When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
MS3x catching rod-knock before I can hear it? For Science. Maybe.
"Wow, I'll pull some timing when I get home..... wait a minute, just one cylinder?"
- Me, $950 dollars ago.
OK, full story. Due to my own stupidity my engine has had a couple bouts of oil starvation. Stock '05 LS, 148k miles, 10W-30. Removed the cat because I melted it, and discovered I must have valve guide seals on the way out. It doesn't really smoke but you can definitely faintly smell it. Which is where all my oil went apparently (consumption is generally OK). While tuning high vacuum areas off of Brain's (read: our ) basemap for my stock '05, it went dipstick-bone-dry as I moved on to doing several WOT pulls on the street. Didn't know until I took an offramp so hard it gurgled air (I admit, not the first time..); that's when I discovered my fuckup. When changing the oil it was glittery but nothing crazy. I pushed the check valve on the filter and dumped it out. Very glittery:
Later I cut the filter open; some shavings. 10mm for scale. I admit - I expected worse:
I dialed in my knock window settings prior with Reverant's "75kpa start, +20 duration" method (fantastic!).
Then eyeballed the per-cylinder gain values to within error. So far it behaves well!
Knock config for reference:
I scatter-plotted my knock readings to set the threshold values, and it doesn't tolerate a damn thing.
As soon as it hints at detonation, it gets squashed immediately.
For reference, detonation has to spike MUCH higher for me to hear it by ear, such as the following.
(Note - this photo was before I had the thresholds configured properly)
So, back to present. I started to notice faint peaks (see: top of this post) on just Cylinder #1. I don't have a screenshot handy, but when I first saw it, it was just happening ONLY between 3300-3450 RPM, and ONLY at ~50-75 kpa. Thing is, knock retard pulled up to 6* of timing. It didn't stop it. The reading is nowhere near high enough to be genuine spark knock IMO? Aside from those datalogs, it continued to run/drive perfectly. Hell, even made it on a ~350 mile trip + idling while I slept in it at a track. It hurts, thanks. 6+ weeks later, and it's finally starting to read much higher.
After total of 1k miles since last change, changed again.
Same as last time: oil was OK, pushed the check valve, dumped - bronze glitter. Lost this one before I could cut it open.
By now, I still don't hear rod knock, maybe a faint tapping sound at 3.3k. But mostly the exhaust just gets super raspy at that time.
Here is a recent datalog... Be warned...
As it's declining, acquired clean '03 junkyard engine.
Taking the opportunity to rough-in a FlowForce 640 VE table, with the donor engine in the trunk... Literally.
#1 is as of late:
Again - even now, I can't hear a thing. Runs/drives perfectly.
But the writing is definitely on the wall. It'll hopefully turn into a rods/pistons only rebuild while I run the donor.
So. Anyone ever see predictive failures from a knock sensor?
Donor assembly stripped down for refreshing. Got all the front oil seals done. Just have to scotch some areas to reassemble coolant; rear seals and bearings; covers; then it's ready for drop-in.
Originally Posted by Mudflap
You should send the oil to Blackstone. You can't tell just by looking at it.
Detecting knock isn't that clear cut.
Respectfully, I'm past the point of this.
With the current oil in the engine, I could send it off (and I'll try to) - but I'll most likely be sampling it with the donor installed, so... I could visually see what was damaged and where. I do want to know the composition though. For how much bronze/copper coloring I'm seeing, surely the bearings aren't that far gone and still silent..?
Finished refreshing the donor and pulled the current one. Shot of the valves which immediately stuck out. I didn't drain it, so the oil I will send out when I remember 😬
Regarding the valves... I'm still fighting a mix of heat soak with Flow Force 640's and probably inaccurate IATs (stock location, with AC on out here I easily hit 160*F). So posting a tune isn't 1:1 with how it's been ran the last couple hundred miles. These injectors take patience for sure. But generally, fuel wise I target:
15.0-15.3 below 70 kPa
Low 14's by 75 kPa
12.5 kPa at WOT
Above 4500, cruise areas become 14.7
Unhealthy (but running) engine pulled. White exhaust valves. No coloration that I can see. Donor from an 03, re-sealed.
What's your timing at 75kPa at 2000rpm? Let's say it's 25*. In your knock window settings, set the start angle for 2000rpm to 25*. 20* might be a little bit on the short side if your start is 25 - I like to be able to listen up to 15-20 ATDC, so in that case set the duration to 45* for 2000rpm. Repeat for the other RPM areas.
What's your timing at 75kPa at 2000rpm? Let's say it's 25*. In your knock window settings, set the start angle for 2000rpm to 25*. 20* might be a little bit on the short side if your start is 25 - I like to be able to listen up to 15-20 ATDC, so in that case set the duration to 45* for 2000rpm. Repeat for the other RPM areas.
Ok, thank you for the reply!
I tried a log with those changes, and it just looks like i'm getting engine noise that increases with RPM.
It even pegs out at 100% at higher RPM.
This can't possibly be knock, as I already have my timing conservative, and the MS3X pulled an additional 10 degrees when it detected "knock" without any discernible change in the noise level.
What should be my next step? Reduce my per-cyl gains?
Why are you quoting 6 gain numbers, ie for a 6 cylinder engine? Only the first 4 should be active. Did you specifically tune this numbers or did they just came up with a basemap?
Why are you quoting 6 gain numbers, ie for a 6 cylinder engine? Only the first 4 should be active. Did you specifically tune this numbers or did they just came up with a basemap?
Sorry, guess I should have mentioned this is on my other car, not my Miata.
BMW M52B28 motor in an E30.
These numbers came from a gent on facebook running the same engine and knock module. Think he just used some trial and error though, considering there wasn't anything beyond the default knock window.
I've since swapped the engine out - and when I get the chance to tear it down I'll definitely post such, as I'll need help reading what I find.
In the meantime, the new engine really wants to knock in 75kpa cruise.
Tune is attached. First thing I did was check base timing on the pulley, and that looked good, even better after a tiny adjustment (reverted; 5.5 offset, 7 offset looked better). I'll pull a plug and check the cylinder TDC this afternoon.
Note that I'm now running a 1:1 of @curly 's VVT table from HERE. The car feels much better with this. In fact with 0* VVT advance < 80 KPA, it feels like MAP just wants to go right up to 80 KPA very easily.
Should I just accept that VVT in cruise areas is lowering my "<100kpa advance" thresholds due to the additional air taken in? I didn't stand a chance running the TruboKitty 05 map, readings are much higher (@Braineack mentioned NB2 testing to me in the past, albiet startup). Ultimately at 18* in that region it consistently shuts right up. Just seems low.
Originally Posted by O R X O N
Hell, even made it on a ~350 mile trip
I have another one of this exact trip coming up this weekend. Unless anyone has any insights, I'll probably re-enable EGO correction and do some large datalogs of nerfing VVT down low to see how soon it'll start taking ignition timing; and averaging some fuel flow values. It's a pretty flat drive.
If there's mid-tens-% of fuel economy to be gained by pulling VVT to 0* < 80kpa then I'll deal with that. But for now I sure am spoiled to the responsiveness.
Just note that from what I've seen on the dyno, the VVT advance is just a relative number. My car for instance idles (0% duty and/or solenoid unplugged) at 20* VVT advance. And where others max out at ~38-40 degrees max, I was finding gains at 46, which is roughly where my 10 year old Walbro 190LPH gave up the ghost on the dyno Friday night. My point is, true VVT gains require a private perfectly flat road for virtual dynoing, or an actual dyno to actually tune. It'll be small gains, but once done, total gains can be 5-10hp and tqs for NA engines.
Just note that from what I've seen on the dyno, the VVT advance is just a relative number. My car for instance idles (0% duty and/or solenoid unplugged) at 20* VVT advance. And where others max out at ~38-40 degrees max, I was finding gains at 46, which is roughly where my 10 year old Walbro 190LPH gave up the ghost on the dyno Friday night. My point is, true VVT gains require a private perfectly flat road for virtual dynoing, or an actual dyno to actually tune. It'll be small gains, but once done, total gains can be 5-10hp and tqs for NA engines.
Gotcha.
Finished this weekend's journey. This is the part where I admit I was wrong (no **** sherlock): zeroing out the VVT table down low actually yielded lower MAP values at a given TPS. Pumping losses/EGR effect, etc?
I haven't looked at fuel consumption, as I was fighting 75kpa knock readings (real knock, mind you) between 2500-3700. I added and removed small values to the 70ish kpa range on the VVT table to try and resolve this (pump 93 mind you, not even 91). Haven't gone back to analyze yet. What was confusing me was to quiet it down, required less timing at 75kpa than 90+kpa. It seems off to me that part throttle would knock so much easier than WOT? Insight welcomed here.
I softened up timing from 65kpa and up, and just lived with it for now. I have an email out to Savington on EFR availability, and I'll just run this lighter timing until it's dyno time down the road.
Hey Curly, sorry it took me so damn long to get back lol.
So that engine was trashed and I swapped in another one. It's been perfect for the last 15k + miles. Developed crank wobble. I just put a Fluidampr/Crank Saver on after finally getting my Kraken kit + EFR 6258 on. I'm seeing something quite similar but only cyl 2 and 3 but it's usually not so bad. From day of the EFR I've been fighting some serious oil leak issues that are intermittent. After adjusting my oil drain to be a straighter shot to the pan, it stayed bone dry while putting through town and even romping hard at empty intersections. I had a 4 hour trip for work lined up and decided to give it a stress test. Everything was fine on the first 3 hours or so.... Some regions where things were getting pretty noisy on those two. I was using cruise control going up and down hills for miles-and-miles to try and get noise thresholds ironed out.
Then when I crossed into Virginia (because of course), I had to disengage cruise, much lower speed limit, and all hell broke loose knock noise wise. A lot of this is 60-70 kpa.
Datalog is ~120mb.
Made it to my destination and of course: See some knock while dealing with oil issues - not what I wanted to see...
Things started to get pretty noisy.
Revisited this thread while making some scatter plots. Going to cautiously put home. Drain line is a flange and hose supplied by Kraken. Haven't seen others with this issue. Suspect it happened when cruising around 4k when the turbo is spooling for a hot while but not enough to leave vac (recirc open; no positive pressure to equalize against the oil seals). Goes away almost immediately at 85kpa+. Knock retard pulling up to 5 degrees doesn't seem to stop it. PCV is factory to IM and breather filter to atmosphere.
Last edited by O R X O N; 03-10-2022 at 04:03 PM.
Reason: Clarify
Made it home last night. Datalogged all the way, but laptop died just a half a mile from my house. For your own entertainment, tune as of last night and datalogs attached.
Stopped at rest stops regularly to check my EFR inlet. Stayed dry the entire time. Wiped it with a shop towel and there was a very thin film at most (which IMO was just residual from when it gushed out on the trip out there). Not a clue what caused it to gush out in the first place. Wondering if I was going downhill and jabbed on the brakes causing a pool in the CHRA cavity? Maybe?
On this trip as you'll see from my tune settings I set the knock thresholds pretty aggressively low. Per-cylinder control is enabled. I based it on scatter plots filtering out everything that wasn't WOT. Surely 70kpa, 13.8-14.5 (!!!) AFR, and 25* of timing shouldn't be enough to cause knock like this?
Above is something I definitely don't want to ignore. Those spikes are too high above ambient in my limited-experience opinion.
Open to any opinion/comments. Cheers lads.
Edit: Triple-checked my timing (SuperMiata 36-2) - all good. Ordering a catch can setup now - will crack open my IM to inspect/clean, replace the PCV valve, and see what's-what with a catch can.
Plz forgive my triple-posting. Not trying to spam but just want to relay findings.
First up... If you knock plots look like the right - respectfully, fix your wiring. My knock readings post-EFR-install were crazy high at low RPM - showing almost 50% as a baseline in some cases:
A post HERE (and the reply) sated my concern over this. However it's definitely something that needs to be addressed... My wideband died Monday and my readings went back to looking flat and low at low RPMs. I replaced it with a genuine Bosch 4.9 LSU and my knock samples at low RPM are back up to high. I suspect my UEGO ground - which is used to heat the sensor - is referencing the sensor ground. I'll be fixing that.
I installed a Mishi dual port catch can last night (TODO: drilling the valve cover; teeing the output to both PCV and finding whatever check valve has least resistance and putting that between inlet and filter). For testing theory that oil vapor was causing this cruise knock, I teed the dual ports into a single inlet. One side IM, one side PCV. Breather has a filter and VTA. The can was bone dry after ~10 miles... but still knocked in cruise.
Went from my VVT map posted in my last attached Current Tune, to the DIYPNP VVT base from the 01-05 map. Drove smoother IMO lol. But, still knocked in cruise.
Tried the TruboKitty VVT table as well. Still knocked in cruise.
Used cruise control to eliminate VVT from the equation (?), commanding between 0-35* VVT advance. All of the 5* increments still knocked in cruise in some capacity.
Eliminated crankcase gasses as a cause by pulling over on the shoulder of the freeway, disconnecting catch can, jamming an extension in 1/2" line to block off the IM port. With the PCV and breather venting to atmosphere - no crankcase gasses piping into my intake - I still knocked in cruise on the way home.
A suggestion to eliminate driveline harmonics was given to me. I went away from Reverants method covered above, and set my knock window to start at 0* (TDC) and listen all the way through to 20* ATDC. Quieted my readings down a good bit - but, you guessed it - still sampled what is, IMO, genuine knock... even if at a low CFM. All above trial and error aside I think I'm going to try to get some audio from the knock sensor itself (or an external mic) and sync it with datalogs in a video. I really am at a loss here.
SOLVED. The solution ended up being a shitton of Seafoam. "One ounce per gallon of fuel" nah - more like 65% seafoam 35% 93oct. The root cause of both issues (Turbo oil + Knock in Cruise) is probably a feedback loop lol.
First: I probably made this worse when I put the Turbo on. The line to my wastegate cracked. I replaced all my lines and accidentally bypassed the VTCS solenoid which I trigger when in closed-loop idle. It was now VTCS at all times in vac... Probably also in cruise. I addressed this first. I know it wasn't a cause because I have datalogs of this cruise knock from *2 ******* years ago*. It's been a while...
I mentioned crank wobble. I put a Crankshaft Saver and Fluidampr on. It helped some but it still wobbles. To rule out mechanical noise ("2 and 3 are on the same phase"), I was under the car rotating my crank/moving the sway bar to re-torque it. I hit a compression stroke. Not 2 seconds later there was no resistance...?! I immediately had a flashback to when I put the '03 engine in and turned the key. The starter wasn't loading up properly. My buddy (TM) told me, "just lay on it - it'll start." And it did. Starter always sounded fine after that. Chalked it up to the engine having sat for a few years and never questioned it again. Until I felt that.
I also remembered I was on BK5RE platinum plugs. I ran to the parts store, low on fuel, and bought some copper BK7RE and Seafoam. A lot of it. Put a splash in the crankcase and put 2 cans in my tank. Way more than recommended. Drove some, didn't see enough white smoke for my liking. 3 more bottles of Seafoam in. Some light smoke. Upped my req_fuel by a tiny bit and putted home. Heat soak, shut down, etc. Went out, putted until the first hint of fuel starving. Pulled over > gas can fill > drove more > very low again > filled up the tank, aaaand hit the highway!
Cylinder 3 was always the worst offender, and I liked what I saw, so I reverted my knock windowing back to the very aggressive Reverant method: Two baselines because I changed windowing mid-datalog.
For reference...
I've since gone through the entire tank and put fresh 93 in.
Now it behaves as I would expect. Spools UNGODLY better than it ever has - even still without EBC wired in. Knock readings doing exactly what I'd expect!
finally, some actual ******* knock. /s
Number 3 no longer spiking in 75kpa region! All-Cyl spiking in high MAP. Fixed!!
So, what was happening? From sitting for years in a junkyard, valves probably had gunk on them and I know they weren't sealing. They were probably exhaust vales. When romping on it, enough cool air comes in to prevent knock. When cruising, exhaust gas was likely leaking back in. This would've been made SO much worse by the VTCS valves not opening as they were supposed to when not actively in CL Idle. This intermittent spark knock was likely compressing the hell out of my crankcase, causing my turbo to spew oil. Which definitely was not helping the issue.
I was also running improper heat range plugs and cruise in somewhat-lean conditions (per your guys' posts). My crank wobble is probably caused by the timing gear not seating all the way back, caused by ferrous material from the keyway getting worn down, and I will address it next time I tear that far in.
07-02-2020, 01:24 PM
MS3x catching rod-knock before I can hear it? For Science. Maybe.
) basemap for my stock '05, it went dipstick-bone-dry as I moved on to doing several WOT pulls on the street. Didn't know until I took an offramp so hard it gurgled air (I admit, not the first time..); that's when I discovered my fuckup. When changing the oil it was glittery but nothing crazy. I pushed the check valve on the filter and dumped it out. Very glittery:
0
