Someone relieve me of my ignorance, please

[j2k4]

I've been playing with my LCD boost control (courtesy of BlackStealth), and it's a nifty as it can be.

Been slowly bumping solenoid duty-cycle to push boost - my assumption is I should be able to run comfortably in the 13-14 psi range, given my minor mods (Stillen DP, 13Cs).

I've done a vacuum reduction and tested for boost leaks - I'm good for 15+ psi with acceptable leakdown.

Here's the deal:

From a dead stop, if I let the clutch out and accelerate at what you might call a normally 'brisk' pace, the likelihood is I will get ZERO knock; maybe a count or two at most.

If I hammer it through the gears, shifting at 6K or so, again, the likelihood is ZERO knock.

However-

If I idle off a stop, and roll up through first and second (no/low load, sub-2K rpm), I get knock - 6, 8, 12 counts...last night, immediately after running up through 3rd to about 6K, hitting 13.4psi and 1 count, I basically cruised for another mile to a stop, then pulled away gently, only to record a maxed-out 28 count!

I haven't got the turbo-car/3S experience or intuition to figure this out - I had suspected a sloppy rear mount or some other anomaly as responsible for this, but the low rpm knock counts are increasing as I bump-up the duty-cycle.

In other words, if I backed off to waste-gate boost, I will get, say, up to 8-10 counts under the same circumstances, but having pushed the dutycycle, this off-hand knock can peg a 28, apparently, and this gives lie to my theory about about mechanical vs combustion-type knock.

What in the hell is going on?

[donniekak]

When the tank is almost empty dump in a 5 gallon pail of some cheap(ish) race gas. Usually 100-107 can be had for a reasonable price. See if the knock continues. If it does, it is almost certainly false knock from old mounts, loose bolts, etc. also if your car is a 91-92 check that the base timing is set properly.

[j2k4]

When the tank is almost empty dump in a 5 gallon pail of some cheap(ish) race gas. Usually 100-107 can be had for a reasonable price. See if the knock continues. If it does, it is almost certainly false knock from old mounts, loose bolts, etc. also if your car is a 91-92 check that the base timing is set properly.

That's a brilliant idea.

It'll have to wait a bit though, since the closest race gas is prolly 85 miles away.

I can get booster, but I've never liked using it, particularly with unleaded gas.

Remember, I'm old.

It's still the best idea I've got, though.

[Boost4VR4]

I would say you probably have phantom knock or a bad knock sensor. Maybe replace the knock sensor? I think OEM is only 40-50 dollars if that.

[j2k4]

I would say you probably have phantom knock or a bad knock sensor. Maybe replace the knock sensor? I think OEM is only 40-50 dollars if that.

Possibility there, too; I can't afford to be in a big hurry, though, so I guess now I'm wondering what people might have to say about their experiences with functional-yet-defective knock sensors.

[futurevr4man]

Well if your knock sensor is "defective" you will read a consistent 7 counts (7 I think it is? Might be 9...) all the time. Again, it will be all the time. From startup on.

Not doubting Greg here, but according to him, our engines are noisy engines and the way the knock sensor is setup, is based off a raw knock voltage. That voltage goes through some filters and is converted into a knock 'count'. That knock count is a calculated number, not a direct amount of times your engine is actually predetinating. Which is why bad mounts can cause you to get 'counts' of knock even if your engine isn't actually detonating in a bad way.

With having said all that, the gas thing isn't going to change anything in all honesty. If your engine is knocking at no load with 93, it will have generally the same result with 116 octane. I'll all buy guarantee your engine always did this, you just couldn't see it on a little screen before. My car does it too, to give you a little stress relief. I'll bet yours goes away with any more or less load too just like mine.

I have been watching threads about this in curiosity and I have talked with Greg about it some too. The correlation between your waste gate duty cycle and the amount of knock is interesting though

[j2k4]

Well if your knock sensor is "defective" you will read a consistent 7 counts (7 I think it is? Might be 9...) all the time. Again, it will be all the time. From startup on.

Not doubting Greg here, but according to him, our engines are noisy engines and the way the knock sensor is setup, is based off a raw knock voltage. That voltage goes through some filters and is converted into a knock 'count'. That knock count is a calculated number, not a direct amount of times your engine is actually predetinating. Which is why bad mounts can cause you to get 'counts' of knock even if your engine isn't actually detonating in a bad way.

With having said all that, the gas thing isn't going to change anything in all honesty. If your engine is knocking at no load with 93, it will have generally the same result with 116 octane. I'll all buy guarantee your engine always did this, you just couldn't see it on a little screen before. My car does it too, to give you a little stress relief. I'll bet yours goes away with any more or less load too just like mine.

I have been watching threads about this in curiosity and I have talked with Greg about it some too. The correlation between your waste gate duty cycle and the amount of knock is interesting though

I can easily buy all that as well, Nate, and also equate that to a reluctance on anyone's part, generally-speaking, to belittle knock, as to do so with these cars can be fatal.

My assumption all along has been that my sensor is functional, but what you've imparted about how the ECU interprets what it's told makes a buttload of sense, especially given how it reads, then reports, knock.

The limit at 28 counts toward a conclusion of phantom/anomaly reporting.

My limited experience can't reconcile the low-rpm/load counts with the high/highest loads and their lack of knock.

Then again, this is the squirreliest car I've ever tried to get a handle on...

Hey, thanks for responding, guys - been wondering all day.

[CoopKill]

Kevin mine did the same thing as well. Sometimes when coming off cruise load. I was freaked at first, but somewhere in some thread on here I asked, and it was concluded phantom knock I believe. This as bpu status on 9b's.

[j2k4]

Kevin mine did the same thing as well. Sometimes when coming off cruise load. I was freaked at first, but somewhere in some thread on here I asked, and it was concluded phantom knock I believe. This as bpu status on 9b's.

Then I have to assume my 13Cs would be at least as touchy, yes?

[striker2]

Just today my CEL flashed (CEL flash on knock with the Chrome ECU) for a second and then right back off while cruising at about 1500rpm. its done this a few times before, and i even managed to log it once but didnt see anything out of the ordinary in the log to induce knock. I figure its from the solid lower mounts and poly mounts allowing the knock sensor to pick up on noises it normally wouldnt hear.