ac4000

Symptom
When the engine is cold and in neutral, the engine accelerates smoothly to around 2500 rpm, then rapidly fluctuates. When on the freeway, the effect is harsh--feels a bit like the transmission is shifting up and down to keep up (which it may be). Here's a video: http://www.youtube.com/watch?v=P3kE_a-703Y .

Goes Away When
Once the engine has warmed up (usually around when the climate control starts sending warm air into the cabin, not just when the needle gets to operating temperature), the problem appears to go away completely. At around 35-40 degrees F outside, it takes about five miles on the freeway at 60-65 mph to get to normal, smoother operation; it takes a little longer on colder days.

Other Possible Factors
I'm getting a CEL and error 13 (RPM signal no. 2), but I do not know if it's related. Low coolant light, probably because of an actual low coolant condition, which will be corrected soon.

What I've Done
Cleaned the AFM as much as I could (it's a Karmen-Vortex type, so there wasn't much I could do) and tested the temperature resistance (it is within spec). Replaced the air filter, which was as dirty as I've ever seen, which seemed to help a little, but only temporarily (leading me to believe it was actually just higher outside temperatures). Replaced one of the O2 sensors recently to clear that code, but that's probably not related.

Searches
I've searched the forums for similar problems, but I've not found anything that quite matches. The hesitation problems that many have described (a pause and then power, often coupled with rough idle) do not match what is happening here.

Question
Any thoughts? I have a few but, since my expertise is with MR2s, I'll not pollute the brainstorming with my ideas just yet. I'm looking for things to try/test with a theory behind each one so I'm not ruthlessly replacing things. I'm happy to dig in there if necessary (there's so much more space than in a '2!). Thanks in advance for any help--I'm looking forward to working through this!

ac4000

Cleaned the throttle body just for the heck of it. No dice, although the engine does seem smoother and there seems to be more seat-of-the-pants torque.

Bradster

Seafoam and fuel injection treatment..

LiCelsior

iTrader: (32)

clean the idle air control valve.

PureDrifter

iTrader: (10)

what trans fluid are you running? also, check the wires goiing to the CPS.

ac4000

Thanks for the suggestions! It should be Toyota ATF in there, but since it does this at idle, I don't think it's transmission related. Checked the CPS wires as best I could on two different occasions: nothing frayed, nothing loose. I could clean the idle air control valve and/or do the Seafoam, but:

I just got back from the car and I think the timing belt is jumping. Here's what I did: 1) Marked the camshaft pulley and the belt at two places on each pulley/belt. 2) Started the engine and revved until I got rev fluctuation. 3) Discovered the belt marks had moved away from the pulley marks.

So here's my theory: The shop that did the timing belt replacement a few years ago (not a lot of mileage since then) did not adjust the tension correctly so when the engine is cold the belt jumps under acceleration (this also explains why the rev fluctuation range moves down when the engine is under load). When the engine heats up, the pulleys and belt expand enough that the belt no longer jumps. The CEL 13 is presented because there is too great an angle between readings (i.e., the belt has jumped). The only thing I can't figure out is how the engine can still run mostly just fine under these conditions.

Any thoughts on that theory? I'm happy to do the other things if the theory doesn't make sense because they're a heck of a lot easier than redoing the timing belt.

ROLEXUS400

iTrader: (2)

ac 400 did you fix your problem? im curious to know what the fix is or was.

ac4000

Not yet, but I'm 95% certain it's the timing belt slipping. Thing I can't figure out--and I want to before I dig into and start disassembling--is why it's slipping. That is, what was it about the installation (done by a shoddy shop) that resulted in a loose belt? I've pored over the diagrams and I can't figure it out. The tensioner is hydraulic and doesn't appear to need adjustment from the shop manual instructions (and it's brand new), so I'm stuck on this one thing. I do, however, hope to get in there and figure it out soon and I will update the thread, so keep an eye on it.

ac4000

Okay, I've unbuttoned everything and I'm at a loss. The hydraulic tensioner seems to be at spec (can't be compressed by hand, which is what the service manual suggests to test it, and it looks brand new, but I haven't checked the protrusion yet) and the pulleys seem to move freely (although they don't look only two years old, which is what they're supposed to be). What I've noticed in there:

1. lots of dirt and debris around the pulleys
2. some oil behind the LH side
3. LH camshaft pulley is much harder to turn by hand than the RH pulley
4. coolant was very viscous when I drained it--almost like transmission oil
5. timing belt cover gaskets don't look to be in great shape
6. no. 1 timing belt cover damaged
7. didn't find a timing belt spacer gasket although the service manual said I would
8. timing belt is Toyota 13568-50010 (same as 13568-59045, as far as I can tell)

So I'm at a bit of a loss. What could cause the belt to slip? It started doing this after the timing belt change (back in '07), so it's likely something the shop did something, but I can't figure out what. What else should I check?

Banshee365

ac4000, I'm all for you ruling out the timing belt slipping scenario. When you get in there you'll see the belt has pretty tough teeth. The belt is installed in one position and does not move from there, it would take an enormous ammount of force and the teeth would most likely sheer if it skipped. I've never heard of one skipping, it could possibly if the tensioner was bad though. It is possible that the last tech installed the belt a tooth or two off timing but your car sounds pretty smooth in that vid. You also answered your own question about the tech not tightening the belt enough, there is no adjustment.

Did you remove your timing belt to turn the cams? That belt has to be on there at the right spot, it's not like a serpentine belt that you just put on and it turns your accessories. There are marks on the belt and cams/crank that show the belt lined up, don't mess with that unless you know what you're doing. I do not thinkg you problems have anything to do with you timing belt.

In you video you engine sounds pretty constant while the gauge is bouncing. Go off the sound of the engine rather than the gauge. Close your eyes and rev your engine up and think to yourself if it is surging. I would get rid of the RPM sensor CEL since that is related to you needle bouncing. That 2,500 spot seems to be a bad spot, whether in the cluster itself or the sending unit.

If your car left the shop with problems they may have damaged a wiring harness or connector and simply didn't make a good connection on something. Get rid of your trouble codes before troubleshooting other random stuff. You'd be suprised what can lead to different thing. What does your engine idle at when cold? It should be running a little over 1,000 rpm or 1,500 or so at first and slowly go down to normal after it warms up. Does yours do that? I'd say look into replacing your ECT sensor for the ECM. It could be reading faulty coolant temps until warm causing the ECM to get wacky signals.

ac4000

Thanks for the detailed post, Banshee. I'll go through the things I've tried in response. The reason I think the belt is slipping is because, before I unbuttoned everything, I marked spots on the camshaft pulleys and on the belt, then let the engine do the rev dance (as in the video) and the marks had moved. Timing belt slip isn't common, but it does happen--the service manual even points to it as a possibility for a CEL 13. I agree that with proper tension in the belt it seems like quite a difficult thing to accomplish without tearing up the belt--and the belt isn't that bad, which suggests it was loose.

I did, of course, remove the belt to turn the cams. Wouldn't make sense to try and turn one by itself with the belt off. I needed to turn them to the match marks because, since the belt was slipping, they weren't in their correct positions. It remains odd that the LH camshaft seems frozen (though I have not tried turning it other than by hand yet; I just offered the difference in turning resistance as a possible problem).

It's definitely, without question, not a gauge problem. The engine actually jumps and hesitates. Would have noticed a gauge problem immediately. Chances are it's not a sensor problem either since I've checked all the wires and all three speed sensors (both cams and the crank) are two years old. Idle's fine. I'm interested in your ECT/ECM theory, but keep in mind the belt is slipping, so I'm not sure it would make sense to follow that one until I know the belt is on right.

Banshee365

When you install the belt at TDC, turn the engine 2 revolutions (since it's a 4 cycle) to double check the timing. Man it would start running really really bad if the belt was slipping. I don't remember if the '90 is an interference engine or not, I don't think it is. It was and the belt was slipping that's a bit dangerous to your valves.

The Toyota timing belt should have cam1 and cam2 listed on the belt for the timing marks. You may be starting the engine and shutting it off to pull the timing belt covers and see the crank mark on the belt not aligned with the cam or a cam mark not aligned with the other cam. You may be looking at it right, I'm by no means trying to say you're doing anything wrong. But man, a timing belt slipping with a tensioner that's in spec would have to be caused by something VIOLENT. If it was slipping more than a tooth you would start getting knock sensors going, spitting, sputtering, and backfiring. Depending on which way it possibly slipped it may juts shut off. You say the car runs fine when it warms up. Do the symptoms that you think are caused by the belt slipping go away when it warms up and come back when it cools down?

ROLEXUS400

iTrader: (2)

If you ohm out cam sensor #2 what does it read? And just a heads up when you put the belt on its a lil tricky to get it right on the money. I had your same exact isssue and had to correct the passenger side cam timing. Be very picky when you get it where you think its right. If it looks a hair off it is. When you get the marks where they need to be,youll know goood luck

ac4000

Banshee, I'm starting to think your theory that it's not slipping might make sense. The belt does look a little worn in some places, but mostly new, and I figure I would have seen more wear. Plus even though the '90-'94 (I think after that is when they changed the head and increased the horsepower) engines are non-interference, I can't imagine the car would run at all if the belt were slipping. I also think my test might have been incomplete: Would the match marks have moved apart and then back together after another rotation?

Rolexus, was your problem exactly the same? I read your thread about it, but I wasn't sure whether you had the same fluctuation (as in the video I posted at the beginning of this thread). I'd definitely like to know whether it was the same problem because if the "slightly off" belt causes that fluctuation, then belt slip wouldn't need to be the issue.

Believe me, there's going to be a lot of staring at and double-checking that belt before I button things up. Waiting for a new belt and pulleys to arrive; no point not changing them now that everything's open. I'll take readings on the sensors when I'm out there next; all three are new though, so I doubt I'll find anything interesting.

ac4000

The length of the belt is approx. 1671 mm (from the Sun Belt number, converted from inches) or 1672 mm (from the tooth pitch, 8 mm, multiplied by the number of teeth, 209). The camshaft pulley has a diameter of approx. 120 mm, so a circumference of approx. 377 mm. That means approx. 4.43 rotations per belt length. I take this to mean that every 30 camshaft pulley rotations the match marks would realign, but that using match marks is an absolutely bogus way to determine whether the timing belt is slipping.