kream-

what are some ways to doing this?
what are the advantages/disadvantages to this?



------------------
17 from Wichita, KS (previously newbieZ28)
1988 Iroc-Z ZZ4 tpi, ttops, 50k miles on chassis (edelbrock tes, stock tpi) free mods etc coming soon after first 1/4 times!
1985 Z28 LB9 option [STOLEN!] Click here to see

http://thirdgen.bawx.org
K.I.S.S - Keep It Simple, Stupid.

JAY87GTA

The only way I've heard of doing it was to run a resistor in the wire, which I heard makes the computer throw in more timing sooner. But doing this risks detonation, Which you probably know can cause all sorts of bad things to happen.

D_Amlee

Hi, kream-... good to see you back here again.

I've never done this for myself, but I've heard that you can desensitize the sensor to some extent by removing it, wrapping a few turns of teflon tape around it, and snugging it back into the block with just enough torque to ensure that it stays put.

Advantages: IF (and only if) your sensor is receiving false knock signals from header noise, gear drive noise, or other "normal" engine sounds, then it would be telling the ECM to pull out ignition advance when your motor really didn't require it. Those sensors are tuned pretty closely to only go off when true knocking or detonation is happening, so having one go off from some other reason is rare, but it could happen, I suppose. Any time the knock sensor pulls out timing you lose power. The engine does this in order to eliminate detonation if it is detected, as an effort to save the pistons.

Disadvantage: if the knock sensor really was detecting knock and you desensitize it, you may be leaving yourself wide open to potential piston damage due to cylinder detonation.

Pablo

iTrader: (1)

Dale is right

first of all you ahve to determine whether or not your car is getting false knock, which is accomplished with a scantool and playing with the timing some.
If infact you do find false knock first try finding whats causing it.. could be as simple as a loose bracket.
If you cant find a cause, you can try wrapping it in teflon tape, but ive heard this doesnt have much of an effect, you can also purchase a brass adapter fitting, with the same male threading as the knock sensor, and a female threading as the knock sensor hole.. this brings it out away from the block some and cushions it from some of the harmonics the block is producing
theres also plans on the web for desensitizing it using a pair of resistors and some simple wiring but i lost my bookmarks list that had it.
in the meantime, dont worry about messing with it

[This message has been edited by Pablo (edited September 10, 2000).]

kream-

k, thanks for all the quick replies
i think ill let this go for now..
the reason i was wondering is because i have a strange noise coming from my headers (no not the leak either )

but for now.. better safe than sorry

WECoyote

my $.02 worth......


I am working on a system that will allow one to isolate the knock sensor reading(counts) the ECM sees. So far from my research I have found that the ECM looks for a 6 volt + signal from the ESC (electronic spark control) module. The knock sensor is connected to the ESC. If the voltage drops below 6 volts from the ESC the ECM 'deciphers' this as engine detonation aka knock and hence retards the timing. Not sure how the 'counts' are determined. If you ever have used a diagnostic scanner you will notice that it reads the knock sensor in counts.

I have to agree with Jay87GTA and D_Amlee....it's a risky thing to mess with!



I hope to have a working system for testing by the end of this month.

Mike

[This message has been edited by WECoyote (edited September 10, 2000).]

Pablo

iTrader: (1)

ah good work Wecoyote...
using the same info it could be fairly simple to make a knock detector i suppose.
Does the severity of the knock increase the amount of voltage subtracted from the magical 6? like slight knock would be 5.1 volts and severe would be 2.3 or whatever.. you could probably make an LED bargraph knock indicator.. even audible
neat
do you have any links to this info? how it works i mean.
i remember some gn site had plans to make one but it since has dissapeared

WECoyote

Pablo,
What I have found comes from my 1986 Camaro GM service manual. I cant say for sure if this same method of 'reading' the knock sensor applies for other 3rd gen f-bodie years but I would harbor a guess that it does.

Not sure if the amount of voltage drop below 6 is proportional to the knock severity. It could be. This is the next hurdle for me to get over! LOL! I am going to guess at this point and say the ECM takes a 'snapshot' of the ESC voltage signal every few milli-seconds and if it continues to see the low voltage adds it to the count 'bin' and then checks again and again and again and........its a an endless loop. When the voltage is above the 6 volt threshold for a certain period of time then the counts are reset to zero. Thats my hypothesis on that at this point....subject to change.

Mike

Pablo

iTrader: (1)

hehe should be the same up till the 90-92 tpi cars which have the est circuit on the memcal

redmet91z

FYI, there are plans on gnttype.org on how to build your own knock sensor display (if you don't have a scantool)It is suppose to beep according to how much knock the knock sensor is seeing. All the parts are only about $12 and are readily available at Radioshack.

------------------
1991 5.7L Z28 - gutted air boxes, K&Ns, airfoil, ported plenum, relocated MAT, accel TPI base, accel 300+ ignition, accel cap&rotor, homemade AFPR, derale fan switch, vette servo ,flowmaster exhaust, torque conv. lock-up switch, and fresh 255/50/16 Dunlop sp800s. More to come!

TRAXION

I have done two things...

1) Wrap the sensor threads with a few turns of teflon tape
2) Run a KS bypass switch. The KS bypass switch is easy to install. You need to wire a resistor in line with the KS signal so that the ECM won't throw a code. Easy to do. Heck - sounds like another tech article to write.

Tim

------------------
TRAXION's 1990 IROC-Z
Best Time = 12.587 @ 107.97mph (1.710 60ftr) on the long runner setup.
All Natural. No Force. No Drugs. Stock Bottom End. Stock Body Panels. Stock Wheels. No Weight Reduction.
Gunning for NA 11's with the MiniRam II and Hooker LT Headers.
-=ICON Motorsports=-

Blade

Actually, just wiring a switch in series with it won't do crap but retard timing even more. The wire fromt the ECM has to be grounded trough a 3900 ohm resistor in order to properly "fool" the ECM. If you just disconect it or just ground it, you get a constant reading of 6 degrees of retard. That's what my Diacom shows.

P J Moran

Here's a link to a procedure for just this. I don't know if it works cuz I haven't tried it yet, but give it a try if you want.
http://www.charm.net/~mchaney/knock.gif



------------------
'87 IROC 5.7l TPI - original owner!

TRAXION

Blade,

Exactly. That is the procedure that I followed to bypass my KS. Soryy - I should have been more clear ... its not really a bypass ... its more like you are 'fooling' the KS into thinking that you are reading 0º of retard.

------------------
TRAXION's 1990 IROC-Z
Best Time = 12.587 @ 107.97mph (1.710 60ftr) on the long runner setup.
All Natural. No Force. No Drugs. Stock Bottom End. Stock Body Panels. Stock Wheels. No Weight Reduction.
Gunning for NA 11's with the MiniRam II and Hooker LT Headers.
-=ICON Motorsports=-

MikeH

iTrader: (10)

Yes...I used the 3900 resistor on mine I was picking up quite a bit of false knock. I get full timing and no codes!

------------------
Black 90 IROC, L98, A4, 323 gear, Edelbrock TES, 3" Dynomax catback exhaust, Performance Resource chip, Accell LT runners, MSD6A ign, ported plenum, cut airbox, Airfoil, K/Ns, AFPR, AGR 12.1 box, soon to have 406 sb.

sparks383iroc

If you use the teflon tape route, test with different amounts of tape (ie. try one layer and check results, then 2, etc.). I've tried up to 3 layers. With three layers, I went to the track and set an engine code (forget the #- but it's the one for lost of contact with the knock sensor). I'm now running 2 layers without any codes being set.

Sparks a flyin'.

Tuned Performance

iTrader: (92)