Who's there?
Knock sensor question.

I'm finally getting around to installing the 7730 ECU into my 87 GT with 3.4L swap.
Ran into a small problem installing the knock sensor in the old Camaro OEM location on the block.
I have Fiero Store headers installed and the electrical connector for the sensor hits one of the header tubes, so obviously the connector and wire will melt.
Not good.
Since I'm obviously going to have to use another NPT port somewhere else on the block, I just thought I should ask if the sensor will function properly?
I am assuming, (which usually gets me into trouble), that the sensor will *sense* knock vibration regardless of where it's installed on the block???


....
------------------

[This message has been edited by randye (edited 06-30-2010).]

Sorry I can't answer your question definitively, but I'd assume you're correct. For example, on my Northstar the knock sensor is in the intake valley between the cylinder banks, yet on my SBC it's on the outside of the block screwed into side of one of the banks. Some locations may be apt to transfer the vibrations more readily than others, but I'm sure you'll be OK.

quote Originally posted by Bloozberry: Sorry I can't answer your question definitively, but I'd assume you're correct. For example, on my Northstar the knock sensor is in the intake valley between the cylinder banks, yet on my SBC it's on the outside of the block screwed into side of one of the banks. Some locations may be apt to transfer the vibrations more readily than others, but I'm sure you'll be OK.

Thanks Blooz.
I don't presently know much about knock sensors, but I'm *guessing* that they primarily read amplitude rather than frequency of vibration.
(Knock being a much higher amplitude than the regular vibrations of the engine)
Unless someone knows better and tells me so, I just have to hope for the best I suppose, but it would make me rest easier knowing for sure.

[This message has been edited by randye (edited 06-30-2010).]

FWIW, I've seen knock sensors on the side of the block (5.0 in '88 Firebird) and on the back of the block above the bellhousing, behind the left head (4.3 in '01 Sonoma.)
I seem to remember at least one of the 3800 guys picking up mechanical noise from the transaxle, that the sensor thought was knock. He installed a transmission blanket to quiet the noise.
Keep in mind that if you have a noisy valvetrain, you also might trigger the sensor. I mention that because my 3.4 has always had a rather noisy valvetrain, ever since day one.
I've also heard that the SBC guys have a difficult time running gear drives for the cams, because of the knock sensor picking up the whine.

[This message has been edited by Raydar (edited 06-30-2010).]

Just for clarification, here is where the knock sensor mounted on the original Camaro 3.4 block.
It clears the OEM Camaro manifold just fine, and would clear a stock Fiero manifold too, but the FS headers caused the problem for me.
The sensor itself clears the header pipe just fine. It's the electrical connector, (approx. 1 inch long), thats the problem.
(Photo before I tore it down and rebuilt it...)


The only place that I have available to mount the knock sensor now is on the opposite side of the block, just above the new starter location.
It wouldn't fit here normally, except I have installed a smaller gear reduction starter, so there's more room.
(Photo taken during rebuild.)


I have the sensor installed now and the starter back in place and everything fits fine and has plenty of clearance.
Keep your fingers and toes crossed. We'll see if this works or not....

Which knock sensor did you choose? From what engine?

quote Originally posted by Hudini: Which knock sensor did you choose? From what engine?

Ryan, (Darth Fiero), told me to use one from a 92 Chevy Cavalier Z24, 3.1 V6, VIN T, because that's what my 7730 ECU came from.
I bought a new one at Auto Zone and harvested a connector, (fan switch connector), for it from another Fiero harness I had.

[This message has been edited by randye (edited 07-02-2010).]

The knock sensor will function anywhere that it is solidly mounted to the block. It works by picking up the specific frequency vibration given off when there is detonation. Since that vibration goes through the entire block, then any mounting location on the block itself will be fine. Just don't mount it to anything attached to the bare block like the heads. While it is likely the vibration will make it through the heads too, it will be weaker which could cause problems.

quote Originally posted by randye: I don't presently know much about knock sensors, but I'm *guessing* that they primarily read amplitude rather than frequency of vibration.

Like Bloozberry, I can't say for sure, but I think that most knock sensors are both amplitude and frequency sensitive. To the best of my knowledge, knock sensors are piezoelectric devices that are usually tuned (or filtered) to the high frequency components (i.e. the leading edges) of the characteristic knock signature; these frequencies and amplitudes are both significantly higher than normal engine "noises."


Edit: A quick scan of the USPTO database reveals that while most knock sensors are indeed piezoelectric, there are both resonant and non-resonant types. Some of them even claim to match real-time sensor data against predefined knock waveforms (i.e. probably frequency-domain "fingerprints"). It's hard to tell how much of this bandwidth shaping and/or postprocessing is being done in the sensor itself and how much is being performed downstream. A resonant (narrowband) or waveform-matching sensor is probably going to be much more sensitive with respect to the choice of mounting location than a broadband, non-resonant sensor. YMMV.

[This message has been edited by Marvin McInnis (edited 07-03-2010).]

quote Originally posted by Marvin McInnis: Like Bloozberry, I can't say for sure, but I think that most knock sensors are both amplitude and frequency sensitive. To the best of my knowledge, knock sensors are piezoelectric sensors that are usually tuned (or filtered) to the high frequency components (i.e. the leading edges) of the characteristic knock signature; these frequencies and amplitudes are both significantly higher than normal engine "noises." Edit: A quick scan of the USPTO database reveals that while most knock sensors are indeed piezoelectric, there are both resonant and non-resonant types. Some of them even claim to match real-time sensor data against predefined knock waveforms (i.e. probably frequency-domain "fingerprints"). It's hard to tell how much of this bandwidth shaping and/or postprocessing is being done in the sensor itself and how much is being performed downstream. A resonant (narrowband) or waveform-matching sensor is probably going to be much more sensitive with respect to the choice of mounting location than a broadband, non-resonant sensor. YMMV.

Somehow I just *knew* you would eventually speak up on this Marvin
I appreciate the research!
As it stands right now, the knock sensor is now the least of my problems.
The new 7730 ECU refuses to run the engine correctly.
I'm starting another thread seeking advice from 7730 installers with experience, and contacting Darth Fiero on Monday.
From what little I can tell, I probably have a bad ECU, (steady SES light on and no codes).

It will work couple years back was pulling a motor and somehow snapped the knock sensor off on my 95 gp. Took the one off the old motor stuck it in the gully under the plenum and it worked. Never got a cel. And was reading accuratly as far as I could tell.

quote Originally posted by randye: Somehow I just *knew* you would eventually speak up on this Marvin

I just thought of this analogy. Think of a broadband (non-resonant) knock sensor as working pretty much like an old crystal microphone, compact and rugged but sensitive to vibrations over a spectrum of several octaves. Then think of a resonant sensor as working exactly like a frequency-standard crystal in a radio transmitter; such a crystal will only generate significant voltage when mechanically excited at or very near its resonant frequency. Resonant piezoelectric devices will typically have a very high 'Q,' with a bandwidth of less than 1/10 of an octave. Both types are very simple devices, so GM could have used either back in the '80s and early '90s.

Sorry, but there's nothing I can contribute to your 7730 troubleshooting task.

[This message has been edited by Marvin McInnis (edited 07-03-2010).]