Knock retard?

[Rich Z]

I'm getting a bit of knock retard logged and not sure what is going on. RPMs are not high, AFR looks good, and timing doesn't look out of whack. I've attached the latest log file, and the KRs are showing up between 6570 and 6620. It's not much, 1.1 degrees, but it concerns me anyway.

Any ideas?

[Gregs]

I'm getting a bit of knock retard logged and not sure what is going on. RPMs are not high, AFR looks good, and timing doesn't look out of whack. I've attached the latest log file, and the KRs are showing up between 6570 and 6620. It's not much, 1.1 degrees, but it concerns me anyway.

Any ideas?


I don't see any knock in that frame range. The only knock I see is at frame 9896 and its not concerning. Its only 1.1* and it doesn't have a sawtooth pattern=false knock.

[ferocity02]

Any aftermarket valvetrain? That can cause false knock.

[joecar]

I see what Gregs sees.



Real knock/detonation usually shows up as a sawtooth waveform

( PCM detects knock, PCM removes timing, knock stops, as PCM restores timing knock returns, repeats... )

altho, a single event knock is not distinguishable (since it looks like a single spike).

[Rich Z]

I don't see any knock in that frame range. The only knock I see is at frame 9896 and its not concerning. Its only 1.1* and it doesn't have a sawtooth pattern=false knock.

Well, that's odd. I definitely saw the KR at the frame rate area I mentioned when I was looking at it. I even made a note on a pad. Yet when I look at it now, yeah, it's over in the 9896 range.

Beats me....

Does this mean maybe the sensors are set a tad bit too sensitive?

[Rich Z]

Any aftermarket valvetrain? That can cause false knock.

Yes, it's a Jesel valve train on the engine.

[Rich Z]

I see what Gregs sees.



Real knock/detonation usually shows up as a sawtooth waveform

( PCM detects knock, PCM removes timing, knock stops, as PCM restores timing knock returns, repeats... )

altho, a single event knock is not distinguishable (since it looks like a single spike).

Ah, thanks for the explanation. Never having seen KR before, I had no clue about what to look for. All I could see was SOMETHING was there. So I figured I had better ask about it so I wasn't getting myself into trouble.

[Gregs]

Ah, thanks for the explanation. Never having seen KR before, I had no clue about what to look for. All I could see was SOMETHING was there. So I figured I had better ask about it so I wasn't getting myself into trouble.

no problem, better to ask then be sorry later.

I honestly wouldn't worry about 1.1* its just so small. Its the only KR shown in the log and its never repeated.

[joecar]

BTW: your log file indicates MAF DTC(s) present, confirming SD mode.

[ferocity02]

Well, that's odd. I definitely saw the KR at the frame rate area I mentioned when I was looking at it. I even made a note on a pad. Yet when I look at it now, yeah, it's over in the 9896 range.

Maybe you had the data filtered when you noted the frame range? Filtering also removes data from the charts in the dashboard.

[ferocity02]

Real knock/detonation usually shows up as a sawtooth waveform

( PCM detects knock, PCM removes timing, knock stops, as PCM restores timing knock returns, repeats... )

altho, a single event knock is not distinguishable (since it looks like a single spike).

Do you have a screenshot of what real knock looks like?

[joecar]

Do you have a screenshot of what real knock looks like?Here's but 4 pics:
attachmentid=12969&d=1334894017 (http://forum.efilive.com/attachment.php?attachmentid=12969&d=1334894017)
attachmentid=6687&d=1259177496 (http://forum.efilive.com/attachment.php?attachmentid=6687&d=1259177496)
attachmentid=15354&d=1371161608 (http://forum.efilive.com/attachment.php?attachmentid=15354&d=1371161608)
attachmentid=7539&d=1270370224 (http://forum.efilive.com/attachment.php?attachmentid=7539&d=1270370224)

[Gregs]

Here's but 4 pics:
attachmentid=12969&d=1334894017 (http://forum.efilive.com/attachment.php?attachmentid=12969&d=1334894017)
attachmentid=6687&d=1259177496 (http://forum.efilive.com/attachment.php?attachmentid=6687&d=1259177496)
attachmentid=15354&d=1371161608 (http://forum.efilive.com/attachment.php?attachmentid=15354&d=1371161608)
attachmentid=7539&d=1270370224 (http://forum.efilive.com/attachment.php?attachmentid=7539&d=1270370224)

wow! what do you have your recovery rates at? Mine are stock and they take a LOT longer to get back to zero.

[Rich Z]

BTW: your log file indicates MAF DTC(s) present, confirming SD mode.

Yes, that is correct. I am logging data to peck away at the VE table. I believe I've done all the steps necessary to disable the MAF and any fuel modifiers, but heck, there are quite a handful of them lurking in there, so I can only HOPE I got them all.

[Rich Z]

Maybe you had the data filtered when you noted the frame range? Filtering also removes data from the charts in the dashboard.

Ah, yes, that is probably it. I was working with a map and did initiate a filter. I wouldn't have thought it would filter out entire frames, though. I'll have to try to remember this in the future.

[Rich Z]

Here's but 4 pics:
attachmentid=12969&d=1334894017 (http://forum.efilive.com/attachment.php?attachmentid=12969&d=1334894017)
attachmentid=6687&d=1259177496 (http://forum.efilive.com/attachment.php?attachmentid=6687&d=1259177496)
attachmentid=15354&d=1371161608 (http://forum.efilive.com/attachment.php?attachmentid=15354&d=1371161608)
attachmentid=7539&d=1270370224 (http://forum.efilive.com/attachment.php?attachmentid=7539&d=1270370224)

Man, thanks! That helps a LOT to see what knock actually looks like.

[joecar]

wow! what do you have your recovery rates at? Mine are stock and they take a LOT longer to get back to zero.
Those were pics from other people's logs I took...

but look in 2002 Camaro/Firebird 12212156 file.

[ddnspider]

Would you say these look like false knock as they're not very jagged or sawtooth?
http://i5.photobucket.com/albums/y159/ddnspider/falseknock2.jpg
http://i5.photobucket.com/albums/y159/ddnspider/falseknock1.jpg

[joecar]

It's hard to say.

[ddnspider]

It's hard to say.

Can you expound at all, or is there an easy way to test it? I figured with such low timing and rich A/F it must be false...

[EagleMark]

It also looks like manual trans noise? That is a shift in the second dash?

Log again with less PIDS and use spark advance, knock count, knock retard, TPS, RPM and .... Depends on the vehicle but sometimes knock count is more precise to pattern when knock retard looks like what you have. Then you know the difference on your rig.

It's only pulling 2. So take out 3 degrees and try again? Same pattern?

[ddnspider]

It also looks like manual trans noise? That is a shift in the second dash?

Log again with less PIDS and use spark advance, knock count, knock retard, TPS, RPM and .... Depends on the vehicle but sometimes knock count is more precise to pattern when knock retard looks like what you have. Then you know the difference on your rig.

It's only pulling 2. So take out 3 degrees and try again? Same pattern?

I have been searching all morning for further info on changes to make to the knock tables to try and minimize its intrusiveness if it is due to background noise. And yes you are seeing a shift. I will look to log knock count but I believe its either:

A) false knock as low 11's A/F with ~12* of timing on a 9:1 compression engine should be WELL below the threshold, or

B) it is actually TOO rich/low timing and is causing real knock in which case I need to lean it out or increase timing.

[joecar]

Real knock is very distinctive, it has a very recognizeable tight sawtooth pattern...

but just because you don't see that pattern it doesn't mean it didn't knock...

if you can hear it then it was real, obviously... but sometimes knock is present without being heard (however, large pronounced episodes of knock will be audible)... and also consider that the knock sensors may be degraded.

When tuning on the dyno, usually knock phones (Chassis-Ear) is used to listen to the engine.


Try what EM said, see if your logs change any. Also inspect under car, try and manhandle the engine/trans to see if the mounts are broken, same with exhaust and heat shields.

[ddnspider]

I definitely cannot hear it, especially with the cutout and when the wastegate opens. I also found out that you cannot log the knock count pid on a 98 PCM :(

Another screenshot from yesterday
http://i5.photobucket.com/albums/y159/ddnspider/knock.jpg

[ferocity02]

Have you tried running a higher octane fuel?

[ddnspider]

Have you tried running a higher octane fuel?Not recently. Several years ago I did play around with adding race fuel and saw no change in knock being there or not. From the latest log you can see timing is in the single digits and A/F is rather rich so I would be shocked that its real, although IAT's are too high for my taste, but then again FL is blasted hot right now lol.

[Rich Z]

Hmm, I'm presuming a manual transmission? If so looks like you are getting the KR when the clutch begins to be engaged for a shift, and the again when in the process of being disengaged. Notice KRs stop when the clutch is fully engaged. Probably just a coincidence, but you might want to check to see if those KRs are consistently at clutch engagements/disengagements.

[ddnspider]

Hmm, I'm presuming a manual transmission? If so looks like you are getting the KR when the clutch begins to be engaged for a shift, and the again when in the process of being disengaged. Notice KRs stop when the clutch is fully engaged. Probably just a coincidence, but you might want to check to see if those KRs are consistently at clutch engagements/disengagements.

Thanks for the insight. Yes it is a rebuilt M6 with twin disk clutch and a Ford 9" w/ locker out back and a custom driveshaft so I know there is definitely some drivetrain related noise. That is interesting info relating to clutch engagement/disengagement, I may have to pay more attention and try and coincide the events, although things happen quite quickly with the new wastegate spring installed :)

I would be interested if anyone has had luck with desensitizing the knock sensors or raising background noise levels to see if it lowers the knock levels.

[hog]

Tech 2 should read out knock sensor counts. On L31 Vortec 350 I was getting some knock retard, tried 104 unleaded race fuel and still had this "knock". I ended up installing an LT4 knock module into the PCM and my troubles went away. If you take a hammer and hit somewhere on the engine, the knock sensor should register, you may need to use a drift of a blunt chisel to access a metal part of the engine. Increasing background wouldn't lower knock, it would raise it, or if your issue isn't real engine knock, it wouldn't affect it at all. Back in the day it was said that wrapping the knock sensor threads in Teflon tape would mechanically desensitize the sensor. I guess you need to determinbe if you are experiencing actual engine knock, or if the knock sensor is picking up "background" noise in the frequency range that resembles actual engine knock. Solid engine mounts, exhaust banging and many other drivetrain noise can mimick knock.

[ddnspider]

Tech 2 should read out knock sensor counts. On L31 Vortec 350 I was getting some knock retard, tried 104 unleaded race fuel and still had this "knock". I ended up installing an LT4 knock module into the PCM and my troubles went away. If you take a hammer and hit somewhere on the engine, the knock sensor should register, you may need to use a drift of a blunt chisel to access a metal part of the engine. Increasing background wouldn't lower knock, it would raise it, or if your issue isn't real engine knock, it wouldn't affect it at all. Back in the day it was said that wrapping the knock sensor threads in Teflon tape would mechanically desensitize the sensor. I guess you need to determinbe if you are experiencing actual engine knock, or if the knock sensor is picking up "background" noise in the frequency range that resembles actual engine knock. Solid engine mounts, exhaust banging and many other drivetrain noise can mimick knock.

If EFI Live wont recognize the knock sensor count pid on a 98 PCM how would a Tech 2 know? Either way unfortunately I dont have access to one. That is definitely a good idea about trying to bang on the exhaust or engine and see if it picks up knock on the scanner though, I may have to give that a shot. I know there are several threads roaming the internet regarding min and initial learned knock thresholds as well as adjusting the amplification levels, but I have yet to see anyone walk through them and explain what does and doesnt work. My fear of changing any parameters other than the recovery rate is that they may overly desensitize them making them useless.

[hog]

If EFI Live wont recognize the knock sensor count pid on a 98 PCM how would a Tech 2 know? Either way unfortunately I dont have access to one. That is definitely a good idea about trying to bang on the exhaust or engine and see if it picks up knock on the scanner though, I may have to give that a shot. I know there are several threads roaming the internet regarding min and initial learned knock thresholds as well as adjusting the amplification levels, but I have yet to see anyone walk through them and explain what does and doesnt work. My fear of changing any parameters other than the recovery rate is that they may overly desensitize them making them useless.
Because a Tech 2 is actually designed for GM technicians to diagnose issues with. EFILive isn't. On my Vortec 350 trucks, there is a knock retard PID that none of the scan tools pickup, that Tech 2 does. I do understand you being cautious, esp. using a force inducted application.peaceHog

[ddnspider]

Because a Tech 2 is actually designed for GM technicians to diagnose issues with. EFILive isn't. On my Vortec 350 trucks, there is a knock retard PID that none of the scan tools pickup, that Tech 2 does. I do understand you being cautious, esp. using a force inducted application.peaceHog

Thanks. Some interesting points....

-knocking on the exhaust etc did not induce knock on the scanner.

-I ADDED timing and LEANED out the A/F and the knock that was picked up was at a noticably lower level. Looks like my option B) may have been happening.

On a separate note, I am tired of this hack way to get my A/F ratio on the 98 pcm and have already ordered an LS7 maf and 4" tubing and will scale the tune. Anyone happen to have a starter MAF transfer function that would get me started?Car is a 383 with 60 lb injectors at 4 bar of fuel.