vette_tweak

Does anyone have a schematic diagram for the circuity inside an ESC?

I've read some info about the ESC/Sensor being designed for a specific application. I'm having knock count problems. My motor is not stock so I figure there's a good chance that the Esc/Sensor is no longer correctly tuned into detecting knock from my engine. I have an idea of what is in the module - a bandpass filter and an analog to digital converter. I should be able to change the frequency window of the filter. I'm going to find the window by hooking up a spectrum analyzer to the knock sensor and over advance the timing to find the spectral signature of detonation in my engine.

Back to the original requirement; schematic anyone? :cheers:

Hendej

Here are schematics of a generic KS conditioning circuit with adjustable sensitivity:
http://members.tripod.com/lyc_42/eke/eke.htm

Note that it is not compatible with the '165 computer since it works by delaying the coil signal directly. Might work independant of the computer, but I can't figure out how much timing it pulls or whether it varies with RPM (fixed delay = more retard at higher RPM).

What the '165 expects is a series of 1 ms pulses when KS is active. These pulses gate on a 16 bit counter which is read by the ECM every 80 ms to decide how much timing to pull and/or give back. The high byte of this counter is what is sent down the ALDL line to your scanner as the "knock counts". It would take some work to interface this circuit to the ECM. Would it be worth it? I dunno.

vette_tweak

Awesome! Thanks for the info Hendej! Those schematics are definitely a step in the right direction. Any info is better than none! :cheers:

72tccracer

Is it possible to just eliminate this circuit? Mine gives me trouble in the middle of a race every so often. the car nevers gets less than 100 octane
so I don't believe taking it out should hurt anything???? Thoughts.

Chuck http://www.tccracing.com
#72

vette_tweak

I ended up attaching a digital oscilloscope and spectrum analyzer to the knock sensor of both my vettes to sort out my knock problem. The noise/signal spectrum is quite broad. I expected a narrow band of noise/signal (n/s) during the knock condition but it was still fairly broad and centered around 4.6KHz.

My 89 doesn't have a knock problem however the n/s level is relatively low compared to my 88 (solid roller, large compression ratio). It didn't take a whole lot of additional n/s to setoff the knock sensor because the n/s level was already considerably high. The 88 knock signal simply looked like an amplified version of the 89 knock signal.

My approach to solving this problem was to attenuate the 88s knock signal so that it was closer to the 89s. All things being equal, the signal strength difference in dB , knock condition vs no knock condition, was still the same in both cars before signal conditioning.

I wired up a 50K potentiometer to the ESC harness. Pin 1 to the knock sensor; pin 2 to the ESC pin the knock sensor was wired to; pin 3 to ground. I then dialed in the attenuation to match the 89 and presto! :D end of problem.

Don't try this without a voltmeter or oscilloscope - you must monitor the signal at the ESC knock sensor pin to ensure you haven't disabled the system by turning the potentiometer the wrong way too far. :cheers:

72tccracer

Sounds good, what volt range are you looking at. My car has loud valve train along with header leaks

EricVonHa

Chuck, you can get someone to burn you a custom prom for your car that will not retard the timing if knock or "perceived" knock is heard. I've had similar circumstances occur with mine...

vette_tweak

I measured about 200mV peak at idle on my "normal" vette, about 500mV peak on my modified vette. During a engine knock, I measured 1.2V peak on the normal vette, 2.8V peak on the other. This is measured with a warm engine idling at 750 RPM. Keep in mind these #s are measured using an oscilloscope. A volt meter will give you an RMS reading that will be close to 0.7 times my numbers. I used the scope because volt meters are calibrated at a certain frequency (usually 1KHz) and I wanted as much accuracy as possible.

Eric is right on - on my stock ECM, I limited the retard to 2 deg and everything works great. My problem is that I'm using Accel DFI #15013 w/ software V6.32 and I can't find a way of defeating the retard (up to 20 degs!).

:cheers:

72tccracer

I am looking into burning my own chips and that is on of the things that came to mind. It really :mad me off when this happens,especially because I know it is a false reading.
Thanks for the advice

How do you like the Accel DFI, I am hearing that it helps make alot H.P.

Chuck

72tccracer

One other thing. Did you ever think about just putting a 4.3k ohm resistor in place of the knock sensor? I have this done on my friends lt1.