18psi

iTrader: (76)

But first you need to upsize that baby efr to a 9180

patsmx5

iTrader: (16)

Looks like it rips pretty good!

aidandj

iTrader: (5)

It feels great. Right now I still don't have working vvt, a definitive misfire fix, and am currently going over logs trying to figure out if I'm getting any det pushing it to 25psi.

​​

patsmx5

iTrader: (16)

Post your logs and tune and I'll take a look. Is your speedometer GPS based or sensor?

patsmx5

iTrader: (16)

Also reading plugs is a reliable way to tell if you had detonation, good lighting and 10x magnification. If you see black specs (oil fused to ceramic) or small ***** of aluminum stuck to ceramic (from combustion chamber or pistons) it's detonating. Not a good realtime indicator, but a definitive after-the-fact way to checking if it detonated.

aidandj

iTrader: (5)

Speedo is wheel speed based. I'll post the logs in a minute when I get to another computer.

​​​​​ill probably pull the plugs tonight. Looks like rain, so can't even tune wastegate boost anymore.

aidandj

iTrader: (5)

Ok so this one was messing with EBC. Spiked to 300kpa, hit boost cut. Knock sensor definitely picked something up on cyl 1. Not sure about the others could just be noise. 5-14_22.08.36



This one was more reasonable. The second first pull has missing data, second pull has either a misfire or some bump in the road it appears. But also some knock sensor readings. 5-14_22.12.07

icantlearn

is 15.2 the max knock you saw?

aidandj

iTrader: (5)

On that pull yes.

But the total percent doesn't matter. That is dependent on gain and other factors. I have my gain turned down, but I think I'll be turning it up a bit more. I also turned cylinders 2 and 3 down a bit, so they all read somewhat evenly most of the time.

I'll pull plugs and try and get some pictures too.

icantlearn

is gain the same think as noise?

aidandj

iTrader: (5)

Think of gain like a multiplier.

The sensor outputs a "percentage" of noise. If the gain is set at .500 then what will be recorded to the ecu will be half of the value of the sensor.

For me 2&3 were closer to the sensor, and louder. So I lowered the gain. Which means they will appear just as "loud" to the ecu as 1 and 4.

icantlearn

ah ok. that makes sense. how did you know that 2&3 were louder aside from the sensor being closer to those cylinders? Like how did you know they were just "loud" as opposed to just high knock. And how did you know to set it to .5? were you just trying to get the %value close to 1&4?

aidandj

iTrader: (5)

I logged them with all of the gains set to 1. And adjusted from there. 2 and 3 were both louder than 1 and 4. And im shooting for like a 30% max baseline noise threshold. So that a spike will show up.

Gain will also mute real knock. So you want the most possible gain. with the least amount of false knock.

I'm actually going to turn mine up a bit after looking over my last logs.

icantlearn

I don't understand what the difference is if you set everything to 1 instead of 0 because the values will increase based on 1, so you don't know what is changing. How do you know when you are muting knock? Is it when the other cylinders have a drastic value difference?

aidandj

iTrader: (5)

A gain of 0 would zero out the signal everywhere.

Look at this screenshot.


1 was way higher than the others. And 4 was pretty loud too. To me that signifies that 1 definitely knocked. 4 im not totally sure on. It has a similar noise spike to 3, so that might be engine noise. Or maybe its more knock.

That pull was accidentally at right around 30psi. Too much boost for pump gas I bet.

sonofthehill

iTrader: (1)

Oops

icantlearn

4 looks very close to 2. not as much as 3. unless you mean 1 is close to 3

aidandj

iTrader: (5)

3 has a similar noise spike. But of a different magnitude.

I pulled a couple degrees of timing across the board. Because I like being safe. And EGTs were still around 1250*F.

curly

iTrader: (25)

What knock sensor are you using?

aidandj

iTrader: (5)

Stock NB. I'm pretty happy with the readings I'm getting so far. Noise increases with RPM, but not drastically, and doesn't seem to drown anything out.