[codingparadox]

I picked up a completely stock 06 non-sport (US, 2ZZ) Elise a couple months ago. The short version of my problem is _absurdly_ bad hesitation at low RPMs (below 2500-3000) when cold, which falls back to just _very_ bad low RPM hesitation once it's warm. When the car is cold, I could go WOT at ~1500 RPMs in first and have it take several seconds to slowly accelerate up to 3000 RPMs. Once it's warm, it's better, but you can still feel a huge torque dip (probably 1/4-1/3 the torque) below ~2500. It makes it annoying to pull out of a stop/etc. Looking at dyno plots of stock cars that go down to 2k and a hair below, it looks like it's still making quite good torque down there normally, so it seems like something's wrong.

Once the car's moving and above that it feels basically fine (maybe slightly slower than my old Elise, but it's been years so it's hard to calibrate my butt dyno).

I have an OBD2 scanner, and there's no codes being thrown of any sort. From reading some other threads tonight from supercharged guys, everyone seems to suggest changing the pre-cat O2 sensor. I'm wondering if that's just an SC issue or any-car problem. Anything else to suggest while I go diving? Checking the air filter and plugs are easy enough, and I even have some coils off my Miata (I'm using the 1ZZ coils, which I assume are the same?) that I could try swapping in if it's they're the same to see if that helps. Given that it runs fine up top and it's just crappy down low makes me more doubtful that it's ignition-related, though, but I've been wrong a pile of times before.

Any thoughts? Later this year I'm going to be doing some serious mods and I'd like to get the car running in top shape before I go down that path, so this is the beginning of my fun...

 

[Gunpilot]

I would say your cheapest route is change the Lambda sensor first.

 

[cyow5]

It sounds like what mine did when I put a CAI with the wrong size diameter at the MAF sensor. Are you SURE it is a factory intake? If so, I'd also listen for an air leak. Next, I'd pull the plugs and make sure all four are burning and burning similarly. With the plugs out, check compression if you have a gauge. All these checks are free (you can borrow compression gauges from Autozone, Oreilly, etc). If they all check out, the lambda sensor would be a good sanity check

 

[glagola1]

Clean the MAF? Reset the computer?

 

[codingparadox]

Intake is definitely stock. The only thing that may not be stock is the air filter, which I haven't checked yet.

Alright, I'll give this all a shot:
1. Pull/inspect coils
2. Pull/inspect plugs
3. Compression check while they're out
4. Check air filter
5. Inspect/clean MAF
6. Reset ECU (is that just a matter of doing an OBD2 code clear, or is that disconnecting the battery for a while or some other process?)
7. If that doesn't do anything, change pre-cat O2 sensor.

 

[shay2nak]

these cars have no power below 3000 (even supercharged) and it's already falling at 3500. Not to mention these motors don't like being below 3000 RPM.

 

[sturgeongeneral]

1. Change plugs
2. Compression test
3. Leak down test
4. Check for wiped cam

 

[codingparadox]

these cars have no power below 3000 (even supercharged) and it's already falling at 3500. Not to mention these motors don't like being below 3000 RPM.

Sure, but there's a difference between 60-70% of peak torque and 25% of peak torque.

 

[codingparadox]

Okay, I forgot what a bitch the air filter on these things are, so I'll wait until later to do that...

Pulled the coils and plugs. Coils had an odd red dust on the bottom, but #4 had rust on it (must have gotten water by it or something):

Next, pulled the plugs, and, of course, #4 was super rusty. There was also a bunch of discoloration on the threads on all of them:

While it was apart, I compression checked the motor. I had just driven home from work. It was up past 190 degrees, but I hadn't run it more than about 15 minutes, so it's probably not super warm oil temps. In any event, I didn't pull the ECU fuse like a moron, so there was some fuel in the air afterward, but I did do WOT for each check. The results were 180/190/180/160. Not ideal. So, the engine's obviously worn. I'm also somewhat skeptical of the quality of my compression tester, since it falls down about 20psi between cranks before settling. More over, I poured a bit of oil into #4 just to check, then ran the test again and the tester would never go above 0, and I tried 2 other cyls too, all of which immediately dropped back to 0 (I saw it quickly losing pressure back to 0, so it did go up) so it may have just given up the ghost on this test. Either way, given that I'm going to replace the motor in the next few months, I'm not worried, but it's good to know that it's not in super awesome shape. I'm gonna go buy a new compression tester at this point. This one's lasted me 10 years...

I had some spare BKR7EIX's kicking around from the Miata that I cleaned up and tried swapping in. I also cleaned off the coil boots and coated them in dielectric grease, as per several other threads' recommendations, and reassembled everything. While the car was already warm, I could tell it was already better just cruising around the driveway. Oddly, it feels like high RPM may be a bit slower, but that may only be because low RPM used to be much worse, and now with lower RPM being better, I can't tell the difference at high RPMs as much. Either way, I'm much happier. I'll prolly also pick up some 6 range plugs so that I'm not cooling down my charge too much.

Hopefully when I let the car fully cool down and try driving it cold it's still much better. It was a complete useless dog for a minute or two before it started to warm up before this...

 

[codingparadox]

For some followup, I've been DD'ing the car since then. While it's definitely better than it was before I did all the above work, it's still VERY weak below ~3k when cold, and still fairly weak below ~2.5k when fully warm.

I bought a much nicer new OTC compression tester and rechecked today, after cruising the freeway for half an hour. 185/190/195/195. I think the cyl 4 being 160 before was actually the old gauge's dying breath.

Those numbers are lower than shop manual specs (203+), but not tremendously low. What else might I check?

Any chance the knock sensor is causing this? On my old 88 Celica Alltrac, it did something similar to this when really cold, which turned out to be the knock sensor detecting piles of knock when cold, but that's a slightly different era technology... If I just pull the knock sensor connector off, will the ECU freak out or will it just not respond to any "knock"? Failing that I could try unscrewing it from the block entirely while keeping it plugged in, and just zip tie it to something while inside a foam envelope or something?

I'm going to do a baseline dyno run on Saturday at a dyno day, just to see where it's at, but I suspect it's going to be a very differently-shaped curve than my old Elise, on the same dyno, 5 years ago...

 

[codingparadox]

Went out to the dyno and was originally pleased with the numbers, until I compared the shape of the curve to my old Elise back in 08, on the same exact dyno... Purple is the original car, the other 3 pulls are today:

At the very top end, it almost catches up, but everywhere down low is way down on torque, which matches my butt dyno feel of the car compared to the old one. You can also see the torque dropping off like crazy below 3k (the dyno starts at about 2300 rpms, each vertical hash is 1000 rpms). Oh well. Low compression it likely is. I don't think a flaky knock sensor pulling timing would cause a curve like that, since it should consistently pull timing all the way through.

 

[holmz]

The ECU has protection modes for WOT when cold.
You would need and ECU or ECU reflash to mitigate that.
Or just do like everyone else and wait until it is warm to put the boot into it.

I doubt it is your DBW, but you may want to check it.

It seems low everywhere, but I would not say it is dropping like a rock below 3-1/2k.

 

[codingparadox]

It's not even WOT. Low-part-throttle (standard granny accelerating out of a stop) hesitates a ton when cold and requires pretty heavy clutch-feathering unless you really just give it the beans. Even fully warm it still is way down compared to where an Elise that was over here last night is.

And it's not _everywhere_, it's everywhere below about 8.2k.

 

[holmz]

Unless you have some quantifiable numbers other than a dyno slip, then it seems like guessing as to what may be the issue(s).
But the compression number tells us it is not compression/rings.

Conceptually that leaves mechanism like:
1) throttle (and/or exhaust flow like failed cat, which is unlikely as it only matches in HP at high RPM)
2) fuel
3) spark advance.

At some point you probably want MAF and lambda numbers.

 

[codingparadox]

Yeah, I won't be able to get anything for that until I go with the BOE stage 2 tuning stuff over the winter. Totally stock ECU, so no datalogging...

Aren't those compression numbers pretty low? Googling finds people mostly getting 230ish on healthy engines.

 

[me73]

Even fully warm it still is way down compared to where an Elise that was over here last night is.

Except at idle, where mine dies

I have obd2 software on my Surface with logging. We can check to see if it's pulling timing.

 

[codingparadox]

Oh, I didn't know you could even do that with OBD2. I've been monkeying with OBD1 cars too long.

 

[holmz]

Except at idle, where mine dies

I have obd2 software on my Surface with logging. We can check to see if it's pulling timing.

^Nice^
Does it show Lambda as well?

One usualy wants an assistant or to log it rather than text-n-crash.

 

[me73]

^Nice^
Does it show Lambda as well?

One usualy wants an assistant or to log it rather than text-n-crash.

It can do lambda. Agreed, logging is invaluable.

 

[codingparadox]

I've been spoiled by always using standalones for the last decade for any of my modified cars. This stock thing is weird.