Reddragon88gta

I was reviewing some old data logs and saw that my car had some knock counts and at times was pulling timing at WOT or when accelerating quickly.

I understand some knock counts being present are fine, especially at start up etc and if it pulls a few degrees of timing occasionally not a huge deal but this seems like a problem.

Examples: At start up I would see low teens like 11, engine would warm up and I would drive normally and the knock count would just stay at 11. Then if you accelerate quckly(up to 4500 rpm give or take) or WOT knock counts would go up to 30, pull up to 9 degrees timing, resume normal driving no knock retard but counts still 30 and not moving. Accelerate quickly again knocks hit 116, pulls timing again, resume normal driving knock stay at 112 until shut down.

What I thought was odd was the knock counts never went away or really moved while data streaming? They would only go up and never down even when rpm was normal/low? I also did a data log while the engine was in park and at 3000 or more rpm’s ( didn’t want to rev too high as it was unloaded) and I had no knock retard or knock counts above the typical low teens present at start up that never went away.

Before I dig too deep I have a new knock sensor on the way as the one on there looks original or I’m thinking false knocks somewhere.

Base timing is 6 BTDC and had had the same results at 8 BTDC. Fuel pressure is set to 48psi on AFPR. I have pulled a few plugs and they look fine with no signs of detonation.


I tried adding the whole video of my data log but it wouldn’t work, here are a few screen shots:


Normal driving, at beginning or drive.




Normal Cruising after WOT.



WOT or close to it.



Idle before shut down.

ULTM8Z

I was battling something similar and ended up tracing it to excessive torque and/or bad knock sensor. On a modified engine where noises can come from things like headers, roller rockers, etc... probably makes the knock sensor's response to things much more sensitive to it's torque.

https://www.thirdgen.org/forums/diy-...tard-real.html

Reddragon88gta

I actually read your thread and found some good info in there. My engine internals are bone stock, just bolt ons with no headers. I need to get under there and really look at my magnaflow Y pipe and see it’s close/hitting the sub frame connectors.

I am leaning towards the knock sensor first as it looks original and pretty crusty. I found a NOS AC Delco so hopefully that does the trick.

Reddragon88gta

Also the fact that the knock counts never go down is puzzling me? Even when normal driving but I do not know what how what the values should be ( probably low I assume) for knocks and if they are supposed to fluctuate.

The fact the knock counts go up and never down seems odd but I’m no expert on reading data logs.

ULTM8Z

IIRC, the knock counts never go down. It'll just keep going up until it reaches 256 or something like that, and then it'll go back to zero and start over.

Tuned Performance

iTrader: (92)

I think you’re right or even 255 then start back over

Reddragon88gta

Thanks for the insight on the data guys, I could not find much info on that myself.

Reddragon88gta

I replaced the old knock sensor, looked to be original and I was surprised it didn’t fight me more. All I had to do was use some PB blaster and it broke loose, I was easy on it as I didn’t want it to separate or strip out. I barely lost any coolant as well, it was tight putting the new one in but I got it and torqued it down to 12ft lbs.

Spec is 15 ft lbs but I read a few threads where Rbob (pretty serious contributor in EPROM stuff) recommended as low as 11ft lbs to help with over sensitive knock detection so hopefully 12 does the trick. Weather permitting in about month in a half I will be able to get it back on the road and run some data scans to see if my knock retard at WOT is taken care of.




Looks crusty





Still made in Mexico like the new ones but I like NOS for sensors when I can find them.




Looks like a genuine GM part.

Reddragon88gta

I had a chance to finally get the car out of storage and take it for a ride/data scan, it’s still pulling timing after swapping in the new knock sensor. Only thing new that I noticed is that BLMs are 129 when it happens instead of 136. Car seems to typically run on the richer side. The car runs great but I just feel with 9 degrees being pulled at WOT I am losing performance.

Also it doesn’t pull any timing in park/neutral which is the same as before.

I took a look underneath while I was swapping the knock sensor and could find nothing hitting or rubbing etc. but I may have missed something.

I am due for my state inspection next month and may have the shop take a look, they are actually really good with TPI/OBD1.

Reddragon88gta

One other thing to add is the knock counts were not as high as before so maybe the new sensor and lower torque setting contributed to that? I may try to lower torque setting down to 10 or 11 ft lbs and see what happens. Another thought is add some octane boost to 93 with ethanol instead of the 91 non ethanol but that feels like grasping at straws almost.

Reddragon88gta

I adjusted the torque a bit on knock sensor and while it has less overall knocks than before it still pulls 8-9 degrees at WOT or heavy acceleration.

I also noticed at times if you give the throttle a quick blip at low rpm it pulls a few degrees. At this point I am really leaning toward false knock. I drove around a bit with the sensor unplugged and the car ran great, no check engine light etc. but I really don’t like the idea of doing that.

One thing I did notice is that on the first ride of the day when engine wasn’t fully warmed up I gave it a shot of rpm while driving and no knock retard…

Also after doing a ton of research on a C4 Corvette site I saw more than a few posts dealing with false knock and it was solved by an “upgraded” ESC module, a Corvette guru called Gordon Killebrew recommended it and many said their issues were solved afterwards.

The part in question is ESC module GM part# 16038331 but seems to be near unobtanium but I found one on eBay (seems like a factory part for 4.3 V6 TBI engine?). It’s a shame they are so non existent as a lot of ppl said it worked great. A few ppl even listed the cross reference numbers for aftermarket but I had no luck there, seems to be out of production. The ESC module will be here hopefully by the weekend and I will swap them out and see what happens.

Tuned Performance

iTrader: (92)

Was the application mentioned? Standard motors and wells still makes esc modules. They fit various engine sizes. Lxe29 is nla too

Tuned Performance

iTrader: (92)

looks like all nla

AC / DELCO / GM 16038331 OEM
AC / DELCO / GM 216-35 OEM
AIRTEX 5F1015 aftermarket
BORG WARNER ESC109 aftermarket
CARDONE 164-404 aftermarket
ECHLIN ESC209 aftermarket
KEM 164-404 aftermarket
MICROTECH 59-0637 aftermarket
STANDARD / HYGRADE LXE29 aftermarket
WELLS ESC108 aftermarket

Reddragon88gta

That's what I found too, they seem impossible to find. I feel lucky to have found one.

Reddragon88gta

This is what I got from one of the GM parts dealer retailers. Pretty obscure group of cars which explains why the ESC module is so hard to find.

Reddragon88gta

I will add that the short ride I took with the knock sensor unplugged the car felt stronger but that may just be the placebo effect but to me it felt better.

Reddragon88gta

The new ESC module arrived today, I swapped it out but ran out of time for a test drive so that will be tomorrow or the next day.

After that I took another look underneath for sources of false knock and may have found something, one of the the 2 auto trans rubber mount bolts was pretty loose, the other just needed a bit to get back to tight.

I really don’t know how long that was loose but I doubt very long. My data logs go back a few years ( picking up knock) also I have done work in that area not too long ago and saw nothing wrong.

Reddragon88gta

More of the same after swapping the ESC module, it’s still pulling timing under acceleration. Right before swapping ESC modules I noticed I had a slightly loose auto trans mount that I torqued back down but still no change.

Reddragon88gta

Did more data logging and even fueled up with 93 and octane boost, still no difference. One thing I noticed is the knock counts and retard can be dictated by blipping the throttle. Even at low rpm if you just stab the throttle quick it registers a knock count and has a small amount of retard.

I already checked the motor mounts and they are fine, I even tried tapping the y pipe in various areas to see if the knock was caused by exhaust but nothing. When you tap the exhaust manifold it registers knock but that is normal. I even pulled a few plugs and they looked great.

At this point I’m disconnecting the knock sensor and may have it looked at next month.

SbFormula

My 2cents worth. There are probably other ways to go about with your issue though.

For analysis, you might want to look at the knock retard (KR). I like using a chart. Every time KR goes up, it's an indication of knock detection. Knock count (KC) could stay still but KR could go up. So KC is not as accurate as KR to detect knock event. There is a thread on that on TGO. There is also good information on TGO, don't remember who, who did some testing with knock. I did my own testing with stock LB9 as well. In both our cases, it was impossible to eliminate KC. It's also quite hard to discern between real knock and false knock.

Bottom line, we shot for "Net SA". Means, trying to minimize knock event and KR to get maximum average SA over RPM range. A knock event that triggers 8* of KR could linger for a few seconds during WOT even if no other knock events are detected. That ends-up pulling timing a lot. To do so, you have to look at:
Knock Attack Rate vs RPM
Knock Recovery Rate vs RPM
Maximum KR in PE
Maximum KR
If you have a '7730 with $8D, you also have to disable Low Octane Knock Routine

Even by pulling as much as 10* of timing across the board (see test#4 below), we still got knock events! We did lose power and throttle response.

Here's a summary of testing with a stock LB9 on $6E, rolling WOT runs on 2nd and 3rd (MT) on 3.08 rear gear. No testing was done on 4th gear as it would have yield way too much speed for public highways.

1) As is, from original owner after sitting in a garage for a long time:
2nd Gear
CT: 200Deg.F.
MAT: 120Deg.F. (heat soaked)
A: 30mp/h to 63mph in 4.8sec
B: 35mph to 63mph in 4.4sec
A: Avg SA 31.4*, Avg Net SA 30.4*, 4 knock events, -0.9* avg KR, Max KR -2.3*
B: Avg SA 31.5*, Avg Net SA 28.9*, 6 knock events, -2.6 avg KR, Max KR -3.7*

A: Chart of commanded SA (ORG) vs Net SA (BLU). Net SA is Commanded SA minus KR.



2) With tune-up:
2nd Gear
CT: 187Deg.F.
MAT: 86Deg.F.
A: 30mp/h to 63mph in 4.5sec
B: 35mph to 63mph in 4.0sec
A&B: Avg SA 31.8*, Avg Net SA 28.9*, 12 knock events, -2.9* avg KR, Max KR -4.6*

A&B: Chart of commanded SA (ORG) vs Net SA (BLU). Net SA is Commanded SA minus KR.

3rd Gear
CT: 184Deg.F.
MAT: 84Deg.F.
A: 42mp/h to 88mph in 10.1sec
A: Avg SA 31.9*, Avg Net SA 28.4*, 19 knock events, -3.5* avg KR, Max KR -5.6*
No chart!

3) After tuning:
2nd Gear
CT: 175Deg.F.
MAT: 93Deg.F.
A: 30mp/h to 63mph in 4.2sec
A: Avg SA 26.9*, Avg Net SA 26.1*, 7 knock events, -0.8* avg KR, Max KR -2.2
A: Chart of commanded SA (ORG) vs Net SA (BLU). Net SA is Commanded SA minus KR.


3rd Gear
CT: 173Deg.F.
MAT: 92Deg.F.
A: 42mp/h to 88mph in 9.7sec
A: Avg SA 26.5*, Avg Net SA 25.2*, 12 knock events, -1.3* avg KR, Max KR -2.3*
A: Chart of commanded SA (ORG) vs Net SA (BLU). Net SA is Commanded SA minus KR.



4) After tuning and pulling 10* SA across the board:
2nd Gear
CT: 181Deg.F.
MAT: 127Deg.F.
A: 30mp/h to 63mph in 4.6sec (loss of 0.4sec)
A: Avg SA 16.7*, Avg Net SA 16.4*, 6 knock events, -0.3* avg KR

3rd Gear
CT: 182Deg.F.
MAT: 120Deg.F.
A: 42mp/h to 88mph in 10.7sec (loss of 1.0 sec)
A: Avg SA 19.9*, Avg Net SA 19.8*, 5 knock events, -0.1* avg KR


After tune-up, before tuning:
Best 0-60mph, 7.0s
Best 1/4 mile, 15.2s @ 90.0mph
AFR at WOT: high 9s to low 10s, (O2 voltage 924-963mv)
Ave effective SA at WOT: 28.9* (2nd), 28.4* (3rd)
Ave KR at WOT: -2.9* (2nd), -3.5*(3rd)

After tune-up, after tuning:
Best 0-60mph, 6.7s
Best 1/4 mile, 15.0s @ 91.5mph
AFR at WOT 12.3-12.8, (O2 voltage 850-890mv)
Ave effective SA at WOT: 26.0* (2nd), 25.2* (3rd)
Ave KR at WOT: -0.6* (2nd), -1.4* (3rd)

https://www.thirdgen.org/forums/diy-...89-camaro.html

Reddragon88gta

Thanks for replying, I have both knock counts AND KR everytime I blip the throttle/go WOT or even just give it a shot of gas to simulate passing a car on the highway. Only time you don’t get any KR or knock counts is just cruising or normal driving, no knock or KR while cruise is set at highway speeds as well. While data logging I can command knock retard just by stabbing the accelerator, low rpm as well (sub 3k rpm). It is consistent too, WOT or heavy accel it pulls up to 9 degrees, lower rpm/quick accel its 3-5 degrees timing being pulled and all the while the knocks go up until it resets. I even ran most of the 91 non ethanol gas out and filled up with 93 and a bottle of octane boost and no change.

The MAT was usually in the 90s, coolant temp around 185, blm/INT typically around the proper 128 give or take during these events, o2 sensor at WOT is over 900. The spark plugs look great and the car runs well. Honestly if I wasn’t datalogging I would have no idea it was doing any of this.

SbFormula

Exactly. I bet you they were like that right from factory. I've seen many data logs with KR being activated on that gas pedal blip or 700R4 downshift. Up to 8* KR that lingers for many seconds because Knock Recovery is slow. Hell, on a '7730 $8D, with knock routine disabled and no knock sensor, I get knock count on start-up!!!

Reddragon88gta

Thats crazy, I am learning a lot about these but that knock sensor seems overly sensitive. It is correct to assume that there is a performance loss with 9* KR? The ECM in my car is a 165 so even older than the 7730’ and has an external ESC.

SbFormula

Well!!! It depends. GM could have had commanded SA at 35* for instance. With 8* of KR, that brings SA down to 27* momentary. But with the ICM Latency tune out of whack (from factory), real SA at balancer could have been 30* instead of 27*, thus 38* instead of 35* creating the knock event in the first place.

Read this if you have time. By the end there is some interesting testing done. You'll be surprise how inaccurate SA can be on a TPI.

https://www.thirdgen.org/forums/diy-...ncy-table.html

Peace

Reddragon88gta

That is a lot of info, it sounds like SA can be inaccurate.