Miatas and premium versus regular gas

Josh,

The '99 and 2000 Miatas engines are the same. The timing is set at 10 degrees BTDC. The timing is determined by a fixed position CAS (crank angle sensor) that is mounted at about 4 o'clock at the edge of the crank pulley. If the CAS is stock you have other problems. There are articles that describe the modification or replacement of the CAS bracket that allows the stock timing to be advanced up to about 18 degrees. If you have access to a timing light you can check the timing.

If the timing checks at 10 degrees you may be running very lean. Inspect the plugs. If they are very white instead of tan you are running too lean. You could also have a carbon build up that is increasing the compression ratio and causing the problem.

My '99 runs great >> There are a lot of differences of opinion on this issue.

Yes that's what I figured.

But of course I did drive a number of '99s (with low miles... under 15K) before I drove MY 2000 model, and my car was decidedly more peppy... very noticeable. Made me think (at the time) that the 2k cars were different in some notable way. However, in retrospect, I think it's likely mine's just deviant in some way.

I haven't checked it, haven't attempted to modify it.

Well it's done this since the day it was brand new so carbon buildup is not a reasonable explanation.

It's done this with every brand of gas available in the Austin area. I've got nearly 60K on the car now and it's always been the same.

It's on the second set of plugs (changed about 30K.. time for another set I guess) but the last set was very tan looked perfect, just the end of the electrode was rounded off a bit as per normal plug wear. No reason to think it was running lean.

I have never really known what the explanation was for this car's need for premium gas but I never even really questioned it until gas prices got to over $2. I've been used to running premium gas for my last three cars so it was no big thing.

Strange. Do you think the knock sensor could be broke?

Mazda might have well fooled around with settings between 99 and 00, but if it knocks under all conditions, either they are braver than I would think or there is something weird about your car.

From your other posts, I understand that you got it new and it did it from the beginning.

Just curious, Leon

I suspect your compression ratio happens to be on the high side of production tolerance, due to pistons, head gasket, whatever. I think my '94 is the same: it's got more grunt than most, but used to ping on 89 in hot weather at 14BTDC, until I installed a Randall CAI. Now I can use

Well it doesn't regularly knock on 91+ octane gas. You have to really TRY to get it to knock (5th gear 20mph WOT).

IDK exactly how the knock sensor works on a NB but if it's like other cars I have experience with, then it's a piezoelectric element embedded in some kind of casing bolted to the block in a certain spot. Failure mode would be disconnect (which would cause an error code, likely retard timing, engine light come on), wiring/connection flakiness (intermittent engine light), overtorqued/cracked element (constantly sense knock, car would run like total crap and retard the timing), or undertorqued or the thing is defective in a way that it never senses knock, in which case you'd think it'd knock all the time no matter what gas.

So... given that it does not knock on 91+ octane gas, and it does knock regularly with 87 octane makes me think the knock sensor is probably working, but the computer is unable to sufficiently retard the timing to prevent knock with 87 octane gas. In that case, it could be excessive compression as Lanny suggested, or excessive static timing advance, some of both, or just somehow Austin, TX has ultra-bad gas at every station in the NW suburbs. Either that or the knock sensor works ok under some conditions (premium gas) and does not sense knock when it's excessive...

My experience with other cars (my VW 16V Jetta, for example, with 11:1 compression ratio, 7800rpm redline and two knock sensors) is that when the knock sensor goes out or malfunctions, it results in the car running extremely bad because the fault mode of the computer is "retard" in more than one sense. This is really not how my Miata acts, but quite the opposite.

FWIW the only mod to the car is cold-air intake (homebrew... K&N filter and plumbing, mostly for "the sound" and not really performance improvement at all). I did the CAI mod within the first 5k miles. Can't remember if it knocked before that or not.

Yes that's correct.

I'd try reinstalling the stock intake. Isn't there a thermosensor in the airbox?

Yes there is (was). It's reinstalled in my intake.

It may not be working properly.

Ya think? ;)

now, if I knew the definition of "correctly" then I could probably debug it. However I wasn't really sure it was a thermosensor... but I can say with authority that I didn't really know or care what it was. I figured any sensor in that location couldn't be ultra-critical (and I guess I was right, since the car's been running fine for five years with only complaint being pinging with crap gas). I guess I shoulda bought a shop manual before monkeying with it. If you pull it out and put a cork in the hole, the car won't start.

Funny how I can't really find much info on the operation of this thing on the net. You google "Miata cam" and you'll get dissertations on the operation of an internal part most folks are unlikely ever to tinker with, but no info on a part that you have to touch if you do as much as change the air intake plumbing like every 17-year-old rice boy on the planet has done.

The temperature sensor is most likely a temperature-variable resistor, also called a "thermistor".

You can probably gain some insight from looking at the Megasquirt EFI page or perhaps at Dick Bipes' ACU documentation.

Cheers, Dana

I too have no information on what exactly this thing does or think it does, but it has been suggested on this group that the advance was not very aggressive. Mazda might have designed the engine control for 91 octane, relying on the knock sensor to take out timing if that proves too aggressive for 87 octane.

Retarding timing by sensing knock is not a very reliable procedure, because when it knocks audibly, you probably already have glowing carbon deposits causing it, and making things worse.

I would not. As a supercharger owner, I know 91 makes a big difference in whether it knocks or not compared to 87.

Leon

My 2002 SE, pre-JRSC, would ping slightly under WOT@3k RPM, like when climbing a steep hill in a low gear at low speed. I think the motor probably needs genuine 93 octane to avoid knock/retard.

Dana

So you're saying that the ignition advance provided by virtue of input from the intake air temp thermistor is not very aggressive? Or the timing curve is not too aggressive on the whole? Frankly I have no clue what the ECU function is but if this intake air temp sensor has much of an effect on timing advance I'd be surprised, particularly in comparison to the input of a knock sensor (or four). My guess is that the ability of the ECU to retard the timing is insufficient to prevent knock on 87 oct. gas given whatever other combination of variables and tolerances on my particular car. I guess a timing light would settle it once and for alL!

I'd think the air temperature in the tract between the air filter and the throttle is reasonably consistent and exact location of this sensor should not make a really big difference in running of the car. Certainly it can't be sensing anything very fine... 10degF or larger resolution? Hard to say. Mainly the only thing different in my car is where this component is located in the intake plumbing. I guess it's possible it's getting an erroneous reading.

FWIW I found some notes... I put this intake in with 600 miles on the car. So having not broken it in 100%, there's no way to recall whether it truly "always knocked" or "always since I did the intake mod".

I don't know much about NBs, but I'd suspect the sensor affects mixture, not timing. Of course, being too lean can cause pinging, too.

I was talking about the knock sensor, but on the link ECU's, ignition is retarded when the intake air temperature increases too. But as Lanny says, the problem with having a nonstock intake might be too cold air for the amount of fuel supplied by the ECU, giving a lean mixture. The stock airbox takes in warm air at a regulated temperature.

I should have somewhere the Link ECU timings for the nonaspirated

Sure. It could be heated by a variety of problems (boundary layer air, heat conduction through the conduit, radiation, etc. But IIRC, you said that your plugs did not show evidence of a lean mixture?

Leon

Come to think of it, that is for the M1 compression ratio.

Leon

That's right.

But I haven't had them out in 30K miles. Worth it to check that too. The orig. plugs just had regular looking plug wear and were nice and tan.