What's new
CK5
Register an account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members.

False Spark Knock?

longbedder

longbedder

3/4 ton status
Staff member
Moderator
Joined
Oct 26, 2005
Posts
6,697
Reaction score
10,281
Location
Ridgecrest, CA
1991 Blazer, stock TBI 350, 700R4

I'm wondering if I might have false spark knock retarding my timing. I'm running a scanner via ALDLdroid, and my engine vacuum is dropping drastically under load. These drops in vacuum correspond directly to the rate at which knocks are being sensed. My theory is that the knock sensing is cutting the spark timing so drastically that the vacuum is dropping noticeably. My ECM is a '747 and I'm not able to record spark advance/retard.

I verified that base timing is set rock solid at 0 degrees (ESC wire disconnected).

The vacuum checks out normal/healthy in neutral from idle to 3000RPM, and I verified it's accurate with a mechanical vacuum gauge (i.e.: the MAP sensor is working correctly).

I cannot hear any pinging coming from the engine, and I think this much pinging is a bit too extreme to be real, especially under part throttle and at cruise. Once I get lower on fuel, I'm going to try running some VP race gas and see what happens - if it shows pinging with 100 octane I've about gotta believe they're false triggers.

Any other ideas?

drivelog1.jpg
 
That's kind of busy lol. But do I see that you come off the throttle at about 328, right where it shifts into another gear, and knock goes way up? And again around 442 you've got a downshift under heavy throttle, with corresponding increase in knock?

Any chance the gas is old? It doesn't take long for ethanol to lose it's anti-knock properties. IME you'll be lucky if it lasts 6 months against a motor with decent compression.

It seems too consistent to be false. Maybe sensitive, but not thinking false.
 
The gas is fresh, and from a high-volume station.

At 328 was just after an accel to 55mph, then once I got there it upshifted to overdrive and started indicating ping at cruise. At 440 the throttle went to zero as I began to coast to a stop. I consider both of those conditions unlikely for real detonation (which should typically be under hard accel, no?).

KTM - I know if the knock sensor is disconnected, the ECM will go into limp mode and pull timing, but it pulls it consistently and doesn't let it come back up. Do you know if yours was giving false triggers or if it was simply not working at all (had gone open circuit)?

I do agree it's consistent - perhaps a mechanical vibration resonance at a certain engine RPM or load condition?
 
Mine was consistant with luggung the engine when in OD. Just after the shift if not accelerating, or durring a slow climb with little throttle input. Changed the knock sensor and it went away.
 
I was wondering about a mechanical issue...loose exhaust, broken flexplate, etc.

Coasting to a stop would definitely seem to be an unlikely time for knock, even with a mechanical issue. You tried running with no EGR? That's the only thing during those times that *might* influence fueling/timing.

Outside of just replacing the knock sensor, not sure how you would test it for sensitivity or failure in this manner.
 
ktmoutfront said:
Mine was consistant with luggung the engine when in OD. Just after the shift if not accelerating, or durring a slow climb with little throttle input. Changed the knock sensor and it went away.
Click to expand...

That pretty well mirrors my situation...I think I'll Amazon @rime me up a new Delphi sensor.

There is not a method I'm familiar with to troubleshoot a knock sensor for false positives (other way around is easy and it'll throw a code).

Thanks guys - I'll report back after the new sensor is installed.
 
I told this story somewhere, maybe here, and caught some grief from some folks that said it would not happen.
But I saw it.
Guy had a really crappy homemade trailer hitch and corresponding crappy trailer. Actually a pin type hitch, not a ball.
He said the truck ran great but would not pull his trailer worth a darn. My friend drove the truck, seemed fine.
Guy came back with the trailer on, and it was a gutless wonder. I figured the trailer was not rolling right.
Nope, you could push it around by hand.
Turned out the vibrations from the slamming and banging of the trailer and hitch were being picked up by the knock sensor as knock.
He fixed the hitch, and all was well.

All the knock sensor is, is a microphone tuned to pick up a narrow range of sound. It may or may not have some processing power to dress up the signal and send it to the computer.

As a quick and dirty test, you could remove it from the engine block, put it somewhere so that it did not sense the engine vibration, leave the signal wire hooked up and run a temp ground wire to the mounting stud.

That would tell you if there is some actual detonation going on. If all the knock signals disappear, I would suspect that you actually have a detonation problem. Of course, it might just be super sensitive.

The classic way to test one, is to watch the timing and hit the engine with a small hammer.
If its working, the timing will back off.
But, that only tells if its working or not. Not if its giving false signals.
 
I have been told I was full of crap many times for this issue, so I don't bring it up. But sence Fordum is posting the weird, I will too.

Had an exhaust leak on a set of headers. The ting noise from them was being picked up by the knock sensor. Fixed the leak, problem solved. A friend swears when he replaced the destroyed engine mounts, his intermittent timing issue went away.
 
J & KTM
I am fully aware of all you write of...hence the title of my post.

You see, I'm not a sort to semi-randumbly throw parts at a problem - I have a fundamental need to understand a given issue, and real detonation doesn't make sense to me in this case (part or off throttle cruise with a healthy engine). I also know that a knock sensor is just a narrow-frequency microphone.

I've known others who chased similar-ish issues and tore their hair out...only to find out something as plain as an exhaust note caused the problem (fixed by changing a muffler to a different series).

I do wonder what in a failing sensor might make it hypersensitive (perhaps a cracked/crackling pickup diaphragm?). If a sensor swap does fix it, I'm going to do my meager best to dissect the failed unit and figure it out...
 
Update: the detonation is real.

  • Installed new sensor, knock rate was the same (datalogged the same route at the same time of day).
  • Filled the half-empty tank with 91 octane and the knock rate fell off significantly (although not to zero). That right there told me the knocks being sensed were real. I should've done that before ordering the sensor, but I had a full tank at the time I considered the sensor. :doah:
Today I put a vacuum gauge on the line between the EGR solenoid and the EGR valve and ran it up to the windshield wiper where I could see it while driving. [I'd previously tested the valve itself and it checked out good, as did the intake passages.] Anyway, the needle does not budge under any condition whatsoever...so the EGR solenoid has either failed or it's not getting a command current. I'll test the solenoid tonight.
 
Update #2: The culprit is the switched ground signal from the ECM which controls the EGR solenoid.

I pulled the EGR solenoid and it bench tested good (pull vac with a hand pump, apply 12VDC to terminals and the solenoid opens and releases vacuum). So I rigged an LED and some leads from the solenoid connector. The positive lead (pink/black) is always hot when ignition is on (I verified the hot side is good - nice bright LED when grounded to chassis with key on).

I then drove around with both LED leads plugged into the solenoid connector - there were a few faint blips, but they were awfully weak and intermittent. So now I gotta figure out if the ECM ground is bad, the switch in the ECCM is bad, or the wire from the ECM to the solenoid is bad.
 
The hell of it is, the wires between the ECM and the solenoid check out and so do the ground leads from the ECM to chassis (as of tonight). The switch is solid state (in an integrated chip) which controls the solenoid. Sooooo...now I gotta replace an integrated cicuit on a printed circuit board or get a new ECM. ****.
 
Some systems, not necessarily yours, don't use a simple on/off system to power the solenoid.
Instead, they use PWM, pulse width modulation to control the amount of EGR that flows.
In those systems, turning the solenoid all the way on would let way too much flow.
So, the computer "blips" it with fast pulses. The pulse turns the solenoid on briefly, then back off before it opens all the way.
As the computer sees the need for more flow, the pulses stay on longer.

If that is your system, then the results you see with the LEDs is about what you would expect.
Do you have a schematic of your system? There is a gadget called a EVRV in the same circuit as the EGR solenoid.
Its called an Electronic Vacuum Regulating Valve. It might be part of the problem.

I vaguely remember something about that setup, but its midnight, I have been up since 4 in the morning, and my leg is protesting some of the things I did today.

Hopefully someone else who is more rested than me can jump in and let us know if you have a PWM system.

BTW, if you already know how PWM works, as I suspect you do, forgive the redundant explanation.
I usually never know for sure, and its easier to explain to begin with.
 
No need to pologize - the biggest point to continuing with posts now is so others can benefit from experience if they have similar issues.

As I understand it, this system is not pulse width modulated. The valve command should be a discrete 0/12 volts - the gas flow is modulated by exhaust pressure, springs, diaphragms, and orifices in the EGR valve.

I'll do some more looking around to confirm...
 
I never could figure out what ran EVRV and what didn't. For some reason I thought EVRV was a "heavy duty" deal, and "normal" EGR was lighter duty, IE 1/2 and 3/4 tons. Let us know what you discover.

You probably don't need it, but if you do, I have the '91 service/electrical manual set if you need me to look anything up in it.
 
Actually, an ECM internal schematic would be nice if it's in your manual. All I have are the pinouts and external schematics.
 
You must log in or register to reply here.

Latest Posts

Top Bottom