Low Timing and Still Knock

[spadepro22]

No leaks, 170 compression with kelford 272s and it idles very good. I can't figure out why I'm getting knock. All basic things has been checked and motor was rebuilt 1,400 miles ago. My profile is up to date. It runs perfect until I try to tune doing 3rd gear pulls. Here are todays pulls.

 

[BrokenTsi]

Something is definintely wrong. You have your knock sensor disabled until 3400rpm which is right where you start to read knock on the datalog. So its probably detonating or reading false knock right from the get go. I dont see an issue with the amount of timing being ran during spool up, however look at the difference between your timingMAX and timingMIN tables. When the ECU detects knocks, it will refer to the timingMIN tables.

Now notice how your timingMIN table is targeting HIGHER timing values under the areas that you're knocking. So basically, when you knock, you're telling your ecu to add more timing which is why you continue to knock thru the entire pull. The difference between these two maps needs to be a smooth transition so the ECU can interpolate the values. I never copy my max over my min tables. I'll take the timingmin table and subtract 3-4* from the max table.



Most of the time, when I encounter this type of issue with a customer, I find that their base timing was way too far advanced. Re-check your base timing with a timing light and confirm that its at 5* BTDC with the timing connector grounded (found in the misc tab).

 

[spadepro22]

Something is definintely wrong. You have your knock sensor disabled until 3400rpm which is right where you start to read knock on the datalog. So its probably detonating or reading false knock right from the get go. I dont see an issue with the amount of timing being ran during spool up, however look at the difference between your timingMAX and timingMIN tables. When the ECU detects knocks, it will refer to the timingMIN tables.

Now notice how your timingMIN table is targeting HIGHER timing values under the areas that you're knocking. So basically, when you knock, you're telling your ecu to add more timing which is why you continue to knock thru the entire pull. The difference between these two maps needs to be a smooth transition so the ECU can interpolate the values. I never copy my max over my min tables. I'll take the timingmin table and subtract 3-4* from the max table.



Most of the time, when I encounter this type of issue with a customer, I find that their base timing was way too far advanced. Re-check your base timing with a timing light and confirm that its at 5* BTDC with the timing connector grounded (found in the misc tab).

I've checked base timing via link and it looks to be at 5*. I'm checking by looking down at the crank pulley(looking down in into the bay). I'll check again tho.

What is too lean? Do I need to lean it out?

 

[BrokenTsi]

Ecmlink assumes you are at 5* mechanical base timing.

During the areas of knock, according to your wideband and assuming that the log reflects what is on the gauge, doesn't seem too lean to me. You are barely at any type of real load at 3400rpm, however I can tell you that are most likely showing knock retard before 3400rpm.

Take a look at your knock sensor tab in ecmlink. Its disabled below 3400rpm but in the datalog, you knock AT 3400rpm. Something is mechanically wrong with your car, and I'm willing to bet that the CAS is cranked too far forward. I need you to confirm this with a timing light that it is at least at 5* BTDC before we can go any further with tuning.

 

[killercolt]

He stopped by my house on Saturday. We checked the base timing is correct with thr light. The car is having some sort of phantom knock.

Here are what I did to diagnose the knock issue:

1. Pulled 5 degrees across whole the timing slider. Still knock as soon as he jab the gas and build in 3rd gear. Still knock.

2. Pulled 8 degrees across the whole timing slider. Same knock issue as soon as soon as he jab the gas hard. Pushing the gas pedal down slowly to build boost did not produce knock.

3. Zero out timing slider and leaned the car out 10% from 3000-5000rpm to rule out rich knock. Same issue as above. Jabbing the gas pedal fast initiated knock.

4. Zero all the timing and fuel set points out. Car is at idle. Rolling into the gas at idle produce no knock. Jabbing the gas pedal produced knock.

5. Condition same as #4. This time the knock sensor is "detached" from the block. With the Knock sensor still connected to the connector, it was touched to the intake manifold to produce ground. Jab gas pedal. Again knock is produced.

6. Condition same as #3 and #5. The knock sensor was connected to the connector and held by my brother in his hand. Jab the gas and a lot of knock was produced.

7. Condition same as #3. Knock was complete detached from connector. Jab the gas and it was knocking the same. How is it pick up knock with a knock sensor.

8. Condition same as #3. Snipe the knock sensor wire near the ecu to rule out wiring issue. Same crap. I have to give ip on this one for now. This thing producing knock with no knock sensor signal of any sort. It only does it if you jab the gas.

9. Took the ecu apart to inspect it. Only thing that concerned me is one of the capacitor is not completely soldered. This is the capacitor closest to the EPROM. One of the leg on the capacitor not solder completely though like I am use to seeing. This leg is soldered to the trace on the bottom but not the trace on the top. He said Ecmlink switched all the capacitors. I am not sure if this is done right or not.

This is problem is new to me. I never seen it before. Hope you guys can help him out. I guessing ecu issue.

 

[BrokenTsi]

Nice work on troubleshooting this issue.

I suppose the next step is to replace the ECU. There is nothing on our engines that will tell the ecu to retard timing, other then the ECU itself. Good luck, and keep us posted.

 

[spadepro22]

Any more ideas on this? I added a wire to the shield that is around the knock sensor (white wire) and took it to ground to make sure it is grounded and no change. I'm not sure if its the ecu because everything else functions correctly and it was checked out by ECMtuning in November when I upgraded to V3. Even gives the check engine light with the knock sensor unhooked. I'd hate to replace it and still have the same problem.

 

[Calan]

A few thoughts:

1. I'm not sure what the impedance is on the knock sensor input, but if the ECU is expecting a sensor to be connected, it could be registering knock due to random electrical noise on the floating input while the sensor is disconnected. I wouldn't automatically assume that random knock with no sensor connected points to a bad ECU. Tom or Dave at ECMTuning would need to verify that for sure though.

2. Jabbing the pedal (and seeing knock) vs. slowly rolling into it (no knock) doesn't necessarily mean it's PK. It's possible to have momentary fueling issues caused by rapid throttle changes that are difficult to notice which can cause knock spikes, but in this case I don't see anything that points to that....unless your wideband signal is slow and heavily averaged.

It's also possible that you never saw knock while rolling into it gently because his knock/throttle threshold is set so high (if you never went past 65% on the smooth acceleration).

3. We really need to see a MAP signal, in order to determine where your timing and AFR targets are in relation to the turbo's spool characteristics. Do you have a MAP sensor you can log?

4. Drop the knock threshold down to 2000rpm and 20% throttle or so for troubleshooting.

5. When the knock first appears under WOT, let off the gas completely and then immediately go back to the floor with the pedal. See if the knock disappears and stays gone, or if it comes back as soon as you are back on the gas. You have to be careful when doing this! Set the CEL to come on at 2-3* of knock, and let off immediately if it comes back when you get back on the gas.

 

[spadepro22]

I contacted Dave today and he said if the ECU declares the knock sensor to be bad it applies fixed amount of timing retard (roughly 3.2 degrees of knock retard) when the engine speed and throttle position criteria configured on the ALS/Knck tab are met. I'd say this is why we were getting knock with the sensor unhooked and when we cut the wire from the ECU. Cutting the wire stills makes the ECU think theres a bad sensor I'm assuming.

Calan I will try what you listed tomorrow and I don't think I have a map sensor I can log. How can I tell?

 

[Calan]

I don't think I have a map sensor I can log. How can I tell?

Look in your spare parts box or on the work bench. If there is a MAP sensor laying there, find a spare input on the ECU and wire it up.

 

[CrackedDSM]

I contacted Dave today and he said if the ECU declares the knock sensor to be bad it applies fixed amount of timing retard (roughly 3.2 degrees of knock retard) when the engine speed and throttle position criteria configured on the ALS/Knck tab are met. I'd say this is why we were getting knock with the sensor unhooked and when we cut the wire from the ECU. Cutting the wire stills makes the ECU think theres a bad sensor I'm assuming.

Calan I will try what you listed tomorrow and I don't think I have a map sensor I can log. How can I tell?

I was about to say this, when the knock sensor is unplugged or if a knock sensor is bad, the ECU defaults to so many counts of knock as a safety precaution. How's your knock sensor look?

 

[spadepro22]

I was about to say this, when the knock sensor is unplugged or if a knock sensor is bad, the ECU defaults to so many counts of knock as a safety precaution. How's your knock sensor look?

Knock sensor is almost new(about 2,700 miles) and looks very good. I have two of them that Im working with and knocks keeps showing up. If I set the threshold to 1200 rpm I will get like 15 counts of knock just revving at idle. Any suggestions would be greatly apprecaited. I can try some of what is posted tomorrow and post back with logs. Also is there anything that I need to add to be logged?

 

[MitsGray97]

So Im not sure if this thread was figured out, but in case it wasn't and the op is still dealing with knock, I would check your wideband readings, cause from what I can see in the log, you may be running to lean.

Also I have my own issues I could use some help with, I am also dealing with knock but mainly while cruising, and I have tried lowering the timing, same with adding fuel. No matter how much I add or take, I still have knock in the 2.5k to 3.5k. knock usually reaches from .4* to 1.4* I know it could be phantom knock but I can see correlations. And I also get sputtering around that rpm range . Any ideas? Would really appreciate the help. Thanks.

 

[Dxc545]

Your wb and afr are pretty far apart. Could use some mafcomp

 

[MitsGray97]

Your wb and afr are pretty far apart. Could use some mafcomp

Not sure if that was directed to me or the op. Either way I have adjusted the mafcomp to get as close to 0%, if anything i should be on the rich side.. Checked my wb readings, and have been checking the 02 readings. But it seems to me to be all in good working conditions. Is there a way to rule out if it is phantom? I know I have lifter tick, but I don't want to rely on the assumption that it "may" be phantom knock. And then later on see damage....