I have a 1986 Accord LX-I with manual transmission. When the car is cold, it starts fine, and it always runs fine. It likes to idle at 2000rpm sometimes, but there's no hesitation or weirdness in its operation. When the car is warmed up, it will crank and crank and crank for quite some time before it starts...kinda like running out a flood on a car with a carburetor. A strange one: if you push-start the car when it's warm, it fires right up. But I totally trashed the left inner CV--tearing a hole in the oil pan in the process--when I did that once, so no more of that. So I thought, if I pull the fuel pump fuse and crank the engine, maybe it will start. Pull the fuse, the car runs for about three seconds...which is all you'd expect to get. Stick the fuse back in, car sometimes starts and sometimes does not. But you do the fuse in-out thing, and eventually the car starts. As an experiment, I bought a 30-amp automotive switch from RadioShack and wired it into the fuel pump fuse circuit. This works pretty well...if you turn the fuel pump on, crank the car for a few seconds, then turn the fuel pump off, it fires. Immediately start pumping the gas and turning the pump on and off, and within 20 seconds the car will be running. And like I said, once it is running it runs great. I think my coolant-temperature sensor (aka "thermowax sensor") is screwed up, always reporting that the engine is cold. This device probably only gets read at startup--if the engine's cold, dump more gas in and it will fire right up; if you dump more gas in when it's hot the engine floods. I'm thinkin' the mixture-enrichment cycle only lasts about ten to fifteen seconds because once I'm out of that period, the engine runs stably in a very short period of time. I've already changed the fuel injector seals, and they needed it. I changed the thermostat because the car liked to overheat, and that also seemed to make it start a little easier. Would it be good to start hitting junkyards looking for a new thermowax sensor? I know the dealer price on one will be astronomical. --jm
google this group for the #1 most common failure in older hondas, especially in summer. 30 seconds research would have saved you a bunch of time and money.
I looked at the relay issue before I wrote the OP and it really doesn't seem applicable. 1) this happens all year round--not just during the summer. 2) I just went outside and started the car. I got three clicks. When the car is dead cold, it always starts within a couple of seconds of turning the key to the "start" position--normally one second. (After I put the new injector seals in, it doesn't even need the two-second pause for the fuel pump to pressurize the rail--stick the key in and turn it straight to "start," it fires right up.) 3) When the engine is hot, if you push-start the car (which doesn't activate the starter, hence doesn't tell the PGM-FI "you're going through a startup cycle") it always starts. 4) When you start the car hot and turn off the fuel pump during cranking, it fires right up--naturally, it won't STAY running because no gas is being pumped, but it does start running. And when you sit there and feed it gas in little increments by turning the fuel pump on and off and pumping the gas so the engine will continue to turn, eventually it gets through that time when the PGM-FI wants to overdose the engine on gas and it runs. 5) While this cranking and pumping thing is going on, there's a very strong smell of gas around the car, like a car with a carb gets when it's flooded. If it was a main relay problem the car wouldn't run at all, and the car runs fine. It's just a livin' bitch to start it when it's hot. --jm
main relay. but it's worse when hot. the seals have /nothing/ do do with ability to start. they just hold the gas in the rail and stop the engine sucking air, depending on which end of the injector you're looking at. er, it wouldn't run at all if the injector system couldn't figure out when to start injecting... ok, that's more diagnostic. do two things - check into flooded start conditions and remove plugs to see if you can find the wet one. that'll give you the leaking injector. unfortunately, the worst thing about the relay is that it's intermittent. i strongly recommend you repair or replace it as a precaution. the chances of this vehicle at this age /not/ being affected is slim to zero. then you have a clear run at your other symptoms without false positives.
Indicating the car is not CRANKING. could be a starter problem. (solenoid contacts?? they can be replaced.) Could be a worn ignition switch;not energizing the starter. PGM-FI (ECU) does not monitor the starter. Not true. MR is the most problem when the car is hot,but once the car starts,it stays running,and always starts when cool. Does the car not crank,or cranks but doesn't start? (when hot)
Not true. Yes,it's SO easy to just resolder the relay and ELIMINATE it as a trouble source,and doesn't cost anything except a little time. Then he could look at starter problems,or whatever.
The car cranks fine. It runs fine once engine starts. It starts fine when cold. It needs to have the gas pedal pumped and pumped and pumped when starting hot. Once the gas pedal is pumped enough during a hot start to get it to run, the car runs fine. It's not an intermittent problem; it happens on every hot start. It gives the indication of flooding only during hot starts; once the car's running, it runs like nothing's wrong with it. It doesn't flood on cold starts. And this is the really weird part of the whole problem: once the car starts, it runs exactly as it should. I'm thinking this is the logic the car uses when a start command (turning the key to the "start" position) is issued: on (startup): if (start_switch) = "on" AND (coolant_temp) < 120 deg F then inject additional fuel quantity for 10 seconds and run starter end if or if (start_switch) = "on" AND (coolant_temp) >= 120 deg F then inject standard fuel quantity and run starter end if But for some reason, this particular specimen can't figure out that it's already warm, so it dumps more fuel into the engine than the engine can use, and floods it out. --jm
shouldn't make any difference on a fuel-injected car;there's no "accelerator pump" to shoot more gas into the airstream. Or the main relay has cooled enough to work again. I don't believe that the change in fuel would be that significant,to flood it out.(At temps other than winter conditions) It's sounding more and more like the main relay. Your other post says you have spark. You say if you pushstart the car,it starts and runs. The ECU would not know that,the starting mixture would be the same as if you cranked it with the starter. If it -were- flooded,pushstarting it would be no different. Of course,removing and resoldering the relay is fairly easy and doesn't cost anything,and ELIMINATES it as a problem. You seem to be resisting this.
dude, you don't "pump and pump" to hot start a potentially flooded engine. you press the pedal to the floor and keep it there while you crank. that tells the ecu to /not/ inject to clear a recognized no-start condition. that is an intermittent problem! exactly like a main relay. main relay. main relay. it's a little more complicated than that, but if you really want to know, look up the open source injector code for "megasquirt". "google ees yor frien'." dude, fix the main relay, then come back and tell us if if your problems persist.
--------------------------------- It's the main relay. Chilling the interior, or vibrating the relay will make it work properly. That's why push starting it works. Smacking the dash on the left side with the key in position II will force it to prime the rail too, or slamming the door. . Just fix the relay. . . . Life is too short. 'Curly'
Doesn't the crank position sensor (which would trigger during a push start, since the distributor is turning) tell the PGM-FI when to inject? --jm
I got up under the dash and resoldered the relay late yesterday. (At this point I am thinking that the placement of this component is payback for the Chester Nimitz overseas-postage stamp...) Today I took the car to work. First start was at 4:30am. It started normally. Second start was at 2:15pm. (Yes, I work weird hours.) It started normally. So far it's working just like it did before I resoldered the relay. Third start, because I had to go home, pick up a bill and mail it, was at 2:30. Back to the ol' twenty-second pump-and-flip routine complete with gas smell. When I got home after my trip to the post office, I opened the hood and plugged/unplugged the coolant temperature sensor about ten times on the off chance that maybe the connection was just crudded up. It happens, right? I then got back in the car, turned the key...normal start. I ran it around some more, came back and shut down. I waited a minute and turned the key...normal start. I didn't want to resolder the relay because what this car is, or at least was, doing isn't like Tegger's description of what the relay does when it screws up. For instance, it always gave the three clicks it's supposed to. But it's resoldered. --jm
I've tried that. It doesn't work. Press the pedal to the floor and keep it there, the thing revs for three or four seconds then dies. And you smell gas. Pump the shit out of it for 20 seconds and you're running. --jm
I resoldered the relay and it didn't solve the problem. Cleaning the contacts on the coolant temperature sensor, by plugging and unplugging it a handful of times, solved the problem. It starts right up now. --jm
of course - you have to let the throttle back to a more normal position the moment it fires!!! there no injection programming to support this action.
let's get this back into context: you originally wrote: "3) When the engine is hot, if you push-start the car (which doesn't activate the starter, hence doesn't tell the PGM-FI "you're going through a startup cycle") it always starts." check out tegger's description of main relay operation - that'll explain everything for you. bottom line, if the main relay is operational, it'll start regardless. if it's got intermittent open circuit, the "start" position will not ground the injectors, hence the problem and the difference between the two modes. the ecu triggers the injection cycle when the distributor tells it to - regardless of how the crank gets to be rotated.
