Ok just unbolt the knock sensor from the car.

Or just unplug it.

ok. what will this accomplish?

the error it is getting is that the knock sensor has weak or no signal...

this is kicking it into safe mode, thus the cycling valve is just open, boost to only 1.2 bar.

alex is being sarcastic, referring back to the 'remove anything that gives an error' method. :)

The knock sensor does not do anything special ....... It is like a microphone. It sends a signal to the KLR. Unfortunately we do not know how the KLR interprets the signal. No one has been able to decipher the code. It probably just compares the waveform to a preset form.

SO lets give it a cleaner signal.... You can check the resistance on the knock sensor ~300 Kohm as I recall.
... You could stick a resistor into the plug......

If you unplug it I am pretty sure the signal is not sent to the KLR either.....

check this out:

the shop manual describes the knock sensor as have a 'piezoelectric crystal'. so looked it up in the dictionary.

pi·e·zo·e·lec·tric·i·ty Audio pronunciation of "piezoelectric" ( P ) Pronunciation Key (p-z--lk-trs-t, -lk-, p-z-)
n.

The generation of electricity or of electric polarity in dielectric crystals subjected to mechanical stress, or the generation of stress in such crystals subjected to an applied voltage.

i bet there is nothing to decipher. the knock sensor simply creates current all the time due to the vibration of the engine. thus, if there is not enough vibration, because the sensor is loose, it doesn't produce enough signal. if the engine starts to knock, it suddenly produces a lot of current and KLR starts shutting things down.

Sure..... exactly what a microphone does.

So a specific frequency is generated by knock then the KLR compares it to a preset frequency and it does/doesn't do X .

So once that frequency is hit....who knows what algorythm it uses....but it sends the info to the dme that decides to retard timing 3 degrees. Its a little more involved but thats it in a nut shell.

Are we getting closer now.

i think so. its possible the sensor isn't tightened down all the way, so it isn't producing the amount of current the KLR expects, then when the block and all gets hot everything gets tighter. or, the sensor is just bad. or it could be a connection somewhere. i've removed and cleaned the connection at the sensor and the KLR plug. i'll keep looking and i'll probably have to get in there and re-torque the sensor bolt.

hey nynor, you ever figure this out?

yeah, i did sort this out. the wiring to the knock sensor was damaged (read: shorting out). i had to splice in a new connector and wiring and its been fine ever since.

excellent! so you've kept the stock cv?

yes. i did keep the stock cycling valve. car runs like a top!!!

I've been reading this. Long story shortened, I'd like to "drill out" my cycling valve so that I can run a boost controller yet still maintain full knock control.

To do this, I need to know "which" port to drill out (front one right?), & how far, & how to test the CV both before & after: plus, should I pressurize & re-seal as I read in another thread??

Suggestions on testing the CV, besides putting power to it & hearing a "click"?

Mark.

'88TS, 3" full exhaust going on, then Vitesse chips & EBC set to 18psi on 7-heat range NGK plugs.

Mark

how are you going to run the CV and boost controller while maintaining knock control? putting any type of boost control in place usually over-rides any control the CV might have.

You have a waste gate don't you. Isn't the cycling valve a 3 way valve that opens the port to the waste gate at certain boost & RPM?

correct. how would you add boost control to that while maintaining the correct operation of the CV?

what about putting a different spring in the wastegate? don't flame me! i really don't have a good understanding of how the whole boost thing is controlled.

assuming you didn't change anything else, that would effectively increase the boost level, and also increase the 'cripple mode' level, over-riding the stock CV protection.

i am not sure what spring is stock, but let's say the spring was simply "upped" 0.2 BAR. i am pretty sure that wouldn't do horrible things. when my knock sensor wiring was faulty, it would go into limp mode and barely reach 1 BAR. the car was pretty much undrivable.

if the wastegate spring was upped .2bar (or ~3psi), it would increase the stock boost from .75bar (or ~11psi) to .95bar (or ~14psi).

if the CV were to detect knock and attempt to pull boost, it would only be able to pull boost to .4bar (or ~6psi) instead of the factory set limp mode of .2bar (or ~3psi).

with engine knock, there will probably be way more damage at 6psi of boost than at 3psi.

my original point, was that there's no way to put a boost controller inline with the CV while still maintaining the factory 2-stage boost cut knock protection of the CV (note that 6psi of boost does not really afford a lot of protection).

the only way i see this as a possibility is to replace the DME+KLR altogether and use a standalone EMS with boost-cut knock control failsafe.

nize, you are correct, of course. i was just thinking that .2 BAR is not huge. also, the boost has a range, correct me if i am wrong, and the stock boost maxes at about .75 BAR, but can be as low as .65 bar.