Here are some of my datalogs I found interesting.
These were all done on the same day on the exact same stretch of road using Shell 93 octane pump gas.

For reference, here's my basic setup:I've been street tuning fuel for ~11.8-12.0 AFR with the very top end still being a little bit richer.

After installing the J&S Safeguard I made several passes at 3 PSI to set the knock retard sensitivity. Here you can see that zero knock retard is being logged.


After increasing boost to 4.3PSI I was surprised to see some high RPM knock already showing up at this low level of boost.


Moving up to 5.2PSI, knock is worse and showing up in a much wider RPM range:


At 9.3PSI knock is, not surprisingly, worse again.


At this point I moved back down to the 4-5PSI range and tried pulling some timing out which didn't seem to make much of a difference.
At 9PSI I was already pulling .75degrees/PSI, so I guess it makes sense that pulling more wouldn't necessarily help.

That led me to the next logical conclusion, that it must be fuel related.
I thought AFRs in the 11's would be rich enough to keep detonation at bay, but I guess I was wrong. So I moved back to my original timing numbers and started adding fuel to the 6-8k high cam maps.

I still have some work to do, but after a few runs it was obvious that more fuel is the answer. I would have never guessed that so much fuel was needed though.
Here you can see that compared to the previous 9.3PSI pull there is much less knock, but even at 10.3:1 AFR I'm still getting a little bit of knock at high RPM.

 

What heat range plugs and what's your gap?

 

This was using NGK BKR7ES-11 plugs with a .028" gap.

 

Don't get set on one particular value trying to set-up the tune. Just go with the engine likes. Have you tried pulling more than .75* timing? How sure are you about the accuracy of the wideband? How do the plugs look?

 

I understand what you're saying, but unfortunately the wideband and the J&S are the only things I have giving me data. With that in mind I'm working under the assumption that the engine doesn't like detonation
I was initially pulling 1.5*/PSI. Moving to 1*/PSI and then .75*/PSI each time made the car feel noticeably faster.
Of course motors often tend to feel their quickest right before they grenade.
I'm running a step-retard right now which looks like this:So, the 3PSI pull with zero knock was using ~1.25* more timing than the 4PSI pull when knock first starts to occur.
I don't have anything to compare it to, but it seems to be working fine. I've pulled it out and re-calibrated it a couple times.
Closed-loop works great, I'm getting ~27MPG without much work, which leads me to believe it's accurate.
I haven't been reading plugs between runs, I need to get a magnifier and it's kind of a pain to attempt while street-tuning.
Last time I pulled them they looked like a typical rich plug, no damage at least.

 

Is there an indicator on the J&S that it's actually sensing knock and this isn't somehow fluke voltage on the D14 input?

 

The factory knock sensor is connected directly to the J&S Safeguard input (I pulled D3 out of the ECU harness).

There's a simple POT to adjust the sensitivity of the J&S. Basically I just started with it at about 1/2 and did 3PSI pulls. Initially it was showing knock constantly. I kept lowering the sensitivity until I was able to pull to redline a few times with zero knock. This eliminates false positives from engine noise, turbo/wastegate etc. If I was actually knocking at that point the sensitivity could theoretically be too low, but I'm pretty confident that it's set correctly.
Here's the instructions I followed if you're curious: http://jandssafeguard.com/Archive_In...reenHonda.html

The J&S Safeguard monitor output is what I'm inputting into Neptune via D14. This is a 0-1.3V with 0.13v steps signal indicating the amount of knock retard the J&S Safeguard is applying. Example at the bottom of this page: http://www.jandssafeguard.com/tech.html

You can see in my logs that the signal sits at zero most of the time. There is slight noise visible in live graphing when I turn on the headlights, but it's well below the threshold for the first 0.13v step.

If it were an electrical noise issue in either the Knock Sensor>J&S or J&S<ECU I wouldn't expect adding fuel to change anything.

Based on the relationship to boost pressure and the response to fuel I'm fairly certain this is real knock.

 

Holy crap...needing to run 10.3 AFR's at 9lbs. Something is not right. Ive tuned a lot of cars and never had to go richer than 11.8AFR at 9lbs. I looked at the J&S, and i gotta ask, what would make you wanna use that? It looks like its something that isnt any real benefit, just adds the ability to verify knock. Which an OEM ecu can do with an OEM knock sensor. And being able to adjust the sensitivity of it just seems redundant. You would need to verify somehow that there is real knock to set up the J&S. Knocks sensors have proven to be, to me over the years so finicky as electrical noise and "phantom" knock leads to information that doesnt benefit the tune.

Does the J&S unit have a way to directly monitor what its seeing for knock? Can you hook it up directly to your laptop and monitor it through the unit itself without getting the info second hand through your ECU?

 

Stock knock sensor is WORTHLESS!!!

Why does everyone believe that knock sensors actually pick up knock in a completely different setup (not stock in any way) than they were designed for?

 

My motor has 100% stock internals. Sure there's additional external noise from the turbo, wastegate and boost control solenoid; but since knock frequency is primarily determined by the cylinder bore size I would expect knock to sound exactly the same.It does more than just verify knock, it automatically retards timing on individual cylinders when it detects knock and then keeps trying to advance back to normal timing. I look at it as cheap insurance as well as hopefully letting me get the most out my street tuning.
The J&S is seeing the raw signal from the OEM knock sensor. I guess I could log that, but I'm not sure if it would be useful or not.

I plan on setting up an electric det can to record sound while datalogging, that way I'll be able to compare the audio to my datalogs.

 

Its not just bore size, but what do i know...good luck, I hope you figure it out.

 

I always wanted one of those js safeguards to street tune but dont they have there own knock sensors that needs installing, or two of em? Its cool to c ur datalogs with it, but i would think running that rich, ur on the line of washing out the rings. Do you have any oil in ur intake piping? what size exhaust r u running? these will also contribute to knock.

 

The model I have (Older Honda specific v1) uses the stock honda knock sensor; I'm not sure what the current models use.

My intake piping is bone dry, no oil. 2.5" DP w/ 1.5" dump piped back in before the cat, 2.5" high-flow cat, Greddy exhaust, 60mm I think, I definitely need to get a bigger cat-back at some point, but it's fine for now, no boost spike or creep.

 

hmm yeah i guess thats size is ok for the low boost pressures ur running. U double check timing? and cam timing? nothing else i can really think of, other than proper fuel

 

To help determine if the J&S Safeguard is responding to real knock I did some back to back test runs with pump gas and then higher octane fuel.

This first run was using Shell V-Power 93 Octane pump gasoline.
Quite a bit of knock retard at only ~6PSI.


At this point I stopped to refuel.

I had a little less than 1/4 tank of gas left. (~2.5gal).
I brought along a 5gal container with some home-brew.

Here's what I used:
2gal Xylene
1qt Acetone
~1/2oz ATF
3gal Shell V-Power 93 pump gas.
I know that's more than 5gal, I went a over the fill line.

Added to the ~2.5gal of gas left in my tank this should yield ~100 octane.
Filling from one of those ventless container takes a while, so this next run was exactly 15min later, same stretch of road.


Bye-Bye Knock Retard.

Knock Retard does show back up with higher boost.


Based on my previous results with increased fuel and now with varying octane I think it's safe to say this is indeed real knock.

I plan on doing some more tests using the individual cylinder trim settings in Neptune to determine if the knock is limited to one particular cyl or if it is occurring in all four. This has also led me to the decision to add a water injection system.

 

Oh the joys of tuning. Glad you're making progress.

 

Nice test. I was wondering if you had more test results to share? Thanks.