Been a while since I started a thread, but here goes because short of changing to another set of stock cams, rebleeding every HLA yet again, and ripping out the knock sensor wiring and redoing the harness, I am really confused.

It all started with this year's no-start. Turned out likely to have been due to a bad battery not turning over the motor fast enough or giving enough amperage to the coils, I did end up taking the head off due to really weird compression numbers. Car previously had 165 across the board only a few years ago and now I was getting between 90 and 120, and redoing the belt caused all sorts of changes. Yet, it never started, even though a motor with 120psi would start fine.

Anyway, head off, new exhaust valve guides, new seats cut on I/E, and finally put in a brand new set of BC valve springs & retainers because I was still running the stockers. In fact, I think the slow spring return rate may have actually caused flat spots on the lobes due to float. Hence why I *might* change cams.. but I still don't think that would solve all of what I have. All HLAs were bled when I put it all back together. Just bled, didn't clean them with diesel fuel because I figured it wasn't necessary as they only had about 26k km on them. Perhaps I have some that aren't staying pumped up.

Ever since I put it all back together, I've had this fairly regularly occuring knock above my ignore threshold of 3000/30% up to about 3500. Usually only .4-.7 counts that hang around for a while then go away on their own, sometimes a spike to 1.1 but below 3K with the thresholds disabled it's anything-goes. Pulled the plugs and they're flawless (but also basically new).

So I took a rawknock log and saw a MASSIVE amount of noise but it's also inconsistent, and one instance in particular caught my eye.



I've never seen rawknock like this all over the place. No balance shafts, 2.0 stock pistons/rods/crank. 2500rpm ought to be about 50-55psi oil pressure. -20inHg at 850rpm idle. That period of low activity is around 12.2s.

The odd thing is even though I don't speed up at all it just suddenly goes crazy. I've retorqued the sensor; it's a 2G on a 2G ECU. There is zero knocking sound coming from the motor itself though I'm not entirely convinced it's not the HLAs because they tick more now than before I pulled everything apart.. but the last time it ran before I did was last year. Runs strong, just went through a couple oil changes already with no flakes in the oil or filter..

Attached the log.. also have a miss I can't tune out even with injector deadtime but I haven't sent them to be analyzed or matched. Not sure how prevalent that is with a 2G ECU with only a 1G CAS.

Perhaps I just need to pull the HLAs again, fully clean them as best I can and use the other set of cams if they've got better lobes but has anybody seen this behavior before? I also redid my alt wiring and added an alt ground and a few more heavy chassis grounds. Wouldn't it be funny if I ended up making the sensor more sensitive in doing so.

A misfire can cause the knock sensor to go off.

CAS lined up properly?
https://www.dsmtuners.com/threads/cas-dash-or-dot.519420/

Yep, ignition timing validated @ 5deg btdc. Mechanical timing is recent and good with a new Gates T167 belt.

Misfire happens at idle every 10-30 seconds and no load freerev, so perhaps.. but isn't that a whole lot of seemingly random noise from the sensor? It seems to also favor cylinder 3 so, if it is PK, I would probably start with those particular lash adjusters. Misfire does not affect cruise or acceleration at all. In fact, it's quite strong and I should was right around 36-37lb/min from this B16g.

This motor was put together 6 years ago and this is the first time the head has been off since. It was not showing these issues before the head was reworked..

What I would see is as soon as I passed 3000rpm with 30% or more throttle, it would pick up .7 or, more often, .4 knock and hold it for a few hundred rpm and then disappear.

The CAS being "off" or 180 degrees out will swap the injector firing events.
Usually manifest in symptoms like a harder start or idle/lo rpm rough running issue.
The ignition timing would be unaffected .

No issues starting.. it's the rawknock values and odd knock pickup under 3500rpm ever since the headwork. Runs strong, idles fine (besides intermittent miss like others have, usually attributed to mismatched injector latency which I can't iron out by trial-and-error), and has no other audible or response issues, although sometimes i get a slight tick out of the top end like a lash adjuster. I'm leaning heavily in that direction but i'm not certain these rawknock numbers correlate to the rotation of the cams.

things done this year:

switched from ngk 7mm plug wires to taylor 8mm
two cv boots
one lower balljoint
new battery
new chassis grounds x2
new alternator ground
new alternator charge wire
new timing belt, tensioner and idler pulleys
new cam gear seals
new valve cover seals all around
bled 3g lash adjusters i've been running for years
new ngk bp7es @ 0.025 then 0.028, no change either way
BC valve springs & ti retainers, installed to 84lbs pressure
new bronze exhaust valve guides, reamed to clearance
new valve seats cut
tip heights set
head decked 0.002
felpro pt composite gasket

i've never had this kind of knock (.7 and .4 sustained) under 3500rpm before until this year and it showed up instantly on the first drive of the year.. so I'm highly suspect that the combination of the above is the fault. I just don't know why rawknock is crazytown. Even at idle..


Here's another one.. I would think if there was an actual mechanical engine issue, the sensor would not be acting like this. The dashed line is rpm, the other is rawknock1-4. ISn't this nuts?

wrencher said:

The CAS being "off" or 180 degrees out will swap the injector firing events.
Usually manifest in symptoms like a harder start or idle/lo rpm rough running issue.
The ignition timing would be unaffected .

Click to expand...

Just went through this (Installing CAS backwards) and my car fired right up and idled great, I was only getting a misfire while cruising or under load.

I can guarantee the cas isnt 180 out, I have validated and retimed the engine a couple times this year.
That last graph I posted shows almost if a switch was flipped and the sensor became 10x more noisy.
What is also possibly strange is the readings, in order of most to least noisy, are 3-1-4-2, very close to the dsm firing order.

Anybody with collapsed HLA(s) have similar experiences?
I am considering the following scenarios:
- hla that bleed down fast or dont pump and stay
- suddenly bad sensor after the last winter freeze
- bad harness wiring
- bad connections to the ecu for knock input
- really, really bad mechanical noise.. no balance shafts and poly mounts

I am just trying to gauge whether anybody has seen this type of rawknock reading before on a 2g ecu i just started logging it, too bad i didnt last year.

nobody yet has any weird shit going on with rawknock1-4 like me, i presume. god i feel special.


going to try my ngk wires, check condenser, check ecu for weirdness, even knock wire to ecu for proper sheath grounding
something is weird and that 12 seconds-then-flip is really getting to me.

Thought I had it straightened out but it came right back. Changing back to NGK wires, zip tied the knock sensor wire to the harness split loom, made sure the alternator charge wire wasn't close to the alternator..

Another instance of what the hell.. holding RPMs steady and suddenly the noise level plummets when I get off the throttle a little, then entirely, and the noise stays low after I get back on throttle and only comes back to nuts level after I shift.



Then, on acceleration, it decides to plummet in noise too. Driving. me. nuts.


Oh, and this one is real good. Noise plummets on its own in the middle of a pretty steady cruise (7 seconds to climb 500rpm):

I'm reviving this because it's my own thread

My Fidanza 3.2 4-puck kevlar clutch disc gave up a spring around this time. I wonder if the spring being so loose was causing this issue right before it let go.. I guess we'll never know now.