Knock Sensor

[bc...@juno.com]

What about the sensor; is it just a microphone, or does it have some
specific frequency filtering characteristics too?

Bruce Roe
++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Switching Back to a Q-Jet? Only kicking and screaming.
From: "'Johnny Bridges' via GMCMH EFI" 30 May 16

Just frequency isn't going to do. You need rise and decay time as well.
Also amplitude floor at least. What you're trying to do is identify a specific
waveform. It's doable, but it isn't easy. I'd try feeding white noise bursts
into the filter from a working system, and see what comes out the other
side, and reference that to a triggered 'knock' and see what it looks like.
If you can borrow a storage 'scope and trigger a single sweep when the
sensor decides 'knock' you can get a pretty good picture of what a 'knock'
looks like >in that engine<. Then get a picture of what a 455 knock looks
like and compare. Some fiddling around ought to produce a setup which
responds to a knock but little else.

I'd expect to find a lot of the discrimination to be done with the sensor
where it would be relatively cheap to do. This would also simplify circuitry
down the line.

--johnny

[Matt Colie]

On 6/1/2016 11:28 AM, bc...@juno.com wrote:
> What about the sensor; is it just a microphone, or does it have some
> specific frequency filtering characteristics too?
>
> Bruce Roe

Bruce,

In very simple terms, yes and yes, but the answer is item specific.
As usual, generalities are typically in error.
Internally they are much better described as both frequency and
directionally specific accelerometers. When we were tuning them, we
had to do all manner of things to try to reduce the sensitivity to all
the other crashing and banging that goes on in a running engine. We
fought and gave up try to get one vehicle to stop retarding the spark
for railroad tracks.

Matt Colie

[Dolph Santorine]

Matt

They are piezo electric transducers, aren't they?

(Or "Piezo Benders" as they say in the speaker biz")

Dolph Santorine

Do...@DolphSantorine.com

Excuse me for not being my usual wordy and sporadically verbose self. This message is sent from my iPhone.

No trees were killed in the sending of this message and few long dead dinosaurs were involved. A large number of electrons were terribly inconvenienced.

> --
>
> --- You received this message because you are subscribed to the Google Groups "GMCMH EFI" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to gmcmh-efi+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Matt Colie]

Dolph,

Both piezo and electret were used.
Piezo being favored for its low total cost as the required pre-amp could
be a part of the main board, but the low output had noise immunity
problems.
The electret were more expensive as the oscillator, demodulator and
filters were all packaged in the unit. This also made them pretty much
type specific.

Matt Colie

[Gordon]

Matt,

Did you put the in the water jacket or in a blind hole?

I was just reading about a Porsche which started with 10 sensors in the
intake manifold and heads. They found 2 positions with a signal they
could work with.

I am guessing taking a GM sensor which came from a 350/454 engine
originally installed in a blind hole and putting it in a Olds block with
the end in water is the WRONG thing to do? That is why it works better
on the front engine mount or on a strap off a pan bolt.

[Matt Colie]

Gordon,

I guess we weren't the only ones to do the porcupine game....
We never tried the intake because it was isolated. This was a OHC
engine and that left little room on the heads.
We tried to avoid going into the cooling jacket. It had nothing to do
with the coolant and everything to do with warranty.
I would also be inclined to believe that the coolant might act as a damper.

Can you make it work on a strap off a pan bolt. That would be a great
place.

Matt Colie

[Bruce Hislop]

My thinking is the sensor is pickup up the "sound" directly (mechanically) from the block and not the sound we hear through the air.  I would think a sensor mounted on a bracket would be more "damped" than directly the block.

Just my farmboy shade-tree or fix it in the field mechanic point of view.

Bruce Hislop
Perth Communications
www.perthcomm.com
Please note our Office local phone number has changed to 519-273-3307
800-565-9983 & FAX 519-273-4111 remain the same.

[Gordon]

The strap off the pan is what some of the Olds (and other) guys have
used in thirdgen.org http://www.thirdgen.org/

Many prefer the front motor mount, especially on the Olds.

[Gordon]

My WAG is no matter how bad the sensor, position of it and filter are, they are still many times better than my ear and wrench on the distributor.  At least you can go to the Whatsup and see what is going on with knock counts, timing and fuel tables.  The sound in the air is like the little red light on the oil pressure.  It is the "game over" sound.

[Johnny Bridges]

This begs the question - maybe Colie knows - does the sensor location matter as to which cylinder is knocking?  Or do they all sound enough alike to set it off?

--johnny

---

From: Gordon <Wiz...@telus.net>
To: gmcm...@googlegroups.com
Sent: Wednesday, June 1, 2016 4:53 PM
Subject: Re: [GMCMH-EFI] Knock Sensor

--

---
You received this message because you are subscribed to the Google Groups "GMCMH EFI" group.

To unsubscribe from this group and stop receiving emails from it, send an email to gmcmh-efi+unsub...@googlegroups.com.

[Keith Vitko]

I've wondered about knock sensor location and sensitivity for all cylinders.

Porsche made cylinder bridges for the knock sensors on their opposed 6 and others have copied the technique on custom v8 and v6 builds.

I wonder how cylinder bridges (2 knock sensors) would work on the 455 attached to the head bolts.

To unsubscribe from this group and stop receiving emails from it, send an email to gmcmh-efi+...@googlegroups.com.

[Gordon]

There is too much valve train noise in the head and block from a 455.  The consensus seems to be the front motor mount?

[Gordon]

Bruce Hislop referenced a write up on analyzing the knock from a 600hp Porsche. 

[Matt Colie]

Oh Boy.....

This is one of those questions.... 
Remember: Generalities are typically in error. 

Simple answers are Yes and Yes  (OK?)
It is largely a specific engine problem.   Not which type engine - What serial Number? 
On a more or less typical America V8, the coolant flow is from the radiator down the banks and back the heads.  This means that the front of both heads are hottest.  That is a major contributor. 
So is mixture.  Again a typical engine runs richer in the center and leaner at the ends.  This helps that front pair can be hotter, but the lower density actually decreases the probability of detonation.
This all assumes that the head castings are good enough so no combustion chamber is very different. 
Still with me here??

In a test lab, we can detect a lot that you can't hear, and the equipment to do it is only a few thousand dollars (per cylinder).  I have seen conditions where some knock was not detected as well in some cylinders as others, but in most cases (another generality) a properly tuned system did detect the event regardless of where it occurs.   This is one of the reason that they are usually located in the cylinder/crankcase.  It is all one casting and there are no buffers like headgaskets to deal with. 

Summary:  Nobody Knows until the test map is complete. 
So, if you have a location that works at all, it will probably work most of the time.

If that was not an effective answer, please try again and I will try to file off the fuzzy stuff.

Matt Colie

To unsubscribe from this group and stop receiving emails from it, send an email to gmcmh-efi+...@googlegroups.com.

[Keith Vitko]

Matt,

Thank you for the follow up, informative and appreciated.

I have a follow up that is tangential to the discussion.

What would the effect be to engine knock and ability to advance timing if we were to add egr to our engines?

[Matt Colie]

The Porsche engines are a very special case.
Unless they have changed a lot since I last looked.  (This is very possible.)
They use a crankcase and individual cylinders with a single gang head.  This made the bridges their only hope as they did not have the processor capability to listen to two individual sensors. 

Matt Colie 

[James Hupy]

Bite your tongue.  (Grin)
Might as well throw air pumps in, too.
Jim Hupy
Salem, Or
78 GMC ROYALE 403

[Keith Vitko]

Lol, I really think egr is one of the biggest reasons today's motors have HP and the fuel economy they do.

It's allowed the compression to go back up (with computer control) which directly effects efficiency.

The only reason I see egr not effecting us positively is the 455 compression on our motors is relatively low.

[Keith Vitko]

Just read last message and could come across rude.

I only laughed and defended it because I think I'm the only one who thinks egr is good.

[Matt Colie]

Whoa Kieth,
You rude or me rude?  If you, I didn't notice - If me rude, I in explaining mode here and a lot of civility get blown off.  I have not intention of offending, but I often leave out things that should get explained to people that are dyno-lab refugees.

(No Jim, I will steadfastly refuse to get into the exhaust air injection discussion again.  That is buried and should stay so.)

Are you of the impression that EGR does anything good for an engine? 
Lets look at what happens:
You add hot almost inert gas after the air throttle.  This increases the pre-compression charge temperature.  This is exactly something you don't need.  What is does do is lower the adiabatic flame temperature in the cylinder so that there is less potential to form NOx.  This does help CAFE a little because the engine is now running at a higher MAP and that decreases the pumping loss.  
That can do nothing for horsepower.

In Simple Terms, Air Equals Horsepower.  (This part of the explanation just ended.) 
You are very correct that CR is a big function in recip engine efficiency, but the actual dynamic CR also has a lot to do with and engine's octane requirement.  The dynamic CR can be vastly effected by valve timing.   There are many other factors in play here.  To get the CAFE numbers that are needed to sell cars, an manufacturer can use valve timing to actually turn down the displacement of an engine when it is in grocery mode.  As this is where the testing is done, everybody is happy. 

Now, you want go-fast mode?  You pull the valve durations big and wide and with a four (or more) valve engine, you can get a lot of air in the hole.  But next, think about the design of our poor intake manifolds - squashed down to fit under the floor.  You don't see anything like that in the last 30 years.  The guy that suggested it today would be walked out to the flag pole and shot.   Flow designs are now a full quantum leap from where they were then.  That alone is worth 20% in horsepower gain.  (I did that with a 2.3 turbo motor.)  It doesn't hurt CAFE at all. 
You/we are still somewhat limited by the octane of available fuels, but with the fast clean flow, the high charge motion in the cylinders and lots of other tricks (and I haven't even gotten to direct injection yet) there are a world of things that can happen that were literally science fiction two decades ago -  Let alone 4 Decades (The BB olds is a 50yo design.  Not Bad? Huh?)

Back to EGR.  A big problem with old hardware calibration is that once it was calibrated for EGR, shutting the EGR off cause a huge amount of grief.  I did do this on an number of vehicles that I owned and on one I had to put it back on because the road-load performance was so terrible and nothing I did fixed it. 

Lets get down to another thing that I still can't get my arms around.  The manufacturing processes of the current OE are now at an unbelievable state.  The dimensional controls are amazing.  The metallurgy is something else, and the control processors in one car could do a moon landing and still tune the radio. 

So, stand back and look out.  There are several new cars out there that I would love to own, and I would.  But, some other people thought I didn't really want the savings I had worked more than twenty years to make so it was "redistributed".      
 
Matt Colie

[Ken Henderson]

Bruce,

I agree with you.  Without Matt's empirical data to full substantiate it, I think every sharp sound anywhere in the engine telescopes to the remotest corner.  Some years ago I hooked a 'scope to a knock sensor and lightly tapped here, there, and yonder with a tack hammer with and without the engine running.  Distance from the KS didn't seem to matter much and my taps were distinctly visible and predominant.  What I did not try, but should have, was the output from an ESC.

The bracket question begs information about the bracket itself.  I don't remember whether I tested the hypothesis, but I figured the size of the front mounting plate, its location, and its immediate attachment to the perhaps dampening motor mount argued against it as a KS location.  Since my KS's thread were different than all those on the lower starboard side of the engine, I made a 1/2"x1"x1-1/2" steel bracket/thread adapter to go there.  WIth a threaded hole in one end for my KS and a bolt hole on the other to fit the block, that gadget's sturdy enough that it should act almost like an integral part of the block.  It's seemed to work well on the 455 and the 500. 

Ken H.

On Wed, Jun 1, 2016 at 4:41 PM, Bruce Hislop <br...@perthcomm.com> wrote:

My thinking is the sensor is pickup up the "sound" directly (mechanically) from the block and not the sound we hear through the air.  I would think a sensor mounted on a bracket would be more "damped" than directly the block.

Just my farmboy shade-tree or fix it in the field mechanic point of view.

Do...@DolphSantorine.com

Excuse me for not being my usual wordy and sporadically verbose self. This message is sent from my iPhone.

No trees were killed in the sending of this message and few long dead dinosaurs were involved. A large number of electrons were terribly inconvenienced.

On Jun 1, 2016, at 11:42 AM, Matt Colie <matt7...@gmail.com> wrote:

On 6/1/2016 11:28 AM, bc...@juno.com wrote:
What about the sensor; is it just a microphone, or does it have some
specific frequency filtering characteristics too?

Bruce Roe

Bruce,

In very simple terms, yes and yes, but the answer is item specific.
As usual, generalities are typically in error.
Internally they are much better described as both frequency and directionally specific accelerometers.   When we were tuning them, we had to do all manner of things to try to reduce the sensitivity to all the other crashing and banging that goes on in a running engine.   We fought and gave up try to get one vehicle to stop retarding the spark for railroad tracks.

Matt Colie

[Gordon]

I always understood EGR was to reduce NOx so this comment through me for a loop. 

I am waiting for the 12V Hydrogen generator which makes the GMC into a 37 mpg cloud racer to come back? 

Thanks Matt

[Gordon]

Ken,

Where is your bolt in the block?  What does it secure?

Thanks,

[Ken Henderson]

Gordon,

I'll have to go look to remember just where the KS is bolted -- and that only on the 500.  IIRC, there was an un-used, 3/8-16 hole available.  It may even be where I removed the brackets for the now rerouted oil cooler lines.

Ken H.

[Johnny Bridges]

It isn't rude, but egr is now passe.  As are air pumps.  My Kia has neither, and develops 35% more power than the same size engine I had in a Datsun some years back (which had an air pump, but the requirement wasn't so stringent as to require egr).  Direct fuel injection, adjustable cam timing, and accurate engine control systems keep peak flame temps down to the point egr isn't needed.  And gets better fuel mileage in a heavier car in the bargain.

--johnny

---

From: Keith Vitko <keith...@gmail.com>
To: gmcm...@googlegroups.com
Sent: Wednesday, June 1, 2016 8:04 PM

[Keith Vitko]

Matt, sorry about late reply but in class. iwas worried I was rude not you.

I appreciate the opportunity to learn from all of you.

As for egr, agreed it's raising temp but it's also lowering vacuum pumping loss at high vacuum (egr generally only on at cruise not under load or idle) and increases charge density improving energy transfer to the cylinder upon explosion of the cylinder charge.

One of the shorter articles on this is http://www.tuneruniversity.com/blog/2012/05/dont-block-or-remove-the-egr-valve-its-saving-you-money/

[Keith Vitko]

Johnny, I believe your new engine effectively does what an egr did by adjusting valve timing.

Still have exhaust gas entering on intake stroke just able to remove one component and save the cost by adjusting valve overlap

[Matt Colie]

Kieth,

The referenced page is interesting, and it is close to right, but the
thing that the writer missed is that by lower the adiabatic flame
temperature of the combustion event, even with the greater mass in the
cylinder, the resultant effect is to lower the BMEP and have to more
mass to carry the energy out with the exhaust. This usually does not out
weight the power not used in the pumping loss, so it can be a net gain*.
*Except for the damage that the combustion by-products do going into the
engine - Like the carbon on the upstream sides of the valves.

Matt Colie

[Johnny Bridges]

Well sure it does.  My point is, it makes an expensive and failure prone part (egr valve) passe.  Once the manufacturer has gone to the expense of phased cams, the ability to use the technology to cut peak flame temp and thereby cut NOx emissions is a freebie added to the need for performance and fewer emissions the cam timing change gets.

And of interest, the only check light code I've seen on the Mouse has been 'knock sensor malfunction'.  I think it's either a poor connection or a marginal sensor which sometimes fails startup check. n Happens very infrequently and goes out by itself if I don't reset it, so I leave it live.  The little engine seems to be running right on the ragged edge to make poower when you spin it up.

--johnny

---

From: Keith Vitko <keith...@gmail.com>

To: gmcm...@googlegroups.com
Sent: Wednesday, June 1, 2016 10:09 PM