Welcome to the North American Subaru Impreza Owners Club Tuesday July 29, 2014
Home Forums WikiNASIOC Products Store Modifications Upgrade Garage
NASIOC
Here you can view your subscribed threads, work with private messages and edit your profile and preferences Home Registration is free! Visit the NASIOC Store NASIOC Rules Search Find other members Frequently Asked Questions Calendar Archive NASIOC Upgrade Garage Logout
Go Back   NASIOC > NASIOC Technical > Engine Management & Tuning

Welcome to NASIOC - The world's largest online community for Subaru enthusiasts!
Welcome to the NASIOC.com Subaru forum.

You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our community, free of charge, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is free, fast and simple, so please join our community today!

If you have any problems with the registration process or your account login, please contact us.
* Registered users of the site do not see these ads.
Reply
 
Thread Tools Display Modes
Old 08-06-2003, 12:24 PM   #1
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default Those of you with high clearance forged pistons - engine noise vs knock detection

Some rebuilt engines with forged pistons basically do not work with factory ECU knock correction, with a device like a knocklink or an engine noise level you could have detonation that is quieter than the background engine noise?

All the cars I have remapped have been OEM engines, quiet running. Knocklink flashing just the warning oranges can be mild detonation, on a noisier engine you can have far more than this through background noise, so would not see the spikes from detonation.

Any information appreciated... will be using headphones with det cans and/or listening to the knock sensor directly, but want in car det monitoring as I can't go around with headphones on

How well does the J&S safeguard cope with noisier 4 thou clearance forged piston engines?
* Registered users of the site do not see these ads.
john banks is offline   Reply With Quote
Old 08-06-2003, 02:58 PM   #2
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Don't be shy BTW just because I've mapped a fair few ECUs. I'm here to learn.
john banks is offline   Reply With Quote
Old 08-06-2003, 10:58 PM   #3
mr2guru
Scooby Specialist
 
Member#: 10334
Join Date: Sep 2001
Location: St. Louis
Vehicle:
91 RX7 LSx swap
'09 CC 3.6 4motion

Default

John,

PM Jorge....he ran into this exact problem with his 2.2 motor. But figured it out. He should have some good info.
mr2guru is offline   Reply With Quote
Old 08-07-2003, 02:15 AM   #4
JustDSM
Scooby Newbie
 
Member#: 17502
Join Date: Apr 2002
Location: Cheyenne, WY
Default

Something that would allow you to view the Raw knock signal would be helpful as it would allow you to see a pattern of knock activity if your pistion slap is bad enough. By looking with careful eyes I'm sure you could distinguish between piston slap and pinging.

Thoughs?
JustDSM is offline   Reply With Quote
Old 08-07-2003, 03:47 AM   #5
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

I can do that with a knocklink, and I could log its filtered output, but if the background noise is louder than the quieter detonation spikes I still have a problem?

I'll ask Jorge.
john banks is offline   Reply With Quote
Old 08-07-2003, 06:59 AM   #6
JustDSM
Scooby Newbie
 
Member#: 17502
Join Date: Apr 2002
Location: Cheyenne, WY
Default

It was late when I made that post, and it's later now.. But I should perhaps have been a little more clear.

Since you've got the ability to see the raw knock signal (can you log it vs. RPM as well?) you could establish a cycle of what would look like knock in constant with RPM. Any deviation or change in that pattern would indicate something from the norm, IE: knock. Thats's given the pistion slap is at a frequency close enough to that of an actual detonation event for the knock sensing hardware to detect it.

Again, thoughs welcome.

(P.S. Perhaps I'll re-read what I have written here after a few hours of sleep and clear up any confusion. )
JustDSM is offline   Reply With Quote
Old 08-07-2003, 07:11 AM   #7
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

I am worried about damaging my motor with knock that is QUIETER than the background noise level though when I am not listening with headphones...

Basically I've mapped all the engines I've done so far with knocklinks on maximum sensitivity.

On most cars this has correlated very well with knock correction up to the RPM where it switches off. The cars have made reasonable power and none of them have blown up.

On these cars anything above 2 greens on a knocklink has responsed to either less timing, more fuel or less boost or a combination of all. Even a flick to 2nd green in the midrange has the ECU pulling out timing.

With my first run of the stroker engine (which was noisy because of a cambelt tensioner problem) I could hit big red on the knocklink at 4000 RPM with no load. At 8 PSI boost it would hit big red about 3500 RPM. The ECU pulled out 8 degrees timing when there was no knock!

When I get the engine back in I need to correlate the knocklink with headphones to see if it is any use any more.

I will need to remove the factory knock sensor because it is useless, and run off a fixed map.

I need something in-car to tell me if I am detting when not wearing headphones, either that or a very conservative tune.

I could log the knocklink's output with a Labjack and Delta Dash, but I may as well just watch the knocklink which is in my line of sight.

I am happy to watch the difference between gradually rising noise and sudden flicks up, but if I have to run the knocklink on very low sensitivity it will not show up what could be knock.

I've read through all Jorge's reports to see how he got around this but can't find any information.

I've been looking at Intersil 9010/1 chips, boards, TMS 320C, J&S safeguard, there is nothing I can see that would fit the bit. All the standalones look quite primitive in knock control/correction too.

Bear in mind I am using the JECS ECU, which really only adds or subtracts a few degrees from zero and usually just runs a base map. But it will be useless since sensitivity is non-adjustable and retards -8 with hardly any boost with very conservative timing.

I feel I've kept engines together so far because of careful timing and continuous knock monitoring. I could lose my new engine in only weeks if I don't know if it is detting. I won't enjoy it is I am paranoid about it detting.
john banks is offline   Reply With Quote
Old 08-07-2003, 07:42 AM   #8
JustDSM
Scooby Newbie
 
Member#: 17502
Join Date: Apr 2002
Location: Cheyenne, WY
Default

Out of curiousity, what are your pistion to wall clearances?
JustDSM is offline   Reply With Quote
Old 08-07-2003, 07:44 AM   #9
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

4 thou on Wiseco as factory spec. My first engine was rattly because of a cambelt tensioner, others have had similar and they are slightly noisy from cold for about 30 seconds, but they still light up knocklinks very easily.
john banks is offline   Reply With Quote
Old 08-07-2003, 09:51 AM   #10
David_Wallis
Scooby Newbie
 
Member#: 22479
Join Date: Aug 2002
Location: Leeds - UK
Vehicle:
94 2.33 Garrett GT
Heads, Cams & AutronicSM2

Default

might be worth speaking to chris palmer.. also the new knocklink is digital.. .

Texas Instruments do a DSP based knock detection IC iirc.

basically a bandpass filter to filter the engine noise from knock may be needed.. I await a solution also.
David_Wallis is offline   Reply With Quote
Old 08-07-2003, 09:59 AM   #11
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

The 9010 or TMS320 are too complex for me to build anything with - I was studying the datasheets last night. There is nothing off the shelf I can find with these in.

Chris's circuit has a tighter filter but I think would still struggle.

J&S can be fooled too I gather.

There seems to be nothing for a good price that listens during a time and frequency window per cylinder with any accuracy. Pressure sensing plugs are expensive for all four and the equipment to read them. Ion sense is not readily available.

David, can you quantify for me what noise levels you get on a knocklink by RPM/load approximately?
john banks is offline   Reply With Quote
Old 08-07-2003, 10:01 AM   #12
David_Wallis
Scooby Newbie
 
Member#: 22479
Join Date: Aug 2002
Location: Leeds - UK
Vehicle:
94 2.33 Garrett GT
Heads, Cams & AutronicSM2

Default

off the top of my head no.. as I was trying to sleep in the car whilst mapping

Speak to mark or Pat..

David
David_Wallis is offline   Reply With Quote
Old 08-07-2003, 11:59 AM   #13
Jon [in CT]
*** Banned ***
 
Member#: 2992
Join Date: Nov 2000
Chapter/Region: NESIC
Location: Connecticut, USA
Vehicle:
02 WRX Sedan
Silver

Default

The Knocklink is little more than a noise meter and doesn't attempt to distinguish the particular noise that knock produces from other engine noises. Watching its lights will not help you distinguish knock from other sounds.

Here's a relatively simple and inexpensive approach you might want to consider. It relies on your ear/brain to detect knock in realtime and a spectrographic analysis of the recorded sound to detect barely audible knock during a post-run analysis, conducted in conjunction with the analysis of DeltaDash data.

The basic idea is to tap the stock knock sensor's signal and convert it to a signal that can be recorded via the MIC input of a cassette recorder or PC sound card and, while it's being recorded, you can monitor it via headphones connected to the recorder.

A simple circuit to accomlish the conversion is described at http://pw1.netcom.com/~bsundahl/knock/listen/listen.htm. The post-run analysis is described at http://pw1.netcom.com/~bsundahl/knoc...nockSounds.htm and relies on freeware called Gram V5.0, which can be downloaded via http://www.simtel.net/product.downlo...rs.php?id=9440.


Using the spectrographic analysis program, it's fairly easy to spot knock because its sound falls in pitch with respect to time. This produces a short slanted line on the spectrograph. This characteristic of knock's sound is decribed briefly at http://www.osu.edu/units/research/archive/knock.htm as:
Quote:
The researchers took advantage of the fact that the pitch of the sound (acoustic resonance) induced by knock changes with the combustion temperatures. As combustion progresses the temperature in the combustion chamber rises and then falls. Knock typically starts when combustion chamber temperature is near its peak. Thus, the pitch of the knocking sound becomes lower with time, and this gives rise to a characteristic vibration signature. The software tells the computer to look for vibrations with this characteristic drop in pitch, and ignore all the others.
When you identify knock via spectrographic analysis, you can edit the audio file to replay that section over and over to help train your ear to recognize it better in the future. It's also frequently possible to identify the RPM at which the knock occured because a cylinder typically knocks for several consecutive combustions and you can measure the time between combustions and compute what the RPM must have been.
Jon [in CT] is offline   Reply With Quote
Old 08-07-2003, 12:51 PM   #14
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Thanks Jon, I've previously recorded .wav files from my knock sensor , but never found the spectral analysis software so I'll try that.

The problem still remains though of an in car continuous monitoring setup that is commercially available.

I am happy that I'll be able to hear it whilst listening with headphones, it is when I don't have the headphones on..... either VERY conservative tune or listen frequently

Thanks David, already quizzed Mark, Bob and Merv. Still no one can suggest something in car, and all just say tune it with big margins. Think that is just what I'll have to do.
john banks is offline   Reply With Quote
Old 08-07-2003, 04:19 PM   #15
motoxphil
Scooby Newbie
 
Member#: 5094
Join Date: Mar 2001
Location: Gaithersburg, MD, usa
Vehicle:
2002
11.96, 1.64, 115mph!

Default

John Banks,

Jorge got around this because he was using the UTEC which does not use the same "knock" system as the factory ecu. The UTEC controls the timing independently and therefore if the factory ecu would normally pull timing due to piston slap it has no effect on the UTEC (since the knock system can be fine tuned).

He used a prototype knock system from us to differentiate between what is knock and what is piston slap, altering the UTEC’s knock settings at every 1000 rpm accordingly.

This product will be released to the public in a couple of months and will have a visual knock output along with a headset jack. You do not need a UTEC to use this.

Good luck.

Phil
TurboXS
motoxphil is offline   Reply With Quote
Old 08-07-2003, 04:41 PM   #16
ride5000
Scooby Guru
 
Member#: 32792
Join Date: Feb 2003
Chapter/Region: NESIC
Location: lincoln, ri
Vehicle:
2003 GGA MBP
12.9 / 105+

Default

that sounds good. i am anxious to hear more about it.

ken
ride5000 is offline   Reply With Quote
Old 08-07-2003, 04:57 PM   #17
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Phil that sounds superb. I would love to test this in the UK for you!

I am also working on a MAF to MAP conversion, and I could easily use a "knock level" input into my uC which would richen and retard the mixture by RPM and boost according to programmable thresholds (would just boost the simulated MAF voltage by a programmable percentage for every knock that went over threshold).

All I need is the box that has an output of a useful "knock level".

Phil, meantime, do you think I am onto a total loser checking for knocklink outputs above a set level by RPM to have a visual indicator of knock? Can you get SUB-background noise levels of knock that are damaging?
john banks is offline   Reply With Quote
Old 08-07-2003, 05:18 PM   #18
motoxphil
Scooby Newbie
 
Member#: 5094
Join Date: Mar 2001
Location: Gaithersburg, MD, usa
Vehicle:
2002
11.96, 1.64, 115mph!

Default

John Banks,

I haven't told you all this device will do on purpose but there will be a configurable 5 volt output with this device, i.e. knock, MAF, MAP, etc.

The Knocklink is better than nothing but the UTEC pulls timing when the very first green indicator flickers. The Knocklink is more of a noise threshold device vs. actually tuned into to the specific frequency of knock. I'm sure you've noticed this as the knocklink can ignore knock at low rpms and light up light up like a Christmas tree without knock (mine lights up over 7400 RPM with no knock heard via a listening device.)

Thanks,
Phil
TurboXS
motoxphil is offline   Reply With Quote
Old 08-07-2003, 05:24 PM   #19
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Thanks.

Meantime then I am thinking of a high order bandpass around approx 6KHz, with gain inversely proportional to det free background noise at 1000 RPM intervals. The knock window thing is too much for me to even attempt.

Please rush your product to the market ASAP!
john banks is offline   Reply With Quote
Old 08-08-2003, 03:15 AM   #20
John at J&S
Scooby Newbie
 
Member#: 2036
Join Date: Aug 2000
Location: Garden Grove, CA USA
Vehicle:
2000 Intrepid
silver

Default

John:

The J&S uses a knock window and a load qualifier. I haven't found frequency filters to be of much use.

Since piston slap is at its worst at low loads or decel, we don't listen until the vacuum drops below 5 inches.

Clearance of .004 shouldn't be a problem for the J&S.

The monitor jack outputs a voltage for the retard display bargraph, which can be logged. The voltage varies from 0v to 1.3v at max retard.

In individual cylinder mode, the voltage changes every spark event. In retard all mode, the voltage updates every two revolutions.
John at J&S is offline   Reply With Quote
Old 08-08-2003, 04:42 AM   #21
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Thanks John. Can you confirm that I need the 4 channel version for a MY00 UK Impreza turbo? On the J&S site it lists one/two channel for the Subaru.

My ignition system is very similar to the US WRX AFAIK.
john banks is offline   Reply With Quote
Old 08-08-2003, 05:14 AM   #22
8Complex

Moderator
 
Member#: 922
Join Date: Feb 2000
Chapter/Region: MWSOC
Location: Schaumburg, IL
Vehicle:
04 FXT
Red

Default

You'll need the version that suits your coil setup. 1 or 2 can use the old setup, the new setup is good for 3 or 4 coil setups.
8Complex is offline   Reply With Quote
Old 08-08-2003, 07:39 AM   #23
Jon [in CT]
*** Banned ***
 
Member#: 2992
Join Date: Nov 2000
Chapter/Region: NESIC
Location: Connecticut, USA
Vehicle:
02 WRX Sedan
Silver

Default

Quote:
Originally posted by john banks
The knock window thing is too much for me to even attempt.
John Banks, the knock window thing need not be "too much." You can take a simplistic approach, such as the one adopted by the J&S unit. It's described in http://www.eng-tips.com/gviewthread....d/71/qid/54276 as:
Quote:
The detector in the J&S unit is in software, running on a 68HC11. The unit controls dwell and timing as well.

Software sets up a knock window, since knock is expected to occur around TDC. The window opens 32° after the ignition trigger, and remains open for 44°.

A load qualifier is used to disable the detector under light engine loads and decel, where piston slap can cause a large interfering signal.

The unit uses a proprietary algorithm to develop a threshold, against which the signal is compared.
BTW, have you also read the application notes for the HIP9010? They're at www.intersil.com/data/an/AN9770.pdf and might make the HIP9010 a little less intimidating.
Jon [in CT] is offline   Reply With Quote
Old 08-08-2003, 10:22 AM   #24
John at J&S
Scooby Newbie
 
Member#: 2036
Join Date: Aug 2000
Location: Garden Grove, CA USA
Vehicle:
2000 Intrepid
silver

Default

Thank you so much. I designed my simplistic approach in '84.

The detector used a combination of bandpass filters, a knock window generator with phase lock loop, and a digital threshold generator.

The finished module measured 3.2" x 4.7", and included an ignition timing and dwell controller.

Dinan Engineering used these in their late 80's BMW turbo kits. They hid it inside a large metal can, and called it "TurboTronicsIV".

In '89, I converted the design to microprocessor control. This allowed me to add digital signal processing to the knock detector, individual cylinder knock retard, and adjustable soft touch rev limiting.

I showed it to the editor of Turbo Magazine, who told me that if MSD and HKS aren't making it, then it must be a product that's not needed.
John at J&S is offline   Reply With Quote
Old 08-08-2003, 11:19 AM   #25
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

Jon (in CT) yes I'v read the Harris datasheets and app notes, the board was available from one supplier with 26 week lead time and a high price. The 9010 is no longer made.

I am interested in the J&S and have searched all the old threads. The last I saw the 4 channel version was being tested on the WRX. John at J&S, is there something available that will work for my engine, and do I want the 4 channel version? I don't need boost retard, nitrous or midrange retard, I can map the base maps on the ECU to what I want.
john banks is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Those of you with Cusco Engine/Tranny Mounts pleaes come in r12rex Southern California Impreza Club Forum -- SCIC 5 03-20-2008 08:29 PM
For those of you with Forge BOV's, need help, what spring are you using? vipernj Factory 2.0L Turbo Powertrain 4 01-21-2006 11:55 PM
Those of you with 02 WRX did you have OBDII test this year in MA? MooseWRX New England Impreza Club Forum -- NESIC 9 05-04-2005 06:53 PM
for those of you with engine/tranny mounts quickscooby Transmission (AT/MT) & Driveline 40 11-16-2003 10:35 PM
those of you with a TXS TMIC, what BOV do you use? im having problems.. M32WRXin3seconds Factory 2.0L Turbo Powertrain 11 10-18-2003 08:38 PM


All times are GMT -4. The time now is 12:12 AM.


Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Powered by Searchlight © 2014 Axivo Inc.
Copyright ©1999 - 2014, North American Subaru Impreza Owners Club, Inc.