ripple is indicative of a bad diode.

Dandy. I hope they tested it for maximum output and ripple. I never took a generator in for testing so I don't know what they look for. I've been told some places just check to see if it can produce some output. That's not nearly good enough. All of my testing is done with a professional load tester with the generator on the car. It must be able to deliver its rated current when engine speed is raised to 2000 rpm. I've never seen a bench tester that can run a generator that fast.

If a professional load test shows low ripple, and maximum output current can be produced, I'd start by monitoring fuel pressure. I like to hook the pressure gauge under a wiper arm so I can watch it while I'm driving. You'll get used to seeing how the pressure changes when you're on and off the gas, then you'll notice it if it acts differently when the problem occurs.

Sounds like a VAT 20/40

Volt / amp tester model 20/40.

A coolant leak shouldn't be related. The test you need to have done is a charging system load test. That must be done with a professional tester to rule out the generator as the cause of the problem. While yours is developing enough voltage, which is normal with one defective internal diode, its the current that is in question. One bad diode will cause an inability to keep up with the current demands of the electrical system under some conditions, and ripple will be high too which confuses computers and makes them do weird things, especially Engine Computers.

Here's a better description of ripple if you care to know more about it. It's the last paragraph on the page:

http://randysrepairshop.net/charging-systems.html

You have some problems that should not have "passed". First of all, look at the red arrow. Your generator should be capable of developing 80 amps. With one defective diode out of the six, you will lose two thirds of its capacity leaving you with 25 - 30 amps. The electric fuel pump takes around 8 amps. Add in the ignition and fuel injection systems, all the other computers, radio, radiator and heater fans, and a few lights and the generator will not be able to keep up. This printout shows less than 8 amps output which is definitely WAY too low and would never pass. The load test does just that, it loads the system down to mimic the electrical systems on your car to see how the charging system responds. Since the generator, (which is like a water pump for electricity), can't keep up, the pressure, (voltage), drops. That is shown by the 14.32 volts unloaded, then dropping to 14.14 volts when it has to do some work. That voltage is used to create a magnetic field inside the generator and since it is lower, the magnetic field is weaker, and the output current goes down as shown by the lower current under load, 7.8 amps.

I suspect it's more likely the mechanic placed the inductive pickup probe on the wrong wire when he made this test. One bad diode would result in a load test of at least 25 amps. Two bad diodes might result in an maximum output of around 8 amps, but then ripple would be extremely high, in the order of perhaps 3 - 5 volts. This is showing 61 millivolts, (.061 volts), which is normal and acceptable.

7.8 amps is absolutely not acceptable for any car. The rest of the numbers suggest the test was done improperly. The mechanic should have seen that number and questioned the results and his equipment setup.

The next number in question is the drain. The only accurate way to measure that is by disconnecting one battery cable and inserting an ammeter but that has its own procedure to prevent blowing fuses inside the meter. It's a pretty involved description and it can take up to 20 minutes for some computers to go to "sleep mode" before the measurement can be taken.

If an inductive pickup probe was used, (those just snap around a battery cable like a clothespin), they are not nearly accurate enough at such small current levels. Unless the manufacturer says differently in their service manual, the industry standard is a maximum of 35 milliamps, (.035 amps), after any computers have timed out and gone to sleep mode. At that rate of drain, Chrysler and some other manufacturers state a good battery will crank the engine fast enough to start after sitting for three weeks. Cadillac allows 50 milliamps because of all their silly computers that have memory circuits constantly drawing current from the battery. Yours is showing 190 milliamps which is very high. I suspect an inductive pickup probe was used, some computers hadn't timed out yet, or the probe was on the wrong wire. It appears this was done with some type of automatic tester that doesn't require a skilled operator who knows how to interpret the results or understands when proper procedures aren't being followed. If your car really had a 190 milliamp drain, the battery would be too low to start the engine if it was left sitting for about three or four days. If you are NOT having that problem, suspect operator error.

All these things considered, I'm even more inclined to go with my original thought of a plugged pickup screen in the gas tank. According to your description, you were at times unable to accelerate, ... But the engine remained running. That proves the ignition system was working, but now you want to jump around blindly and start troubleshooting the ignition system. The logic escapes me.

Nothing you said up to now points to an ignition system problem, but everything can be attributed to a fuel supply problem. You might also look at the mass air flow sensor in the fresh air tube. It is responsible for measuring the weight of the air going into the engine, then the Engine Computer commands the correct amount of fuel to go with it. As long as the signal voltage from that sensor is within acceptable limits, it will not set a diagnostic fault code, but it could send the WRONG signal within those limits. That will result in the wrong amount of fuel entering the engine and the symptoms you described. Be sure there are no cracks in that tube and any hose clamps are tight. If any air sneaks in that doesn't go through that sensor, you won't get enough gas to go with it.

Sorry; forgot my sad arrows.