Still Getting KR under hard acceleration(sometimes)

[RedLT4]

I've been trying to dial in the tune on my 5.3, and it's running really well. The problem is it's reading knock when I go WOT, but only sometimes. I cannot pin down why it's reading this, let alone why it's intermittent. I've chased this a lot since getting this running, but nothing has worked to make it go away completely. I cannot find anything rattling on the engine, I also have never heard any engine knock when it happens in the log. I've pulled timing in the areas where it happens, I'm running 93 octane, I've disabled burst knock and believe I desensitized the sensors in the tune. Also installed a catch can due to seeing oil in the intake. I know it's not going lean when this happens either, the wideband is reading around 12.2 AFR at WOT.
Hopefully someone have been down this rabbit hole before and might have some new ideas, because I'm out of them Help please. Tune and a couple of logs attached.

blazer 122820-INJ35psi-speedo-timingpull-trans.hpt
blazer123120.hpl
blazer123120a.hpl

[blindsquirrel]

Why are the trims adding so much fuel basically everywhere? I wouldn't worry about 2* of KR until the MAF/VE are taken care of.

[RedLT4]

Why are the trims adding so much fuel basically everywhere? I wouldn't worry about 2* of KR until the MAF/VE are taken care of.

Thanks for the reply. I'm not sure why the trims are adding fuel. It could be the NB sensors placement, or maybe some voltage offset issue when comparing the NB with the WB sensor readings? But I have done the MAF and VE tuning(with closed loop turned off) using the WB readings. It's also possible my WB readings are off and I should trust the NB more, but that just seems wrong.

And that isn't just 2*, sometimes it's none, or really low, but occasionally it spikes to 8* which is all it can pull.

[blindsquirrel]

Is this what the graph should look like after it's tuned with a WB?

screenshot.01-01-2021 13.43.45.png

[RedLT4]

Is this what the graph should look like after it's tuned with a WB?

screenshot.01-01-2021 13.43.45.png

No, and I'm new enough to this that it's entirely possible I totally missed something when I did it. I used the tutorials found here and on youtube to get the tune and histograms setup for the purpose. But according to the WB error percentages when I did it(commanded EQ compared to what the WB actually saw), it was close. So what do I do about it? Trust the wideband readings or ignore it and tune using the narrow band and LTFT? And just look at the WB for WOT adjustments?

While all of this could prove to be useful to me and the process of trying to learn tuning, the question I had was regarding possible sources of random knock readings. I know the the tune being off can lead to knock(too lean for example). But I do not see this being my specific issue. I'm logging 11.99-12.4 AFR when the knock is being picked up. Of course this assumes I can trust the WB readings.

[blindsquirrel]

If this is N/A, 11.99-12.4 is too rich. Too rich will cause knock too. Like I said, I would not waste time on random knock events until the thing is running like it should, at +3%/-5% trims, or whatever targets you're happy with. The computer shouldn't have to be working that hard to bring things back in line.

Think about this: if you made no change to the tune other than forcing it into open loop, it would suddenly be between 5%-18% leaner almost everywhere across the map. The trims are adding fuel everywhere. Why that is, whether it's because the data you collected from the WB was off, or the narrowbands are now off and jacking up the LTFT/STFT, I don't know.

[blindsquirrel]

Where is your wideband data? I see raw EGR voltage in the channels, but how are you getting usable data that can be turned into something that can be applied to VE/MAF?

[Cazbah362]

REDLT4 - I just went threw tuning my MAF and dropping my LTFT's and here is how I did it:

--- I followed this video, below are the steps etc I created to follow this approach: https://www.youtube.com/watch?v=xO_b8T6d_nI&t=569s

1. Force into Open Loop (Here are the steps I followed)

Fuel Tab
Oxygen Tab
LTFT - Disable
Long Term Fuel Trim
- 283 Scale all away across

Closed Loop Enabled
- ECT vs IAT - Set to 284

** Disable SFST

Fuel Tab
Open Loop / Base
- STFT Open Loop - Disabled

Fuel Tab
Temperature Control
- COT Disable - Note this is turned off already

Fuel Tab
Cutoff DFCO
Deceleration Cut Off
- Enable ECT Set to 280
- Stall Set to 12000

Airflow?
Dynamic Tab
Dynamic Airflow
- High RPM Disable - 200

Dynamic Airflow
Steady State
- Hi/Lo RPM Thresh to 200
- Hi/Lo Map Thresh to 0

Exhaust
General
- Enabled - to disable

2..

Once in open loop, I went to my VCM and added an AFR Err portion to it (attached is the VCM .xml you can use)

3...
I fired it up, did idle, short drives and a couple of spirited drives to capture data (I did this about 10 times - I am a newbie, so better tuners probably only do it once or twice)


Once in the 3% range, (I could only get about 5% ish) place back into close loop and tune the as appropriate.

I believe once you get the LTFT's in like Blindsquirrel is saying, it may help you with the knock? and it may not?

[RedLT4]

The 0-5V output from my Innovative LC-2 is being fed into the EGR input on the PCM. That is being converted to an AFR reading using [2811.10]/0.332446809+7.35 there were a couple of different versions of the math I found for doing this conversion. Don't know if this is best, but it seems to be working.

I did read about ground loops causing issues with WB readings, so I did ground the LC2 controller very close to where the PCM grounds as I saw it was recommended to ground the sensor/controller in the same place as whatever you are feeding the data into.

[RedLT4]

REDLT4 - I just went threw tuning my MAF and dropping my LTFT's and here is how I did it:

--- I followed this video, below are the steps etc I created to follow this approach: https://www.youtube.com/watch?v=xO_b8T6d_nI&t=569s

1. Force into Open Loop (Here are the steps I followed)

Fuel Tab
Oxygen Tab
LTFT - Disable
Long Term Fuel Trim
- 283 Scale all away across

Closed Loop Enabled
- ECT vs IAT - Set to 284

** Disable SFST

Fuel Tab
Open Loop / Base
- STFT Open Loop - Disabled

Fuel Tab
Temperature Control
- COT Disable - Note this is turned off already

Fuel Tab
Cutoff DFCO
Deceleration Cut Off
- Enable ECT Set to 280
- Stall Set to 12000

Airflow?
Dynamic Tab
Dynamic Airflow
- High RPM Disable - 200

Dynamic Airflow
Steady State
- Hi/Lo RPM Thresh to 200
- Hi/Lo Map Thresh to 0

Exhaust
General
- Enabled - to disable

2..

Once in open loop, I went to my VCM and added an AFR Err portion to it (attached is the VCM .xml you can use)

3...
I fired it up, did idle, short drives and a couple of spirited drives to capture data (I did this about 10 times - I am a newbie, so better tuners probably only do it once or twice)


Once in the 3% range, (I could only get about 5% ish) place back into close loop and tune the as appropriate.

I believe once you get the LTFT's in like Blindsquirrel is saying, it may help you with the knock? and it may not?

Thanks for the help. That is the process I used(even GRG's videos). I'm leaning toward changing my formula on my WB voltage readings to offset it and get closer to what the NB sensors are telling me. Looks to be about an 8% disagreement between the 2.

[chris35moto]

Do you have a gauge you can compare your scanner to? I took videos of my scanner and the aem live gauge to pick the right math equation and dial it in. I also was able to watch it with everything stock as a sanity check that the computer was seeing the same value as my gauge.

I'm pretty new to this as well but according to my graph your sensors seem to jive with the wideband at stoich. The oscillation is even similar to my 03 8.1. Where are you seeing the 8% difference?

Blazer Stock Vs Wideband.JPG

[blindsquirrel]

Also keep in mind, narrowband sensors do wear out over time. You cannot just assume that since they aren't setting codes they are OK, just like you can't assume spark plugs are in good shape just because it's not setting misfire codes. The time for replacement is long before they completely fail. If you don't know how long the O2 sensors have been in there, replace them before you spend a lot of time perfecting the tune... or else a month from now when one does fail outright and you replace them, you'll find that suddenly your tune is all jacked up for seemingly no reason and you have to start over.

[chris35moto]

Here's Commanded AFR vs Wideband with your math.
Attachment 106195
Wideband looks good as far as I can tell. Seems like Closed loop is working too. PE is enabling. Like others said this looks like VE or MAF that the fuel trims are fighting. Make sure you're logging enough hits on your histogram when the engines fully warmed up. I normally shoot for 25-50.

New O2 Sensors are a good idea. I think I might pick some up next time i'm at the parts store.

[RedLT4]

OK, I went back and re-did a bunch of VE and some MAF tuning. Went for a long cruise and it seems much more in line with the LTFTs. No idea how I screwed it up before(NOOB). Unfortunately the roads are crap here right now, so WOT was not a good idea! But in some of my attempts to get into PE, it still saw knock. The WB is showing pretty rich when the knock happens. 11.6 in some cases when 12.45 is being commanded. Also saw the commanded would occasionally slip down to 11.77 briefly, then right back up to 12.45. Not sure why, and if I should try to address it by adjusting the PE ratio?




blazer010321.hpl

[blindsquirrel]

As far as PE, I don't like how you have it set up. Don't be such a tightass with it - let it work when it needs to. A bunch of times in the latest log it was like (from memory) 50% throttle, .70g/s, 98 KPA MAP, 3500 RPM, and still not in PE. Then when it did come in is was a full point too rich for 300ish HP N/A engine. You don't need to go richer than 12.6 for this. Did you forget my earlier mention that too rich can cause knock too?

Make the PE ratio table a flat 1.16 across the board, it does not need to go to 11.7:1/1.25 like that around 4K. Raise the MAP enable to like 75 or 80, and drastically lower the TPS enable, to like 40% tapering down to 20%. I cannot think of a scenario where an engine would be at ANY throttle position with RPM above 1500 and MAP above 70-80 where it would NOT need to be in PE, except for meeting some dumb emissions requirement. It's under significant load there, it needs fuel.

Those nitpicks aside I think your KR is a random mechanical issue, engine mount, exhaust rattle, etc. There's no single set of conditions in the log that says "when running conditions are [X, Y, & Z], it will create KR." Sometimes it does it, most times it doesn't. Get under there with a soft-ish rubber mallet and tap on stuff until you find something that rattles when it vibrates.