Welcome to the North American Subaru Impreza Owners Club Wednesday October 22, 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 > Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo)

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 09-05-2003, 04:38 PM   #1
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default Jorge's Daily hash: Knock sensor, false knock, filters and electronics...

OK... Been doing tons of research lately on false knock issues with people with noisier internals. I remember going over an article that Austin discussed quite some time ago in regards to knock sensor solutions for his TEC.

I've also been scouring the internet looking at the woes others have had in various markets, from VW's, to DSM's, to F-bodies. One solution that seems to work well is redefineing what the ECU sees.

Raising the "floor", as it's been refered to, serves to pronounce a relatively "dead" area to the ECU until a threshold (caused by detonation) is reached.

http://www.zzperformance.com/zzp/inf.../kr_graphs.htm <-- that link shows a good example on what I'm talking about. If there was a way of reprogramming the ecu (which I didn't see in the ecutek software), or filtering out voltage under a certain, user defineable point, you could in effect "cancel out" background noise, until that sharp rise occurs, thus signaling the ECU to pull timing.

Possible things I've considered.

A: modifying my knock link since it has and adjustable "gain", only thing there is finding someway to modify the knocklink to properly output again in a means that the stock ECU can accept and acknowledge.

B: relocating the stock knock sensor. I typically scan the different regions of my engine with a mechanics stethascope listening for things, and know that there are areas of the motor that are "dead" by comparison to other areas. The piston valley, for example, is a NOISY place, where as the intake manifold is this muffled purr in the background. Even the heads themselves are a relatively quiet affair. Question is this... Is it possible on to desensatize the background noise via distance? Have others tried something similar yet?

C: Sourcing a very specific knock sensor to the 92.5MM/79MM bore/stroke I have in my motor. I know most of these sensors are key'd to a certain frequency (I.E. 6900Hz), but I wouldn't have a clue trying to figure out how to find that frequency in my internals case which is knock, vs. what is slap/rock.

D: Building a "bandpass" type circuit to allow an adjustable filter under a certain knock sensor voltage.

EDIT: what I'm thinking here is a simple circuit that may feed .2V or something menial or even a percentage of the true signal to the ECU until a limit is hit. Then once that threshold is crossed, you get full activity to the ECU... this should be a relatively "simple" type of circuit.. so any electrician/EE type folk.. please chime in

Testing this is a relatively harmless affair. I have a tuned UTEC map that is knock free... but the ECU at those timing levels wants to run almost no timing at all. I can safely move the sensor around, and see if the reflash is adding in timing to match my tuned curve.... and conversely, see if it's exceeding what I know is max timing, cause it's dead to the "knock" world.

Jorge (RiftsWRX)
www.ProjectWRX.com
* Registered users of the site do not see these ads.

Last edited by RiftsWRX; 09-05-2003 at 05:02 PM.
RiftsWRX is offline   Reply With Quote
Old 09-05-2003, 04:49 PM   #2
Jaxx
NASIOC Supporter
 
Member#: 177
Join Date: Aug 1999
Chapter/Region: NWIC
Location: Boise,Idaho,USA
Vehicle:
The 93 Imp W/EJ20K
flat black

Default

have you seen where the tec kit mounts its knock sensor?
it replaces one of the engine mount bolts
Jaxx is offline   Reply With Quote
Old 09-05-2003, 04:56 PM   #3
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default

Quote:
Originally posted by Jaxx
have you seen where the tec kit mounts its knock sensor?
it replaces one of the engine mount bolts
Interesting.. I know that the TEC is a range from 0-100, with 30 being an average... just out of curiosity, has anyone tried running a TEC with a sensor in the stock location to see what that 30 would translate to up there?

That's good to know though! If there is a difference (as I'd hope), that can help found the theory of sensor relocation as a solution.

Jorge (RiftsWRX)
www.ProjectWRX.com
RiftsWRX is offline   Reply With Quote
Old 09-06-2003, 04:07 PM   #4
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default

UPDATE:

So I moved the knock sensors and noticed a couple of interesting things...

First... The stock sensor was attached back onto the block at around 60 lb/ft ... That's what happens when an electric impact is used....

This is relavent because your supposed to have around 8-10 lb/ft on that sensor.

I moved it and my knock link from the piston valley to the #4 and #2 runner on the intake manifold. Using Blue Loctite, I set them both at 6 lb/ft... Then, cranked the sensativities on the knocklink back to max (I had it turned down to around 1/4 sensativity. Now, on my tuned 93 map it only highlights the 1st yellow on rare occasion (it would peg red before....)

Also... Noticed the computer was much more receptive in this location to adding timing on non UTEC controlled areas.

I've actually setup an 87 octane map for my trip to the shootout, and ironically.. it's ecutek controlled. I have the UTEC limiting boost to 13 PSI, and then the ECUTEK is controlling fuel and timing.

At 11.2:1 AFR's and 13.8PSI at 6800 RPM the computer doles out 22 degrees of advance, and as far as I can tell... it's still learning.

That is with Water Injection of course.

The car feels slow from any point of view, but after hopping in my wife's stock WRX, I know I still have a significant edge hehe.. I wouldn't be surprised if I'm around 230-240 WHP. Not too shabby for 87 octane.

The biggest test is time... Once I go through the rest of this 87 I'll see how 93 on ecutek control works out.

Jorge (RiftsWRX)
www.ProjectWRX.com
RiftsWRX is offline   Reply With Quote
Old 09-06-2003, 05:04 PM   #5
8Complex

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

Default

There's no way that the knock sensor was tightened to 60ft.lbs., more then likely it was put in at about 15-20 and heat and time froze it in there.

Regarding the use of loctite on a knock sensor... why? You're just going to desensitize it by putting something between it and the part you're fastening it to.

If your heads are a dead zone, by all means, put one knock sensor in each head and if I'm not mistaken, you should be able to hook the signals together and put them into the ECU. The only concern there would be that possibly you'd need a resistor in-line to make sure the signals are cut down enough as to not overload the ECU. I don't know enough about electronics to really say for sure there.

If you're looking for a tuned knock sensor, turn to GM and find a motor with your equal bore and trace to it's knock sensor. That is what the RS guys do, and I believe they traced the bore back to an LS-1 block.
8Complex is offline   Reply With Quote
Old 09-06-2003, 06:38 PM   #6
blackie
Scooby Specialist
 
Member#: 42979
Join Date: Sep 2003
Location: houston tx
Vehicle:
03 wrx
blizzzack stage IV

Default

kinda like an adjustable knock sensor that can be tuned to respond to certian frequincies only caused by detonation, brilliant idea! i know on some turbo cars just unpluging the knock sensor can make like 20hp atw's this would just be so much safer and you could aleways run your motor on the edged safley.
blackie is offline   Reply With Quote
Old 09-07-2003, 05:14 AM   #7
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default

Quote:
Originally posted by 8Complex
There's no way that the knock sensor was tightened to 60ft.lbs., more then likely it was put in at about 15-20 and heat and time froze it in there.

Regarding the use of loctite on a knock sensor... why? You're just going to desensitize it by putting something between it and the part you're fastening it to.

If your heads are a dead zone, by all means, put one knock sensor in each head and if I'm not mistaken, you should be able to hook the signals together and put them into the ECU. The only concern there would be that possibly you'd need a resistor in-line to make sure the signals are cut down enough as to not overload the ECU. I don't know enough about electronics to really say for sure there.

If you're looking for a tuned knock sensor, turn to GM and find a motor with your equal bore and trace to it's knock sensor. That is what the RS guys do, and I believe they traced the bore back to an LS-1 block.
Loctite is to keep the bolt from backing out when it's torqued down to almost nothing

Using the knock link as a guinnea pig, I was trying different materials between the sensor and the manifold to "muffle" the vibration and noise that the sensor reads. I think I've found something that's pretty darn close.

A: in the stock location the knocklink would peg RED anytime after 24 degrees of timing on 93.

B: Moved to the manifold it was yellows..

C: Using two rubber isolators with washers on each side of the sensor it's now the second green worth of back ground noise. Purposely over timing the map produced a quick and sharp red.

Did the same thing on the factory sensor and now have only had to desensatize the UTEC from it's stock factory settings 5 points.

------

Now, going to bore... I went from 92 to 92.5MM bore.. is .5MM going to throw off the readings of the sensor that bad?

Jorge (RiftsWRX)
www.ProjectWRX.com
RiftsWRX is offline   Reply With Quote
Old 09-07-2003, 11:43 AM   #8
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default

Here's an interesting article on how to desensatize the knock sensor electronically.

http://www.anythingcars.com/members/chane/ks.htm

Quote:
The LM7 contains two 100kOhm knock sensors. In order to do this correctly, you need the series resistor to be 25kOhm(in pic, white) and the parallel resistor to be 500kOhm(in pic, blue). Form the two resistors into a Y shape as can be seen on the left. You may have to use multiple resistors to gain the appropriate resistance as I did. I then used heat shrink tubing on the multiple resistors as can be seen. Note, the two resistors are different lengths because I ran out of resistors and had to get them close, but not exact. They do produce different overall resistances, but once installed it is only a 0.03% difference. This is no big deal. (This mod will produce 25% of desensitiztion.)

Jon in CT, or any other electrical savvy folk. What is the range of an EJ20 knock sensor, and what do you think we'd need based on this article to desensatize at 10%, 20%, and 30%?

The original LT1 owner who did the mod came up with the software at this link http://www.charm.net/~mchaney/knockit.zip <-- If someone can come up with the resistance equations to the software, it can produce the resistors needed.

Thanks!
Jorge (RiftsWRX)
www.ProjectWRX.com

Last edited by RiftsWRX; 09-07-2003 at 11:50 AM.
RiftsWRX is offline   Reply With Quote
Old 09-07-2003, 02:06 PM   #9
8Complex

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

Default

Well, it used to be a large problem with the MY98-99's... they'd get hesitation in the midrange due to an overactive knock sensor. Most people ended up putting a few rubber washers between the knock sensor and the block and it helped out quite a bit.

I'm still concerned that you're going to be getting detonation that you can't hear over your block noises, though. Yes I do the same thing during tuning, but I don't have your normal ambient noise levels and my block is easily replaceable.
8Complex is offline   Reply With Quote
Old 09-07-2003, 04:08 PM   #10
turboICE
Scooby Specialist
 
Member#: 33979
Join Date: Mar 2003
Chapter/Region: Tri-State
Location: N.NJ
Vehicle:
2006 STi,'11 OBS
WRB & Black/Silver

Default

Xede has a map that will change the impedence to the ECU from the knock sensor. I don't have sufficient understanding of knock sensor parameters to know what it is intended to communicate in electrical signals to the ECU (how sound frequency is determined or communicated), but I do understand that the intent of the Xede function is to reduce the ECU sensitivity to possible false signals.

More information from Vishnu or ChipTorque on this would be desirable and possibly contribute to your efforts and the resulting shared knowledge.
turboICE is offline   Reply With Quote
Old 09-07-2003, 04:12 PM   #11
8Complex

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

Default

You know, just a thought, but maybe you should give John at J&S a ring this week and ask him what he thinks about it. He's got lots of experience with all types of cars and he's pretty much the knock filter king.
8Complex is offline   Reply With Quote
Old 09-07-2003, 04:13 PM   #12
I am mattjk
*** Banned ***
 
Member#: 36411
Join Date: May 2003
Location: I have my old screen name back
Vehicle:
I can be PM'd at
"mattjk"

Default

I had the same problem with the 93 mr2 turbo, over sensative knock sensor. I put a dab of silicone sealant on the inside of the sensor. Worked like a charm.

You know, using loc-tite is going to act like a dampener as well.
I am mattjk is offline   Reply With Quote
Old 09-07-2003, 07:39 PM   #13
Need4Boost
Scooby Specialist
 
Member#: 15194
Join Date: Feb 2002
Chapter/Region: Tri-State
Location: 1^2 Mile, NJ
Vehicle:
2004 Lightning,
2001 BMW 330ci

Default

Quote:
Originally posted by RiftsWRX
UPDATE:

S

I moved it and my knock link from the piston valley to the #4 and #2 runner on the intake manifold. Using Blue Loctite, I set them both at 6 lb/ft... Then, cranked the sensativities on the knocklink back to max (I had it turned down to around 1/4 sensativity. Now, on my tuned 93 map it only highlights the 1st yellow on rare occasion (it would peg red before....)

Also... Noticed the computer was much more receptive in this location to adding timing on non UTEC controlled areas.

Jorge (RiftsWRX)
www.ProjectWRX.com
So after you relocated the knock sensor, you mentioned the car was more receptive to adding timing since. Are you saying the ECUTek was able to make its adjustment and run more timing?

I would take it that there would be a few more hp to be squeezed out if that's the case.

Would you say this is a better location rather then the stock? Pics of where you mounted it? Sorry, this is new to me. I have never really played with knock sensors at all. Thanks.
Need4Boost is offline   Reply With Quote
Old 09-07-2003, 08:12 PM   #14
RiftsWRX
Scooby Guru
 
Member#: 6124
Join Date: Apr 2001
Chapter/Region: MWSOC
Location: Glendale Hts, IL, USA
Vehicle:
2000 NFR AP1 S2000
'07 Honda FIT sport (5MT)

Default

I suffer from two possible issues.

A: as timing increases (and subsequently cylinder pressure), background noise rises.

B: While the computer is clearly not hesitating at the 25 degrees of advance I run at 22 PSI with WI, once I switch to a 33 degree high octane map, it's a different story.

One thing visible is how the computer doles in timing. I have the knock link setup so you can see it's "attempts"

You get a solid green, then as your accelerating you'll see a sharp second green... back to one... maybe a sharp second and a flash of a yellow... back to one green... etc.

Looking at the logs you can see timing where it's adding a .4 here and there (it's .4, because that's what I programmed into the ECUTEK software when I flashed my ECU). But we're talking a level of resolution here in the 18-20 degree ranges when I was building my 87 octane map (87 +WI, 14PSI max for long distance trips).

So, my dillema is this... do I just leave it at the sensativities it's at now with the way I've "detuned" the sensor, try an electrical solution for the said scenario, or both...

Or... do I attenuate for my highest possible scenario.. problem there (as Nick stated earlier), do I run the possability of detuning so far that it won't hear the low range/low torque knock situations? Who knows...

I'll try the resistor issue next and see what happens.

The only thing I could imagine is that the ecu logic is smart enough to filter out background noise and notice SHARP increases in sensor activity. Somewhere I think the solution will be to find that point where I'm not completely detuned to the point where I can't really hear low end stuff, but enough to where I can run my high octane map without freaking out either. That's the magic answer I guess

Jorge (RiftsWRX)
www.ProjectWRX.com
RiftsWRX is offline   Reply With Quote
Old 09-07-2003, 10:23 PM   #15
thng
Scooby Specialist
 
Member#: 8062
Join Date: Jul 2001
Location: Moreno Valley, CA
Vehicle:
2002 WRX
Silver

Default

First and foremost, don't assume that the ecu is smart enough to determine what is and what isn't knock. We had a $50K ecu in the gtp car and for the life of me getting the knock sensor to work properly was a serious PITA.

You are in a bit of a dilemma. The further away you are from the sources of detonation the harder it is to hear. That just increases the possibility of detonation at high boost (as the ecu is trying to figure out if it is or it isn't). All it takes is the one lucky shot and all the electrons will be in there high fiving themselves cause they got by you. That's how I felt sometimes.

All our traces showed that the ecu just didn't react fast enough in deciding whether to bail out or keep going. This we did by trashing a head by putting pressure sensors in all the chambers and running. That was a bad day... I'd like to tell you about it one day.

It just missed by fractions of a millisecond.

It's there (the knock sensor) but we ended up not using it and just mapping every point along the top of the curve and coming back 1 degree. That was with rocket fuel. I can imagine with pump gas it'll be worse. That stuff is so unstable it isn't funny. Anyway, the motors lived more than they died on the track except for endurance races where the pistons failed at the end or showed signs of emminent failure from fatigue and only the piston rings held them together. Durn but those were weird days.

Last edited by thng; 09-08-2003 at 08:32 PM.
thng is offline   Reply With Quote
Old 09-09-2003, 03:58 PM   #16
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

My underlying anxiety is missing det that is quieter than the background noise. I know little flashes to second green on maximum sensitivity on a standard engine on a knocklink can be the start of det on a standard engine, and they respond by the ECU pulling even 0.5 degrees. However, I got that much noise on my engine on fast idle.

The important question:

DOES KNOCK ON A NOISY ENGINE GO UP IN VOLUME ENOUGH TO BE HEARD OVER THE BACKGROUND NOISE?

Apparently the setup for the Safegaurd is to just desensitise it so it doesn't pick up piston slap. That discourages me slightly, but apparently det this quiet or short lived should not damage the engine. I guess there is one way to find out
john banks is offline   Reply With Quote
Old 09-10-2003, 02:06 AM   #17
thng
Scooby Specialist
 
Member#: 8062
Join Date: Jul 2001
Location: Moreno Valley, CA
Vehicle:
2002 WRX
Silver

Default

Knock noise didn't go up and I don't believe it will. The occurrence of knock takes place in progressively shorter and shorter time frame with the increase of power, which ever way it is created. Therefore the knock sensing has to be more and more accurate. This becomes increasingly more difficult to do.

You can make it more "accurate" if your motor is a truck engine designed to produce maximum power at mid-range rpm. If your motor is being designed to produce power using boost and rpm (or just plain boost). The odds are against you but good luck with your efforts.

Back ground noise increases with speed and power. More power means parts need more mass; springs need more strength which adds to more mass and still more power and so on.

The word anxiety works well in this situation.
thng is offline   Reply With Quote
Old 09-10-2003, 09:30 AM   #18
tcs007
Scooby Specialist
 
Member#: 4103
Join Date: Feb 2001
Chapter/Region: MWSOC
Location: Menasha
Vehicle:
2003 WRX Wagon
Reddish

Default

Quote:
but I wouldn't have a clue trying to figure out how to find that frequency in my internals case which is knock, vs. what is slap/rock.

Hey Jorge.
I've got a spectrum analyzer here at work we could use to determine the frequency to within a 1/3 octave(31 bands from 20 hz to 20 khz). That's probably all the resolution you would need to determine that.

-Todd
tcs007 is offline   Reply With Quote
Old 09-10-2003, 10:05 AM   #19
Wheeler Bement
Scooby Specialist
 
Member#: 15652
Join Date: Feb 2002
Chapter/Region: TXIC
Location: sugar land, katy
Vehicle:
2002 2.5L WRX Wag.
160K club

Default

Quote:
Originally posted by tcs007



Hey Jorge.
I've got a spectrum analyzer here at work we could use to determine the frequency to within a 1/3 octave(31 bands from 20 hz to 20 khz). That's probably all the resolution you would need to determine that.

-Todd
i was about to suggest that. does anyone know exactly ho the knock sensor works. is it basically an accelerometer, when the voltage output goes to high, it thinks it detected knock. i know you can get accelrometers for different frequencys, but my point being is that you should be able to take the signal and use the spectrum analyzer, or if you have labview(like me) you can FFT it and see all the different frequencies and amplitudes....to find out what is normal
Wheeler Bement is offline   Reply With Quote
Old 09-10-2003, 10:39 AM   #20
john banks
Scooby Specialist
 
Member#: 11669
Join Date: Oct 2001
Location: UK
Vehicle:
Evo IX 345 WHP

Default

You can plug a separate knock sensor into a microphone input on your laptop, and put it through free spectral analysis software http://www.visualizationsoftware.com/gram/gramdl.html . However, J&S says that simple frequency filtering is useless for knock detection. The best systems use DSP and gating plus thresholds that are mapped by RPM/load.
john banks is offline   Reply With Quote
Old 09-10-2003, 10:49 AM   #21
turboICE
Scooby Specialist
 
Member#: 33979
Join Date: Mar 2003
Chapter/Region: Tri-State
Location: N.NJ
Vehicle:
2006 STi,'11 OBS
WRB & Black/Silver

Default

I have recently become interested in the knock sensor area as well and wish there was something specific on the WRX systems.

However Jon [in CT] did give me these that I found helpful and informative - at least for a knock newb like me:

Quote:
You should browse around this site: http://home.netcom.com/~bsundahl/kn...KnockSounds.htm and if you decide to record your knock sensor, then you can download the mentioned freeware Spectrogram PC software from: ftp://ftp.agt.net/pub/winsite/win95/sounds/gram50.zip
The DIY that he used to a cassette recorder should work equally as well on my laptop mic - I will probably make one circuit to both connect to my laptop mic and a set of earphones. $20 bucks for det cans and recorder not bad in my view.
turboICE 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
Jorge's Daily hash: Solenoids, ABC's, Boost Control, and Concept RiftsWRX Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) 38 07-28-2003 10:11 PM
Jorge's Daily hash: Cooling RiftsWRX Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) 98 06-26-2003 02:42 PM
Jorge's Daily hash: Intakes, Plenums, and Power RiftsWRX Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) 41 03-31-2003 05:58 PM
Jorge's Daily hash: Bearings, Buildups, and oil RiftsWRX Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) 88 03-19-2003 06:28 PM
Jorge's Daily hash: Senders and electricity RiftsWRX Factory 2.0L Turbo Powertrain (EJ Series Factory 2.0L Turbo) 26 08-07-2002 11:43 PM


All times are GMT -4. The time now is 12:34 PM.


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.