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

[dwb]

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.

 

[danl]

Stock balancers can/do separate and slip. An easy test would be set base timing to 0 deg (not 5 deg) and drive it to see if you still have the same issue. Idle quality should suffer a bit. Otherwise you can verify TDC using TDC piston stop and degree wheel to make sure your pulley is correct: https://www.summitracing.com/parts/pow-pow101330

Just saw a picture of your plugs, sorry for not seeing earlier (I was on my phone).

You are going to want to check oil pressure to your head. This sounds a bit like ticking lifters and you can't always hear them. Low cylinder head oil pressure (due to milled head) can cause these issues. I know you said your car is mostly stock, but just to clarify, still a 1G 6 bolt head correct?

This is worth having, just to verify your oil pressure at oil filter housing and cylinder head is good, its only $30. https://www.summitracing.com/parts/sum-g2934

 

[danl]

I am looking at your log now. It looks like most of your knock is before the coolant temp reaches 150 deg. After the engine reaches 180F the knock is significantly less. This really does mimic the HLA's ticking when the engine is cold (even if you can not hear it). Step 1 would be to check oil pressure to the cylinder head at the stock turbo feed location (1/8" BSPT plug). If it is low, there are a couple things you could do. If oil pressure is fine, step two would be to install known good HLA's.

What weight oil are you running?

 

[dwb]

Stock balancers can/do separate and slip. An easy test would be set base timing to 0 deg (not 5 deg) and drive it to see if you still have the same issue. Idle quality should suffer a bit. Otherwise you can verify TDC using TDC piston stop and degree wheel to make sure your pulley is correct: https://www.summitracing.com/parts/pow-pow101330

Just saw a picture of your plugs, sorry for not seeing earlier (I was on my phone).

You are going to want to check oil pressure to your head. This sounds a bit like ticking lifters and you can't always hear them. Low cylinder head oil pressure (due to milled head) can cause these issues. I know you said your car is mostly stock, but just to clarify, still a 1G 6 bolt head correct?

This is worth having, just to verify your oil pressure at oil filter housing and cylinder head is good, its only $30. https://www.summitracing.com/parts/sum-g2934

The first repair I ever did to this car was replace the original factory OEM balancer, which fortunately separated while still in the driveway. I replaced it with another OEM and 20+ years later it hadn't separated but showed some age so I replaced it proactively.

My head was milled a fair amount but I also performed the oil port mod, about halfway down this post shows a pic: https://www.dsmtuners.com/threads/1991-tsi-awd.540276/post-153867776
Yes, 1g 6-bolt head.

Oil pressure gauge is also on my wishlist. I'd really like something digital so I can run into the cabin.

 

[danl]

For now, get that summit mechanical one. You can just tape it to the windshield wiper and look out the window at it while you are driving. It is great to have around for checking oil/fuel/trans pressure on anything you are working on.

I've been thinking of doing a writeup for installing an oil pressure sensor to the EGR temp connector on a 1G. You can then use DSMLink to log it.

 

[dwb]

I am looking at your log now. It looks like most of your knock is before the coolant temp reaches 150 deg. After the engine reaches 180F the knock is significantly less. This really does mimic the HLA's ticking when the engine is cold (even if you can not hear it). Step 1 would be to check oil pressure to the cylinder head at the stock turbo feed location (1/8" BSPT plug). If it is low, there are a couple things you could do. If oil pressure is fine, step two would be to install known good HLA's.

What weight oil are you running?

I agree with that observation about temperature.
I have 3g lash adjusters and only notice them ticking on startup. Still have my original 1g lifters on the shelf, but those ticked all the time. I've also done a BSE without porting the oil pressure regulator. Therefore pressure *should* be adequate or on the high side but I'm only speculating. I still feed my turbo from the stock head port, so I'd probably need to T off that or find another pressure source. The opposite end of that gallery is blocked by the cam gears.

I have a 4:1 mix of conventional 10w30 and synthetic 5w30. Typically I run only conventional 10w30.

For now, get that summit mechanical one. You can just tape it to the windshield wiper and look out the window at it while you are driving. It is great to have around for checking oil/fuel/trans pressure on anything you are working on.

I've been thinking of doing a writeup for installing an oil pressure sensor to the EGR temp connector on a 1G. You can then use DSMLink to log it.

Good call, for 30 bucks it's a great option to have in my toolbox.
I had no idea the EGR inputs could be leveraged for oil pressure. hmmm...that would be handy!

 

[danl]

You can remove the 1/8 Bspt plug to measure oil pressure at the head. If the head was milled and have you done the oil pressure mod to the head, you should be fine.

The hla’s do wear out over time. They can also become scored if they ingest dirt/metal. They then require replacement as they need more oil pressure and flow to stay pumped up.

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

 

[dwb]

You can remove the 1/8 Bspt plug to measure oil pressure at the head. If the head was milled and have you done the oil pressure mod to the head, you should be fine.

The hla’s do wear out over time. They can also become scored if they ingest dirt/metal. They then require replacement as they need more oil pressure and flow to stay pumped up.

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

oh yeah, THAT port

Come to think about it, these 3g HLAs (prevously used and unknown condition) were installed the same time I did the BSE. Could they cause a vibration if one or more are "bad?" They all seemed ok when I bench bled and cleaned them.

 

[danl]

The vibration is possibly coming from your BSE and/or a bad engine mount.

After you verify oil pressure to the head, I'd be looking at the HLA's next. I'm not sure of a way to test them without replacing them.

 

[dwb]

The vibration is possibly coming from your BSE and/or a bad engine mount.

After you verify oil pressure to the head, I'd be looking at the HLA's next. I'm not sure of a way to test them without replacing them.

Mounts/rollstops are OEM and have good rubber. I bought that Summit oil pressure gauge and I'll take a reading once it to arrives.

 

[dwb]

Here's a couple videos of oil pressure at the head using the OEM turbo oil feed gallery. Video quality seems to be lacking, hopefully enough to read the gauges. When I get a chance I'll do the same tests from the OFH.

This is cold start:

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

This is hot start:

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

Oil gauge for reference:

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

 

[We're on Boost]

I'm waiting myself to see what people will say about the 2nd video. But 0 psi at hot idle 800 rpm sure seems wrong to me!
Is this one of those gauges that doesn't do anything until you have about 10 psi on it?

 

[dwb]

I'm waiting myself to see what people will say about the 2nd video. But 0 psi at hot idle 800 rpm sure seems wrong to me!
Is this one of those gauges that doesn't do anything until you have about 10 psi on it?

I agree, seems low to me too! I think it's just *barely* above the peg, but still below the 5psi line at hot idle. This is a brand new gauge so I'm not sure what threshold it starts reading. I'll hook up my air compressor to the gauge and see how it compares to my regulator reading and at what pressure it starts reading.

 

[danl]

A couple things.

1. One of your previous logs showed more knock when the engine was warming up and less as the engine was up to temp.

2. The pressure at the turbo feed location is the pressure going TO the head. It still has to go to the "regulator" which bleeds pressure further. That means that the pressure post "regulator" which feeds the lash adjusters will be less than measured at the turbo feed location. You only need 15psi to the HLA's in the head. https://www.kigglyracing.com/product-page/hla-pressure-regulator-4g63

3. High pressure does not mean high flow. High pressure means you have a restriction and pressure is able to build up. That you have high pressure when the engine is cold does not mean that you have good flow to the HLA's. When the engine is warm, the pressure is lower, but you are getting more flow.

With the engine cold, are you getting good oil flow to the cams? If you remove the oil cap with the engine running you should see the cam lobes coated with a stream of oil and oil should splash out onto the valve cover, making a bit of a mess. If you are not getting good flow to the cams, you are not getting good flow to the HLA's.

With the engine warm, are you getting good oil flow to the cams? Same as above.

It has been a decade since I measure oil pressure at the cylinder head turbo feed location and I forget what it should be. It will be a couple weeks until I can do this on my car. In the meantime hopefully someone can chime in on what normal pressures at the head are.

 

[dwb]

Is this one of those gauges that doesn't do anything until you have about 10 psi on it?

Testing the oil gauge with compressed air shows it won't register anything until somewhere between 5 and 10 PSI.

1. One of your previous logs showed more knock when the engine was warming up and less as the engine was up to temp.

Yes, but still not completely temperature dependent either. I still get PK when up to full temp but more noticeable when cold.

With the engine cold, are you getting good oil flow to the cams? If you remove the oil cap with the engine running you should see the cam lobes coated with a stream of oil and oil should splash out onto the valve cover, making a bit of a mess. If you are not getting good flow to the cams, you are not getting good flow to the HLA's.

With the engine warm, are you getting good oil flow to the cams? Same as above.

I'll check this again to verify hot/cold conditions, but I definitely saw plenty of oil squirting around last fall after I rebuilt the engine. I don't normally remove the cap while it's running, because, like you said, it can make a mess.

It has been a decade since I measure oil pressure at the cylinder head turbo feed location and I forget what it should be. It will be a couple weeks until I can do this on my car. In the meantime hopefully someone can chime in on what normal pressures at the head are.

I'd be interested to know, but some searching says as low as 5psi at idle and up to 30psi while hot, which is about where I'm at. I couldn't find much for cold pressure.

I'm tempted to pickup a set of the GSC "no-tick" HLAs, since my current 3g HLAs are unknown mileage. I've put 75k miles on them and bought them used. I've also started to notice a random tick beyond startup from the top end. I tried to catch some of that on the end of my cold start video but my mic didn't seem to pick it up.

 

[dwb]

Plenty of oil flinging around when hot. That made quite the mess! If I have time in the morning I'll check with cold start.

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

 

[danl]

If you can hear random tick after the engine has sufficient oil pressure for 10-20 seconds, then it is likely the HLA's need replaced. They should be silent when the engine is up and running. I have had dsm's tick because they didn't have enough oil pressure to the cylinder head, so it is always good to check that first.

The Kiggly HLA pressure regulator limits pressure to 15psi. So 5-30 psi sounds resonable. https://www.kigglyracing.com/product-page/hla-pressure-regulator-4g63

 

[dwb]

Got an opportunity to check the inside the oil cap on a cold start. Kind of hard to tell how much oil is there but it's definitely not flinging around like when hot. At this point I feel like new HLAs are probably a worthy investment.

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

(sorry for vertical video)

 

[Stapl3]

I think there are two possibilities here. One is phantom knock. If you're confident that nothing is rattling around that shouldn't be, then the sensor could simply be bad. Get a new one, OEM if you can [good luck], and as said above torque it lightly. I use locktite and hand tighten it. Also check to be sure sure the sensor connection to the harness is clean.

Secondly though, your timing is super aggressive. And it's not your fault, it's just where you landed on the timing table with the low airflow. Ironically I think if you ran higher boost (high teens) and 93octane you might be better off than you are now as it would throw you a row or 2 down on the table.
With that said, I think I can tune it out where you sit. Load in this file attached.

 

[dwb]

Still the same knock issue on another ECU.
I installed my original non-eprom. Turned the boost down to about 10psi due to 1g timing maps. Ran the same route as my previous logs and logged it with old-school TMO datalogger. 43 counts at 2500 RPM and less than 30% throttle at around 190 seconds in the attached CSV (if you want to look at the log).

I think my next step will be to loosen the knock sensor and re-torque it. Then some new HLAs.

 

[Blinkyblinx9]

Have you thought about listening to the knock sensor to hear what kind of sound its picking up? You can not only listen using a small headphone amplifier and some wired headphones but with even a cheap USB oscilloscope in place of the amp you can visually inspect the signal. I don't know what that data would be worth to you but if you or someone you know has the know how perhapse you can narrow in on what could be causing it. There are also magnetic microphones you can attach to the engine to record sound and particularly the volume at different areas of the engine. If you truly cant find the cause then perhapse it might be useful into.

 

[We're on Boost]

Have you thought about listening to the knock sensor to hear what kind of sound its picking up? You can not only listen using a small headphone amplifier and some wired headphones but with even a cheap USB oscilloscope in place of the amp you can visually inspect the signal. I don't know what that data would be worth to you but if you or someone you know has the know how perhapse you can narrow in on what could be causing it. There are also magnetic microphones you can attach to the engine to record sound and particularly the volume at different areas of the engine. If you truly cant find the cause then perhapse it might be useful into.

This is cool. I've been thinking that he should have a way to "see" the voltage signal that the knock sensor is putting out, to see if it is a bunch of spikes (which would either be knock or some mechanical noise like valve train) or if it is just a steady high voltage (electrical problem). Seems like the "cheap USB oscilloscope" would be a good way to do that.

With the cheap USB oscilloscope, do you mean that he would tap into the wire coming from the knock sensor with a wire going to the input of the oscilloscope, then you USB the output of oscilloscope to a laptop to see it? Is that the idea? Would he be able to record it, like a logger? I see a few USB oscilloscopes online for $50 or so, and of course the prices go up from there. Is there a particular one you have in mind?

 

[Blinkyblinx9]

Yes, you would have to add a properly insulated port to the knock sensor harness as to make listening and viewing as easy as plugging in your oscilloscope or amp.

The oscilloscope unit processes the inputs then sends the data to a laptop to be viewed. I got mine from Amazon for like 80$ because Amazon doesn't hastle you for returns so if it didn't work I take it to Kohl's and order another one.

The software that mine came with sadly doesn't allow you to log and the resolution isn't the greatest over longer sample sizes but if you poke around Google for reviews and such im sure you'll find a few under a hundred that log to an extent.

It has made diagnosing injector and sensor issues easier as well as allowing me to use inductance from spark plug wires to diagnose misfires vs fuel cut without being able to log before I got link and could only poll PIDs over our crappy OBD2 protocol. Not a bad tool when your all out of ideas and throwing money at a problem could prove expensive if you don't nail it the first or second time.

I however stupidly never used the tool in this manner which would have probably saved me a lot of money over rebuilding my head so as far as experience using it this way I can not help you. I do however have a couple ideas as to how it could be done.

 

[We're on Boost]

I however stupidly never used the tool in this manner which would have probably saved me a lot of money over rebuilding my head so as far as experience using it this way I can not help you. I do however have a couple ideas as to how it could be done.

Do you mean you never used it to diagnose knock?

Have you seen my post #29 here where I show what knock sensor voltage looks like when there is real knock?

 

[Blinkyblinx9]

Well I've seen what knock voltage looks like. I was saying that I've never personally used this specific method to view knock sensor voltage.

 

[dwb]

All good stuff guys! I wouldn't have to reinvent the wheel either because this has been done before. Wish I could find the article he was talking about where they did this to a DSM.
https://www.k20a.org/threads/how-to...ensor-to-listen-for-real-vs-fake-knock.27258/

I also noticed it was about a quart of oil low after my log on Sat, so I topped it off. I haven't logged it since but I swear it feels better (but not fixed) than it did last week. Last Friday it was a dog with more PK than usual. Ambient temps were around 95 that day so I correlated some of the PK with the high ambient temps, but now I'm thinking lower oil levels could be a bigger factor. Which could point blame at the HLAs or something with the oiling system.

On that note, this car has ALWAYS ran (felt) better after a fresh oil change. It's also a little happier if I run the oil level a little over 5 quarts.