86 Accord LXI Starting Issue
 

Tried to create a new thread - wouldn't work - so I'm going to try it via a replay to a thread close in content...

I'm having a heck of a time trying to pin down a problem and wondered if anyone might have an idea I haven't tried.

Car: '86 Accord LXI, FI engine (20A4), 5-speed manual

Symptom: Car won't fire-up (start) using starter. Car will start with a short 10 foot push and a pop of the clutch. Runs fine for mileage once started ('flat spot in 3K RPM range only anomoly)

Replaced or verified working:

• Temp sensor to ECU replaced (it was bad)
• New distributor (complete, all components)
• New coil
• New fuel injectors
• New fuel filter
• New plugs
• Wires checked
• Ignition switch bypassed (12V applied directly to condensor high side - verified dash engine lights lit, even tried to start it this way in case problem with ignition switch in "start" position)
• Main fuel relay pulled and verified on bench. Backup fuel relay tried. Loosened nut on fuel filter, turned engine on, fuel squirted out, retightened nut).

Original problem was traditional non-start on a hot day after sitting in the sun, would start after cooling off. That led to the fuel relay and fuel filter items above.

One day on the way home, turned a corner and engine died. Wouldn't re-start. That led to the coil and distributor replacement. No change after coil replacement, originally started after distributor replacement. Drove over weekend, let sit in sun, all seemed fine again.

Next day, went to leave work - no start. First time tried push start - started and drove fine.

Next up was verifying the ignition switch in the "start" position was supplying power to the coil. Also tried the direct connection batter to coil via the hot side condensor lead. No change.

Next up was the fuel injectors. No change.

I'm out of ideas... Starts easily with a push (I can push it 10 feet, jump in, pop clutch, and go). Today leaving church I parked with rear wheels on a foot high mound of dirt. Just rolling off the mound was enough to start it with the clutch.

Anyone have any thoughts on what could cause the symptoms - fix this?????

Made a new thread for you.

When you turn the key to start the car, does the starter spin the engine, or does nothing happen?

Thanks for the thread :cool:


Knew I forgot a "checked this" piece of information :o Yes, starter spins. Checked via the oil fill on the valve cover that the starter is engaging and the engine is turning over as well (camshaft turns, valves open/close).

Since the engine runs after popping the clutch, you can rule out timing and compression. Also the starter is likely off the list, since it spins the engine at the normal rate.

You need to determine if you are missing spark or fuel when the car will not start.

If you spray starting fluid into the intake hose, can you get the engine to fire up and run? Do you hear the fuel pump prime when you turn the key to the II position, but do not try to start the vehicle?

Plug one of the spark plug wire into a spark tester (or old spark plug with the threads grounded) and see if you get spark when the engine is cranked normally.

:)
If I turn the key on and off multiple times in a row (triggering the 2 second main fuel relay activation), I can also smell gas (btw no leaks - all connections, injectors, fuel filter, etc. dry externally). Don't have any canned starter fluid. Tried with some carb cleaner, 1/2 second shot via the PVC port into the throttle body.

Yes, you can hear the main relay click on, then after approximately 2 seconds turn off. Can also hear the pump. Question for future reference - in reviewing the electrical system diagrams, can't see anything that would cause fuel to not flow when starting via the key, but would lead to fuel flow as a result of pushing the car for a few feet.

Ok, tried 2 different old plugs (always keep last set during tune-ups for emergency use). Verified electrode grounded via resistance check, tried both #2 and #4 wire. Turned to start position for 2 seconds in a pitch black garage - no spark.

Would have never thought based on all the other testing and that it runs with a push... Just goes to serve as a reminder - regardless of other indicators, always do the simple tests regardless.

Now the question becomes - what can cause a lack of spark while attempting to crank with the starter, yet full spark and immediate engine start from a couple of feet rolling start? Keep in mind two things... 1) dash engine light stays illuminated while cranking and 2) a direct connection from the positive terminal of the battery to the positive lead of the condensor doesn't result in starting when cranking via the starter. (also note that correct lead of the condensor verified by engine light and oil pressure light turned on entire time connection was made and fuel relay turning on upon initial connection.)

Oh - and thanks...

I'd recommend not to jump directly from the battery for future testing. You can possibly do some damage to the wiring or electrical components if you jump to the wrong connection.

Your problem is not spark when turning the key to the start position. Need to do some testing to identify the problem.

I think your next step is to test for voltage on the blk/yel to the coil while the coil is plugged in. You will have to back-probe the connector. See what happens when the key is in the II position (like when you pop the clutch) and also when you turn the key to the start position.

I'm looking at a shop manual wiring diagram and the radio condenser appears to control the spark timing? Do you know how this operates? I am more familiar with 90+ accords that had the ECU controlling spark timing.

I had a fuse in line when I did the connection to the battery - and had verified the proper connection with an ohm meter between the coil and the chosen condensor lead before doing so ;)

The blk/yel wire is in parallel with the condensor (the black/yel wire from the ignition switch splits coming out of the wiring harness at the coil feeding both).

The condensor is just a capacitor for eliminating RFI (that annoying static one hears on the radio when if it isn't connected). The spark timing is all controlled in the distributor in 86's. Both the ignitor and cam position sensor are in the distributor (they moved the ignitor external to the top of the fender well on VTEC engines I think). The ECU on an 86 only controls the fuel side of the equation (throttle body FI).

Here's the ignition side from my shop manual:
https://www.hondaaccordforum.com/for...ine=1468217963

The ignition switch being bad would make sense giving that it'll start with a push (alternator then supplies power to the coil, even though the ignition switch doesn't) - but that's why I did the direct connection to the battery. Based on that test, I'm at a loss... Note also that the alternator light (charge warning light on the diagram) also comes on while attempting to crank (bright when in the on position, dims when turning starter as it pulls juice), which is another indicator of voltage reaching the IG terminal of the switch. A real puzzler...

Tomorrow after work I'll see about carefully piercing the insulation of the wire going into the coil connector with a straight pin to see if I can monitor the voltage while trying to crank.

You could check whether the ignition switch opens the IG circuit when the key is turned to START. I remember 1 or 2 threads where that was the culprit (although they were newer Accords).

Couple ways to test:
- Disconnect the switch & use a continuity tester.
- Unplug the little wire (S on the starter) so the starter doesn't actually run. Then measure voltage at the bk/y wire at the distributor in both positions. Wiggle the key around to see if the connection is flaky while in the START position.

It's curious why the charge warning light would light up, but maybe there's a way to backfeed that lamp through the voltage regulator while the alternator is spinning? So I wouldn't trust that lamp to prove the IG circuit is live while cranking.

Jim - your comments are the reason why I originally jumpered straight over from the battery to the condensor lead - to insure there was voltage at the coil in the start position. There is a 'feedback' (or back feed if you prefer) path for the alternator to supply the voltage, which is why I tried the switch bypass I mentioned in post 1. Here's the path highlighted:
https://www.hondaaccordforum.com/for...ine=1468288258


Here's a pic of the connection when I did the jumper - basically bypassing the ignition switch while trying to start:
https://www.hondaaccordforum.com/for...ine=1468288258

-----------------------------------------------------------

PAHonda - took a number of measurements. Pic below shows the test points I used, with TP1 being the condensor lead and TP2 being the connector at the distributor (it unplugged, meter into connector contact for the coil):
https://www.hondaaccordforum.com/for...ine=1468288258

With the engine running, I disconnected the condensor and measured the voltage at the condensor lead (TP1) with the engine running - was 12.1 volts.

With the engine off and the key in the "On" position, measured the same spot (TP1) and got the same 12.1 volts.

With the engine off and the key in the "Start" position and starter spinning, measured the same spot (TP1) and got 10.1 volts.

Disconnected the plug on the coil and inserted my meter probe into the connector end for the Bl/Yel wire (TP2) and repeated the measurements for the "On" position and "Start" position (with starter spinning) - exact same measurements I had at the condensor lead.

The measurements basically give the same "test results" as when I jumpered across from the battery to the condensor lead - voltage at the coil, no fire.

Still baffled... (me, not the car).

Seems like there has to be something else (another connection - maybe to the ignitor in the distributor) that is there when doing a rolling start and that isn't there when doing a starter crank... but I'll be darned if I can find any hint of it in my shop manual.

There are 4 connections to the ignitor in the distributor. Three of them are shown in the ignition wiring diagram I've been using. The 4th connection isn't shown - is a green wire - and I suspect that it's the feedback to the ECU unit, but I can't find it shown anywhere.

Wouldn't think the 10.1V when cranking would be the issue, especially in light of the attempted start with the jump from the battery to the condensor lead (using 16 AWG wire with a 10A fuse inline btw).

:shrug:

*bangs head on engine header*

What is the voltage across your battery terminals when you try to start the engine? You may be jumping the 10.1 volts to the igniter.

You may want to charge the battery and see if your voltage measurement changes.

You can also try charging the battery to see if it gets above 12.4 volts. Turn on the headlights and see if it drops voltage slowly.

Here is a good video on doing some electrical diagnostics.

OK I misunderstood exactly how you checked early on...
If the problem still exists with that jumper from the battery, then it cannot be caused by a flaky ignition switch.

Jim - no problem, do appreciate the thought and effort both :)

PAHonda - I've actually done the tests with my after the battery was at 100% charge (per my battery charger). My charger is also one that has a "start" function that allows you to start the car using only the charger (it'll supply cranking current for a _short_ period of time).

Your query did get me to thinking, so I checked what the voltage range should be in the shop manual. Here's what Honda's shop manual has to say (my apologies for the picture sizes - didn't realize how big they were until after I put them in the reply):

For testing the battery they give the following under 'full' load (note that my original measurements were done after extended attempts to turn the engine over trying various things):
https://www.hondaaccordforum.com/for...ine=1468370065

There's also a check for the starter and conditions while cranking the engine:
https://www.hondaaccordforum.com/for...ine=1468370065

Note that the 10.1V I'm reading are both well above the minimums per the shop manual.

In answer to the question after a full charge - the voltage while not in the "start" position, but with the key in the "On" position with lights and other non-starter loads is 12.9V (a tad low, but not bad - I may yet swing by Autozone and have it checked).

In regards to the video (yes, I did watch it) - the guy actually does a very good job in explaining things - I actually use water concepts when I explain electrical concepts to others as well.

For those that may be reading this thread, some more information that'll maybe help them some time and help understanding... In the video he left out (and may not know due to his acknowledged weakness on the electrical side) the reason why his testing of the example bad cable misled him and why he saw what he did (other than it being a bad cable).

The meter he was using only gave him resistance readings down to 0.1 ohms. When cranking, Honda starters will pull a continuous current between 83 amps up to around 125 amps - that's with the starter spinning under engine load. The initial current (while coming up to speed) can be more than triple that.

If the cable actually had 0.1 ohms resistance, at 125 amps it would drop 12.5V (Voltage = Current x Resistance).

In the case of my '86 Honda accord, I have the Mitsuba 1.4KW starter. That means it can pull around 117 amps of current at 12 volts while spinning (Power = Voltage x Current or in this case, Current = Power / Voltage) - and the shop manual sets a limit of 350 amps during initial spin-up (yes, three HUNDRED and FIFTY :eek:).

Short tidbit - it would only take 0.034 ohms resistance to hit the minimum 8 Volts the shop manual calls out in the test (at 117 amps - would only be about 1/100th of an ohm at 350 Amps). On a meter that only had 0.1 ohm accuracy, it would read no more than 0.1 and might even read 0.0. This is why with most meters doing a test for voltage drop under load is a much better way to do it. They can easily measure and show differences measured in volts like he was getting, even though they won't provide good sub one ohm measurements.

A note here - smaller batteries like the standard sized ones in Hondas will show a decent voltage drop when cranking an engine over - this is expected (and why the shop manual has 8V for the minimum cranking voltage). Actually all batteries will to some degree, just the smaller spec'd ones for Honda more so. The reason - batteries have an internal resistance and as I talked about above, with the cranking current required by most cars (between the starter, ignition, etc.) it doesn't take very much resistance to get a decent voltage drop. Another informational tidbit - those CCA ratings on batteries? They are spec'd as the point where the individual cells of the battery still put out 1.2V, which would mean 7.2V across the entire battery (here's a link to an article by Optima on the subject: What does Cold Cranking Amps (CCA) mean? | OPTIMABATTERIES)



Thinking about this did also lead me to do another test - one of the simple ones that can be done to verify some starting conditions on a manual transmission. The results also eliminate one of the things that can happen to any of us when testing things to solve a problem - what I call "the law of expectations".

Earlier in this I noted I had taken the oil filler cap off the valve cover and verified that the valves were moving when attempting to start with the starter. Since I was doing it alone, it involved looking at the valves, reaching in thru the window and bumping the starter, then looking at the valves again. I was "sure" they moved and were in different positions each time (did it a few times).

But - that could have been "expectations" - I expected (and admittedly wanted) to see the valves move. So my mind could have been playing tricks on me and they weren't moving at all, but because I expected to see movement, I did.

Anyway, so I did another test. This afternoon when leaving work I left the tranny in 1st gear and turned the key briefly to the start position. The starter obediently tried to start the car and the car lurched forward a foot.

Verification that the solenoid is engaging the starter with the flywheel and that the starter is generated enough power (and the battery is supplying enough voltage for the starter to do its job). Also verifies I hadn't fallen prey to my "law of expectations" ;)

Unfortunately, I still don't have the actual cause identified and problem solved :( The discussion has led me to think of other things to test and eliminate - and led me to find out that it's spark I'm missing when trying to start it with the starter. Just don't know why it's missing :/ I do know I'm missing something (besides a key piece of info on the Honda ignition system that would explain the results, maybe some good ole fashion sense ... and will likely smack my forehead once it's figured out.

Anywho - hope you guys don't mind the super long post. Thought some of it might help others in the future. Figure since I'm here asking for help, I should share some info too that might help others.

In the meantime, if you or anyone else has any ideas or possibilities, throw 'em on out there. I'm still stumped....

If you look a few pages past the diagram you show, I think the green wire is for the pickup coil (signal for the ignitor to switch off ground).

I can think of two easy tests to try, because I'm out of reasonable electrical tests. What happens if you unplug the radio condenser and try to start the car? What happens if you unplug the alternator and try to start the car?

The condenser unplugged I've already done several times, once early on (didn't remember to list it - was checking for an internal shorting condition of the condenser), and then again when I monitored the voltage to the coil via the condenser lead while attempting to crank it (had unplugged the condenser and plugged my meter probe into the wiring harness connector in its place).

I'll try it with the alternator unplugged tomorrow night.

An additional test I did tonight that I'll do a follow-up test for tomorrow - my meter has a clamp-on for current. Put it on the battery and did several readings.

With the key off, verified their isn't any current flow.

With the key in the "On" position and the engine not running - 4.3 Amps of current.

Can't measure cranking current - outside the range of my meter.

I also tried clamping on the bl/yel wire to the coil both with the key on and while attempting to crank the engine with the starter - never saw more than 100 milliamps (meter has a peak hold function). This is the one I'll do again tomorrow night when I first get to the house, and see how much current goes to the coil with the engine running.

Regarding the green wire - you're likely correct, as it isn't present in the external wiring harness after closer examination. Tomorrow night (or near future anyway) I'll partially disassemble the old distributor and verify.

On a semi-unrelated note: my ancient timing light gave up the ghost last year. I've been looking into getting a new one, but have read a lot of negative things about the less expensive ones I've been able to find (mostly things along the lines of 'quit working after second use'). Can anyone give brand/model recommendations on current gen decent simple timing lights that aren't too expensive which they've personally owned?

As always - sincere thanks.

Update (still no luck in solving the problem)...

With the alternator disconnected, it still does not start with the key/starter. I didn't try a push start in this state due to concerns of accidental shorting of the dangling connections.

To check off another piece for comprehensiveness and to provide reference values for others, this time on the fuel side of things, I checked the throttle position sensor. It read 4900 ohms between the outside 2 contacts. Shop manual says 4000 to 8000 ohms is good. Also checked from one side to the center terminal. Reads 512 ohms with the throttle at idle, increases in resistance as the throttle is opened up, with full throttle reading 4900 ohms.

I was looking over the ignition system wiring diagram again and started making some other checks.

Disconnected the negative plug to the condenser (black wire) and measured to ground. Read 0 ohms as expected, so getting a good ground there.

I forgot to take the current measurement of the battery supply to the coil while the engine is running... next time.

Still no joy in resolving the puzzle....

My gut feeling is that the voltage is too low when you are cranking to have the igniter work properly.

I didn't see if you tested the voltage on the battery posts when cranking or just at tp1.

The other thing you can try is to verify that blue2 wire is getting a switching ground signal. Does the tachometer move when cranking the engine?