Ignition timing problem & weird idle behavior

Maintenance/Repairs
1

I have a 2001 Mazda Protege LX 2.0 with 121,000 miles. All of the maintenance (and then some) has been done in a timely manner - I am extremely anal retentive with this car. Last week, I was driving along at low speeds, and the car was idling smoothly, as usual. All of a sudden, when I pulled up to the next light, it acted like it wanted to stall - the idle started bouncing all over, and eventually my CEL came on. The code was P0300, for a random misfire, but other than the weird idle, the car acted fine - smooth acceleration, high speed driving, no power loss - not what you’d expect with a misfire. I should add that the spark plugs, wires, and ignition coils were all replaced about 5,000 miles ago, as were the EGR valve, vacuum hoses, and intake manifold gasket.

My scan tool showed that everything looked normal, except for the ignition timing, which was bouncing around between 10 BTDC and -2 BTDC (I assumed this was the reason for the erratic idle). I hooked up a vacuum gauge, and while it was completely steady, it was about 4-5 mmHg lower than what it normally is (reading ~17, normal is 21 or 22). I also did a compression test, and was getting ~180 psi across all cylinders, as per normal.

Perplexed, I took the car to a well-respected garage in my town, which has fixed many a problem on my cars quickly and honestly. They spent an entire day with it, and couldn’t figure out what was wrong with it. Their only idea was that maybe I had gotten some bad gas. Since I was on 1/3 tank, I filled up to see if I could dilute what was already in there.

Yesterday, I was driving the car at 55mph on the highway (road construction, ugh!), when I hear a slight squeal/squeak. It is more metallic sounding than a belt, and then all of a sudden I heard a little rattle, almost as if something got spit out the tailpipe. I didn’t see anything in my rear view mirror.

Since the car was still driving smoothly, I waited until the next exit to pull off, and as I came to an idling stop, I noticed that my idle was now again very smooth. It had been getting progressively better over the previous few days (no more CEL), but it seemed now like it was back to normal.

Finally, I hooked my scan tool up to the car again today out of curiosity, and I noticed that while the ignition timing is now steady, it is reading steady at 0 degrees BTDC. I confirmed this with my timing light. The spec from my repair manual is 8-10 degrees BTDC, but since everyother sensor seems to be reading correctly, and the car is driving smoothly, I am at a loss for what to do. I also hooked uo my vacuum gauge again and it is reading about 19mmHg, which still seems slightly low. I know there is something wrong, but I don’t know where to start looking.

Any suggestions?

2

I’d hook up a timing light to see what the timing actually is. I would not use a scan tool for this. Next, look for a vacuum leak around the intake manifold gasket. Something could have gone wrong with the install or it could have a material defect.

Do you have a timing belt, has it been changed according to the maintenance schedule?

3

In addition to the timing belt I would look at the cam/crank position sensors.

4

keith, I mentioned that I already confirmed the timing readings with my timing light. As for the vacuum leak, I have had propane all around the inside of the engine compartment, trying to find the source of the leak, with no success.

the timing belt was changed, along with the tensioner, idler, water pump, cam/crank seals, and serpentine belts, 15,000 miles ago, per the manufacturer’s recommendations.

circuitsmith, the camshaft and crankshaft sensor readings (according to my voltmeter) are within the allowable range, according to my repair manual.

5

nobody has any ideas? Where are all the “expert” mechanics on here?

6

Keith and circuitsmith have made a few suggestions but it’s not always easy to figure something out over the net because with any problem there are usually a few dozen details that may be needed to sort things out. It can be tough enough with car in hand much less sight unseen.

Random misfire can often point to a crank sensor. You have checked that with a VOM but that doesn’t mean it’s within the allowable range all of the time nor does it mean that it’s communicating with the ECM all of the time either.

The bouncing timing can cause a stupid idle and could be caused by a crank sensor circuit problem. Note “circuit problem” which includes the wiring and connectors.

Just curious about the vacuum reading. You state that you have 17" instead of the normal 21-22. Have you checked this car in the past and found it to actually have 21 or so inches of vacuum or is that figure plucked from some repair manual?

I ask because I can’t remember ever seeing a vehicle with 20" much less 21 or 22. (I’m in OK at around 1100 feet elevation but any 21 or so reading that exists should probably only be happening around sea level with everything spot on.

Maybe a knock sensor circuit problem seeing as how (if the figures are right) the timing is running about 8-10 degrees retarded?
With the engine idling take a hammer and firmly tap the cylinder head near the knock sensor to see if the engine stumbles or not.

7

ok4450, thanks for the comments - I apologize for coming off a little short - it’s just that I’m at the end of my rope with this car. The garage that I took it to really does have competent mechanics (they have figured out some pretty screwy problems on some of the older cars I’ve driven), but this one has them stumped.

I’m not sure how I can check the crank sensor besides measuring the voltage, which is what the shop manual indicates.

As for the vacuum, yes, the car has traditionally held 21 or 22inHg of vacuum at idle. I’m in Iowa, at about 900 feet above sea level - I’m not really sure what to tell you, but that’s what it’s always been at.

The knock sensor is an interesting idea. I’ll look into that. Thanks…

8

Just curious, what are the symptoms of a timing belt that has jumped a tooth or two? Would I still be seeing a smooth idle and somewhat lower vacuum, or would it be bouncing around all over the place?

9

Timing belts usually don’t slip…they break…Chains usually slip.

If a belt has slipped…it could run very rough…or on acceleration it’ll ping a lot…Might not even start due to lack of compression.

10

Even if the belt was one tooth off? I know the belt was changed 15,000 miles ago with a Gates timing belt kit, but I am running out of ideas as to what this could be, and I’m trying to think of anything that the garage didn’t do.

11

Even just 1 tooth off can screw things up, yes. If the timing is that retarded and you haven’t adjusted it manually it could be a belt that’s a tooth off. It could also be, as ok mentioned, the knock sensor pulling the timing back, either because the sensor is malfunctioning or because the sensor is sensing detonation and reacting to it.

What really intrigues me is the bouncing timing you had at first. I’ve seen that before, and it was caused by a dual-point fuel injection system’s fuel injector going bad. I’m not too familiar with Mazda products, but I doubt you have DPFI (that’s where you just have a couple of injectors in the throat of the TB rather than direct cylinder injection). It’s something to verify, though.

12

An engine can run fine with the timing belt off a tooth, or maybe even two. This can have a dramatic effect on manifold vacuum so rechecking the timing marks is something to seriously consider.

13

ok4450, I’ve read that ignition timing that is retarded due to a slipped timing belt will cause a bouncing reading on the vacuum gauge. Any idea if this is true? I’m just trying to avoid having to pull everything off to check the timing belt marks if I don’t have to.

shadowfax, I don’t have DPFI - it’s your standard sequential multiport fuel injection.

14

If a timing belt is a tooth or three off, you could not tell the difference by how the engine runs. You might get the reading that the timing is off but the engine would run the same.

You could get other results but there isn’t a 100% type of deal going in that situation. Finding the problem could be tough until an obvious accessory failure happens.

You know the scenario; a bearing that was siezed releases and makes everything seem to be good. Then it blows out completely and you have the noise or the puddle of fluid that lets you know what was causing the noise or the odd performance.

In other words, I don’t know.

15

Lol, it looks like I’m going to pull the valve cover when I get home tonight.

16

Sorry I didn’t get back sooner. I missed that part about confirming with a timing light. I would recheck that timing belt but I would be very shocked and appalled if it were off a tooth or two.

If it had been installed a tooth or two off, you would have noticed the problem right away. If it jumped a couple of teeth, then either it is a defective belt or the tensioner was not set correctly or tightened down properly.

I am liking the knock sensor only because the alternative ain’t so cheap. The alternative is a bad computer. The fact that it jumped around and now is steady at 0 BTDC makes me think that the computer has lost its ability to calculate the necessary advance.

It could also be that the computer has lost one of it’s necessary inputs, like the MAP sensor. Seems like that would throw a code though.

17

In addition to the knock sensor and potential computer glitch (which I agree with) let me pose another possibility.

Wht if the ignition module is not providing a signal to the computer and which the ECM relies on to control timing advance? An erratic fault there could explain the bouncing timing which has now settled down apparently to full retard.

The analogy might be when someone is setting the timing on a distributor equipped, ECM controlled engine and a test lead must be grounded, SPOUT connector unhooked, etc. In those cases the timing drops back to the retarded position and this is often in the 8 to 10 degree range.

The fly in the ointment though is that one would think there would be a code set if there was a loss of signal but with electronics like this you never really know.

18

These are some very good suggestions. I probably won’t have time to investigate this more fully until this weekend (I’m a full-time PhD student in civil engineering), but I appreciate the advice.

19

Just tested the knock sensor (resistance and continuity) per the workshop manual. Everything checked out normal.

20

Problem fixed. Turned out the MAF Sensor just needed to be cleaned. I’m kinda pi$$ed though, because the garage I paid $100 for “complete diagnostics” missed something so obvious. Part of me wants to go back and ask for a refund/give them a piece of my mind. Thoughts - is this something that absolutely should have been caught?