The Top DSM Community on the Web

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. Log in to remove most ads.

Please Support RTM Racing
Please Support ExtremePSI

ECMlink RESOLVED - Crazy knock under part throttle/low load/open loop

This site may earn a commission from merchant
affiliate links, including eBay, Amazon, and others.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

dwb

Proven Member
813
560
Sep 9, 2021
Broomfield, Colorado
I've attached a data log of my commute to work this morning.

I've been experiencing lots of knock under low load, part throttle, ~2k-3k RPM, and usually open loop. In this log, knock continues into closed loop for a bit. I have to release throttle and roll back into it for knock to stop, and even then it doesn't always stop. This same behavior happens every time the engine is cold, up until either closed loop or full operating temp. Once at full operating temp I primarily see knock under higher load/RPM/throttle as expected which you can see later in the attached log.

Stock fuel system, 91 octane. Stock MAF. Mostly stock car except ECMLink and MBC. Running stock 2g timing maps on a fresh 6bolt with 2g pistons. No other changes to ECMLink from a stock 1g config. No boost leaks, no exhaust leaks. Lifters do not tick (3g revision). I do notice some extra engine vibrations around 3k RPM, which I speculate might be a very slightly bent CAM or sprocket (these CAMs came from a 1g that was rolled over/landed on top of the hood and broke one sprocket). I'll be swapping CAMs and sprockets in the upcoming week(s) to verify this speculation.

I'd appreciate it if someone could take a look at the attached log and let me know if there is something amiss that can be adjusted. I'm an ECMLink noob so please forgive me. Hopefully I've captured the pertinent fields and values.
 

Attachments

  • log.2023.05.17-01.elg
    418.5 KB · Views: 52
Solution
Oh yes, the mod-wish-list goes on and on. I'll take intake temps into consideration. Definitely a wideband. It is currently stock with the exception of the mod list in my profile - not much.

I'm just super stoked with how the new HLAs have improved the overall feel, not just throttle response (which is waaay better btw), but it's less 'buzzy' too. I would have replaced them long ago if I had known!

No more crazy knock! You guys ROCK! @danl @We're on Boost @Stapl3 and a few others too, I very much appreciate your time, effort, and patience. I've learned a lot in the process. You guys have given me some good pointers to continue playing with on my own (or a new thread). What a great community to be a part of!

For...
Not yet. What would I be looking for on the plugs?
If it’s real knock (pre-detonation), your plugs will look like they’ve been drug through gravel.

I once had phantom knock 43 counts maxing out the sensor. Weird but running sea foam through the intake fixed it 100%. Don’t ask me why 🤷‍♂️
 
I once had an ecu with a bad knock board that would do as you described, worth taking a look at if you don’t find the culprit. I’ve also had mine pick up random rattles and what not.
How did you determine it was the ECU?
Mine was an Eprom I bought used, so I had no prior knowledge of it's function. It needed CAPs and some board work on the idle circuit. Not sure if ECMTuning looked at any other circuits when they repaired and socketed it.
 
Absolutely *something* is wrong, but the low load/low RPM knock goes away after it's up to temp. I seem to remember the old TMO logger software did log knock count and possibly voltage. Not sure if that program would still work on an ECMLink chip, though. I still have that cable and software however...
Knock sensor is MD141510, which is the 1g PN.

Yeah, the first knock that looks "normal" to me is the knock at about 354 seconds where the coolant is at 193 deg, throttle is 100%, and the knock retard is only 1.4 degrees. All the knock before that looks like something that shouldn't be there.

Here in this log (pic below) my engine was stock original (not rebuilt) but had some external mods like GM MAF and DSMchips ecu.
I was logging with ecu+.
My knock sensor was new OEM at the time.

Compare Knock Voltage to Knock Sum (the bottom 2 charts). That's the idea of this.
Everything is fine as long as the knock sensor voltage is below about 0.38 volts.
Those sharp voltage spikes that start happening are knock. They are picked up by the ecu and show up as "Knock Sum" about 0.1 seconds later.
My "knock sum" here got up to 24 (which you could read if you have the software to work with the log). That's not degrees. It's what we see as RawKnock in ECMlink.

It turned out my problem was that I didn't have a boost gauge hooked up yet, and my boost was hitting 25 psi without me knowing it. 25psi wasn't in the plan! I found a blown off boost hose to the boost controller. Fixing that and limiting my boost to 15 psi was all it needed to get Knock Sum down to flat 0. Didn't make any changes to fuel or timing maps. Those stayed the same as what DSMchips put in it.

So this was real detonation knock that needed to be brought under control.
I don't know what is going on with your car but if you could somehow log the actual voltage in the wire that goes from the knock sensor to the ecu, you could then see if the voltage has these spikes and goes over about 0.38 volts. I think you would need a logger with a high input impedance so it doesn' t corrupt the signal. The ecu+ had input impedances of 1 million ohms (very high) and it logged good accurate voltages.

You must be logged in to view this image or video.
 
I checked an old TMO log and it only shows knock count (0-40), not voltage. Probably not worth attempting to log with it.
 
I checked an old TMO log and it only shows knock count (0-40), not voltage. Probably not worth attempting to log with it.

Yeah. Well, it would be interesting and maybe helpful if we knew what the actual voltage was on that input (knock sensor to ecu) in those areas like from 156 to 162 seconds in the log. That's 6 seconds where the RawKnock is continuously stuck all the way up at 43, the max.

Maybe it's a bunch of spikes. That would probably be from something mechanical bashing around in the engine, making a big ruckus. Bang bang bang.

Or maybe it's not spikes. Maybe it's just a constant high voltage that is steady at some number like, oh I don't know, 4.69 maybe?
I say 4.69 because on my 1g ecu, when I have an input that is not connected to anything, it logs as 4.69 volts in ECMlink (the raw value).
Weird, I know. And my multimeter confirms that there really is about 4.7 volts there.
That's if the ecu input is not connected to anything, or if it's connected to a dead thing.
So if we saw a voltage reading like that in those periods where the ECMlink log shows maxed out RawKnock, you could be pretty sure that it's an electrical problem. Could be an intermittent electrical connection someplace, or a problem inside the ecu. I doubt it's the sensor itself because yours is pretty new and is OEM.
 
Last edited:
I swapped camshafts but it made no difference. Fortunately everything under the timing cover is in excellent shape and completely dry. Cam caps and journals all looked good. I haven't yet tapped on the block with a wrench or whatever to check the sensitivity of the knock sensor. Nor have I tuned it out with ECMLink.
Here's a pic of the spark plugs, they look good to me.
You must be logged in to view this image or video.

Who needs special tools? Not me! Here's my eccentric timing belt tensioner pulley "tool" ROFL
You must be logged in to view this image or video.
 
Here's a pic of the spark plugs, they look good to me.
Good pic of the plugs. Looks like you might have done your own resize of the original before uploading, to keep maximum quality. 🙂
Anyway they look ok to me too. When people start raising boost above stock they usually switch to BPR7ES which is one range colder than the 6. If you are still on the stock 1g MAF you might not need 7's yet.
7's wouldn't hurt though, and a gap of about .024 inches if you are going above stock boost levels.
Anyway these plugs don't look bad to me.
 
I've tuned out knock at 3500/25%. Made things a lot better but still seeing phantom knock above that. I did get it to log a bit of knock when just free revving in neutral. I probably could dig up and post that log if it means anything.

I mentioned previously I have some engine vibrations that are apparent from ~2500-4000 RPM, but also noticeable at idle. I wish I could get a clear video. I'm kind of stumped why I have this vibration but I feel that's where I need to investigate. I first noticed it when I eliminated the balance shafts (OEM stub shaft and removed front shaft) and installed a fresher head probably around 10 years ago. I assumed the vibration was just the nature of a BSE, but now I have my doubts.

Since then I preformed an entire engine rebuild, which included new water pump and timing components. I had the crank checked for balance and straightness when they polished the journals. New pistons but re-used the rods. I failed to weigh each rod/piston combo before install to make sure they were balanced so that's looming in the back of my head. However, the vibration is exactly the same before and after the rebuild.

I've swapped cams, new harmonic balancer, new alternator, new AC compressor, clutch has been replaced, all had no affect on the vibrations. The flywheel is original/OEM and has been machined several times. I doubt it's ever been balanced. No idea if that could be the culprit. Crank sprocket was reused but looked to be in good condition. What else in the rotating assembly could cause vibrations?

One last thing that I've been thinking about are mismatched injectors. Currently running OEM 450cc injectors, but a combo of both 1g and 2g versions. Long story short, I currently have three 2g 450cc and one 1g 450cc. Could that mismatch cause odd vibrations? I do have a new set of FIC 1000cc that I plan to install and tune soon, so that would tell me for sure.
 
So it's still pretty awful then right? Anything above 25% throttle at 3500 rpm and it knocks a lot still? I would imagine that your load factor at 25% throttle and 3500 rpm is not much. You shouldn't have to baby the throttle at 3500 rpm to keep it from knocking.
On the other hand, I wouldn't go heavy throttle until after the engine has been warmed up for a while anyway. But there still shouldn't be heavy knock.
Anyway, yeah, if you could post a log of driving conditions similar to what you showed in post #1, it would probably help a lot to see what you changed in the tune and how much difference it made in the knock.
 
Last edited:
So it's still pretty awful then right? Anything above 25% throttle at 3500 rpm and it knocks like crazy still? I would imagine that your load factor at 25% throttle and 3500 rpm is not much. You shouldn't have to baby the throttle at 3500 rpm to keep it from knocking.
Anyway, yeah, if you could post a log of driving conditions similar to what you showed in post #1, it would probably help a lot to see what you changed in the tune and how much difference it made in the knock.
Sometimes yes, I have to back off throttle and give it another go. Better since changing knock settings but not fixed. I'll take/post another log next chance I get.
 
You check all your grounds? Idk it's probably something stupid that's been overlooked? Some people throw grounds in from the engine, the trans, the exhaust, pretty much anywhere with a non critical bolt that could build up a charge gets grounded.
 
You check all your grounds? Idk it's probably something stupid that's been overlooked? Some people throw grounds in from the engine, the trans, the exhaust, pretty much anywhere with a non critical bolt that could build up a charge gets grounded.
Yeah, all good on OEM grounds. Exhaust...I thought maybe a rattle was part of it. I clamped down the heat shields until no more rattle but no joy.
 
Just in case, super basic thing... Did you tighten the knock sensor properly?
And did you try to "relocate" the knock sensor? (Not replace the sensor with another one) if you doubt that the sensor is actually detecting some mechanical noise, maybe you should try to relocate the sensor to somewhere else that is less noisy to see if that would affect the knocking level/pattern in log. This may narrow down the issue.
 
Just in case, super basic thing... Did you tighten the knock sensor properly?
And did you try to "relocate" the knock sensor? (Not replace the sensor with another one) if you doubt that the sensor is actually detecting some mechanical noise, maybe you should try to relocate the sensor to somewhere else that is less noisy to see if that would affect the knocking level/pattern in log. This may narrow down the issue.
I didn't have the ability to put a torque wrench on it so I guessed the torque with an open-end wrench. You could be on to something if that really does make a difference. I'll have to go look in the FSM to see what the proper installation procedure is. No, I have not tried relocating it, it remains in the stock location. A good thought and something I will consider but I'm not sure where another suitable location might be. Here is a pic after I torqued it down, not that this says anything but I like pictures. BTW, I did remove the engine enamel underneath the sensor and made sure the surface was clean.

Anyway, yeah, if you could post a log of driving conditions similar to what you showed in post #1
You got it, see attached. Same drive as last time. There's a good one that pulled 15.1deg at 227 seconds that has to be phantom knock. It hits right at the 3500RPM activation. Gave it a couple WOT pulls near the end to see some actual knock.

I once had an ecu with a bad knock board that would do as you described, worth taking a look at if you don’t find the culprit. I’ve also had mine pick up random rattles and what not.
I still have my original stock ECU, a non-eprom, that I could swap and log with my old TMO logger to see if this phantom knock still exists. Besides uploading 2g timing maps to support the 2g pistons/compression, and now knock sensor control, I've done very few configuration changes with Link.

You must be logged in to view this image or video.
 

Attachments

  • log.2023.07.13-01.elg
    446.9 KB · Views: 27
Just in case, super basic thing... Did you tighten the knock sensor properly?
OK, I give up. What's the proper procedure? I can't even find it in the 1g FSM and googling I found "hand tight + 1/4 turn" or "just tighten it" or up to 17ftlb. No way am I getting a torque wrench on there even if I had a 24mm slotted socket.

Very possible I over torqued it (I tried to go easy on it), so might be best to loosen and reset it.
 
I didn't have the ability to put a torque wrench on it so I guessed the torque with an open-end wrench.
I understand, it's PITA with intake manifold. The torque would be around 17 ft/lbs. But it doesn't really need to be exact. It would work if it's not too tight/loose, that's what I meant. A common mistake I have seen is over tighten and the sensor gets broken at the thread part.
No, I have not tried relocating it, it remains in the stock location. A good thought and something I will consider but I'm not sure where another suitable location might be.
It's just a test in case if the sensor is detecting some mechanical noise at the stock location and to see difference on signal in log as you can't control the sensor sensitivity through the stock ECU, so just somewhere different (farther) from the stock location but you should avoid on the cylinder head since it may detect more noise from cylinder head components. Maybe somewhere on the intake manifold may be fine.
 
I still have my original stock ECU, a non-eprom, that I could swap and log with my old TMO logger to see if this phantom knock still exists.
This would probably be a really good thing to try since you still have the stock injectors and MAF in the car.
I think if you do this, limit it to trying to reproduce the low power knock that seems like "Phantom" knock.
Don't use it at high power - because as I remember the 1g timing maps are more aggressive than the 2g, and more aggressive timing is not what you need right now!


You got it, see attached. Same drive as last time. There's a good one that pulled 15.1deg at 227 seconds that has to be phantom knock. It hits right at the 3500RPM activation. Gave it a couple WOT pulls near the end to see some actual knock.
Oh I see, it looks like the only change you made to the tune is changing the knock sensor activation RPM from 2200 to 3500, and the knock sensor activation TPS from 0% to 25%.
The timing maps and the fuel stuff all look the same as before.
So the low speed knock is still probably there same as before, we just aren't seeing it because it's "turned off".

Anyway I'm not coming up with anything much in the way of ideas! But I'm pretty crappy with DSMlink.
The way the car is now, it seems to want the timing down around 15 degrees, even later in the log when it's warmed up. Timing numbers that are more like what you have in your "Min Octane" timing map.
When it's warmed up it doesn't do the "43 RawKnock" like it does earlier in the run, the knock retard is less, and it happens when the MAF hz is getting up around 1800 to 2000 hz. And 2000 is too much for the 1g MAF as far as I know. I switched my car to a GM MAF when it got to that point, with 850cc injectors, Wally 255, and Fuelab FPR (with everything else still pretty stock, even the stock sidemount intercooler). SD would be even better. I don't necessarily think that the MAF running out is what causes that knock though.

It would be nice to have a wideband and a boost gauge that can be logged, or even just a logable MAP sensor without a gauge. For that matter even my first wideband was just a logable controller and sensor with no gauge and it was better than the stupid one I've had for the last few years!

Your new injectors, fuel pump, and FPR look awesome. I may have missed if you said it somewhere, but that 1g MAF has to go!
Oh also, in your log I couldn't find the item "Learned Octane" which might be handy to have logged, so you know which timing map it's using more.

Do you know what your boost gets up to in these runs?
 
Last edited:
I understand, it's PITA with intake manifold. The torque would be around 17 ft/lbs. But it doesn't really need to be exact. It would work if it's not too tight/loose, that's what I meant. A common mistake I have seen is over tighten and the sensor gets broken at the thread part.
OK, so I was pretty close in my research at 17ft/lb. I could sense that overtightening it would damage the sensor, so I went pretty easy on it. I would have torqued it properly when I had it on the stand but a large slotted socket was not in my toolbox. I think I'll crack it loose, and then try hand-tight + ~1/4 turn before relocating it.

This would probably be a really good thing to try since you still have the stock injectors and MAF in the car.
I think if you do this, limit it to trying to reproduce the low power knock that seems like "Phantom" knock.
Don't use it at high power - because as I remember the 1g timing maps are more aggressive than the 2g, and more aggressive timing is not what you need right now!
Pretty easy to do, I'll just turn down the boost so I don't accidentally get a little overzealous.

Oh I see, it looks like the only change you made to the tune is changing the knock sensor activation RPM from 2200 to 3500, and the knock sensor activation TPS from 0% to 25%.
The timing maps and the fuel stuff all look the same as before.
So the low speed knock is still probably there same as before, we just aren't seeing it because it's "turned off".
That's exactly right, I haven't messed with fuel yet, and I loaded the 2g timing maps in April.

The way the car is now, it seems to want the timing down around 15 degrees, even later in the log when it's warmed up. Timing numbers that are more like what you have in your "Min Octane" timing map.
91 octane so I would guess that's the reason for it matching closer to the min octane map. Maybe some 93, I'm sure it's around somewhere, would give some better results.

When it's warmed up it doesn't do the "43 RawKnock" like it does earlier in the run, the knock retard is less, and it happens when the MAF hz is getting up around 1800 to 2000 hz. And 2000 is too much for the 1g MAF as far as I know.
That's what I've noticed too, and I thought maybe it has something to do with open loop vs closed loop but now I have my doubts. Could just be a mechanical rattle that quiets down after it gets warm. I still see "phantom" knock after it's up to temp, but not as bad as cold. Hard to believe that a mostly stock setup is overrunning the 1g MAF. I do see that in the log, though. I may have to keep my eye open for a 2g MAF.

It would be nice to have a wideband and a boost gauge that can be logged, or even just a logable MAP sensor without a gauge. For that matter even my first wideband was just a logable controller and sensor with no gauge
A WB is on my list. I was hoping to use the stock o2 location and simulate NB with Link.

Your new injectors, fuel pump, and FPR look awesome.
Thanks, I think so too! I've already installed the FPR and will do the pump next. Hoping to get some of these gremlins worked out before the injectors.

Oh also, in your log I couldn't find the item "Learned Octane" which might be handy to have logged, so you know which timing map it's using more.

Do you know what your boost gets up to in these runs?
I'll look for that setting and add it to the log.
I hit about 14PSI, give or take, might be a tad too much.
 
Do you know the oil pressure at the cylinder head at 1000, 2000, 3000 rpm?

Exactly what part # are your spark plugs?

Do you have stock harmonic balancer? Can you verify #1 piston is at tdc when the crank pulley says 0? Can you double check base timing is 5 deg at idle with timing connector grounded, or using link?
 
Do you know the oil pressure at the cylinder head at 1000, 2000, 3000 rpm?

Exactly what part # are your spark plugs?

Do you have stock harmonic balancer? Can you verify #1 piston is at tdc when the crank pulley says 0? Can you double check base timing is 5 deg at idle with timing connector grounded, or using link?
Sorry, I do not have an aftermarket oil pressure gauge.
BPR6ES plugs
I set the base timing at 5* with an OEM balancer, but I have an aftermarket on now. I had to make my own timing mark and used the OEM balancer as reference. I can definitely double check base timing and where the mark is at TDC.
 
Support Vendors who Support the DSM Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Innovation Products Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications MyMitsubishiStore.com RixRacing RockAuto RTM Racing STM Tuned

Latest posts

Build Thread Updates

Vendor Updates

Latest Classifieds

Back
Top