10v of fury?

 

ROFLCOPTER

I was wondering though if it was something like the Subaru/Nissan sensor that is pretty high resolution for 5v. Maybe I'll have to order one to see whats so different about it.

 

A wideband knock sensor just means its sensitive to a wider range of frequencies. Would be interesing to compare the 8/9 sensor to the 10 sensor on some sort of sound generator.

The knock control tables in the 10 ROM are identical in form to the ones in the 8/9 ROM, so I'm guessing that if the Evo 10 knock sensor has the same form factor as the 8/9 knock sensor, then it can be bolted on to an Evo 8/9 and the knock tables adjusted to make it work like it was there from the beginning.

 

Knock frequency is mathematically equated, why would we need a wider range of potential frequency other than larger bore?

I'll check CAPS and then see what they run. If it doesnt break the R&D budget I'll pick one up and send it over to you M.

 

I think what the Works guy was implying is that although the knock frequency spectrum (its likely a bell shaped curve) is set by the cylinder dimensions, the engine and other things attached to the engine can mask or damp some portions of the knock frequency spectrum and thereby affect what frequencies reach the knock sensor. It seems (still guessing at what the Works guy was implying) that with some basic changes to engine components its possible to strongly damp the signal in the frequency range that the narrowband knock sensor can detect resulting in poor knock sensing capability. A sensitivity to a wider range of frequencies may help deal with changes in signal damping due to changes in engine components. This is all my interpretation of what the Works guy wrote. Would be great to hear what he has to say.

JB, if you decide to send me an Evo 10 sensor, can you send me an Evo 8/9 one as well? I should be able to rig up a sound table to test the frequency response of both.

 

Im at 4 degrees at peak torque right now but need to lower it back down to where it was at 1 or so. Since I keep getting a few random counts, does it seem reasonible to start at 1 or so at peak tq, then a slow climb to 6 or 7 degrees by 7500? To the guys that run this little of timing, what is your ramp rate like? As in, timing vs rpm during a pull. I know all cars are different, Im just looking for a good starting point.

Thanks

 

Random counts? Like what?

 

Can anyone shed some light on how the knock sensor filter tables correlate to frequency and/or voltage? What is the function of maps 1-12?

 

Nothing. They don't change anything.

 

Those maps are all bogus. jcsbanks determined the true knock tables about six months ago. None of them have to do with "filtering" frequencies.

 

The only issue I have seen with the factory knock control is at higher revs with "noisy" motors. I have no idea if the EVO behaves the same way, but I remember a post over on dsmlink forums several years ago from Tom talking about how the factory knock control worked. He showed some datalogs of the raw knock levels of each cylinder and then he showed a trace of the noise level that the ECU considered to be detonation.

When a detonation event that produced a high enough output to go above the allowable noise level, the ECU responded by pulling timing. Pretty straight forward. The interesting thing happened after that. The allowable noise level would increase slightly. The affect this has on a loud motor is that mechanical noise can actually be loud enough to make the allowable noise level increment up to a level that is higher then even heavy detonation. Essentially, it gets to where the motor is just so noisy that the ECU can't distinguish between detonation and mechanical noise and it stops pulling timing.

I have datalogged this first hand where looking at individual cylinder raw knock showed the motor was so noisy, that you really couldn’t tell the motor was knocking at all. However, the spark plugs told the story of detonation.

I have also noticed that as timing advance drops, the engine has a tendency to get louder on the knock sensor. If you are running very low timing, you may be getting close to this level where the ECU can no longer determine what is knock and what is noise.

As for what was possible 5+ years ago. I had a friend that was running 34 psi on 91 octane on an SC61 (T350 turbine wheel, GT35R compressor wheel, 0.63 A/R T3 turbine housing, T04E compressor cover). Not exactly high tech stuff there. The car ran like 125-127mph traps on ~28 PSI and was a HEAVY car. Full interior, sub/amp, etc. This was a stock internal 7-bolt 2G with like 100k miles on it too.

Why was he able to run the car like that? Because he flat out didn’t give a crap if it let go on him. He ran that car on probably 13:1 AFR (datalogged like 130% IDC on 720cc injectors) and 15+ degrees of timing up top at 34 PSI revving out to 8500-8800 RPM. I think when it comes down to it the 4G63 is just one tough little bastard. The motor eventually tossed a rod from a stretched rod bolt after about 8 months of that abuse. Had he tossed in some ARP rod bolts, the thing may have lasted a LONG time.

 

Above a certain rev range there isnt enough time for a motor to knock. The factory knock control works out to at least 8k though as I have seen it pick up noise to there. Above 8 I havent ever seen any noise except when lifting after a pull.

One thing that we have noticed is that when you are the edge on a pumpgas tune because of boost vs timing, if the revs are high enough leaning it out some will help prevent low timing misfires. I have seen low timing misfires show up as 4-6 counts of knock (further compounding the problem) and other than the dyno sheet you'd never know it was dropping 100whp during these events. That dyno sheet I put up of my last high boost/pumpgas attempt on the stock intake showed it at 537whp the car just lost fire and came back at 7800 or so. Pretty violent looking on the dyno, a slight hiccup in the car.

 

I searched hi and low for the above info with no success. Does anyone have a link explaining the true knock tables and/or xml info for either 94170015 or 96260009 (USDM 03 EVO VIII)?

 

With 4 degrees at peak tq climbing to 7 degrees at 7500, I will mostly see no more than 1 count through the pull, but when doing constant pulls or just 1 out of 4 pulls on average I will see anywhere from a 5 count up to 10....this usually isnt at the same rpm and it usually doesnt hold throughout the pull.

 

I would just reduce your boost level. You cannot force a car into accepting too much boost. I will either take it and like it or it won't. There are many factors contributing to what an engine will accept.

Your timing numbers are an indicator that you are trying to force the car. I would revamp the whole map based on lower boost.

7* at 7500 is ridiculous. There are pumpgas cars running 35psi boost at 7*.

I would reduce the boost until the car will accept a minimum of 11-13* at that rpm.