problems!! knock for NO reason - experiments

[cjridert1]

severe tim retard problems b/c I am getting lots of knock. So I change the sensor, no different. I JB welded over the hole in the sensor (1g), no different. I then hooked up the new knock sensor, wrapped it in a papertowell/washcloth combo and didnt put it into the block (go figure). Knock is substantially reduced, but when free-revving, still get knock spikes of 9 (always nine) - The knock sensor is hanging statically and wrapped in cloth - what do you guys think the problem is? (fyi - the engine is not knocking @ all). WB O2 is good and this is done just free-revving in neutral.

logged a 3rd gear pull (11psi boost) with the knock sens out of block (attached to harness and wrapped in cloth).

RPM TIMA knock wbo2 (lambda)
2100 21 9 .86
2700 21 9 .78
3300 16 9 .75
4600 17 9 .72
5100 18 9 .72
5800 18 9 .72
6200 18 9 .72

Knock drops to 0 when I let off gas. Furthermore, any time there is load on the system (like lightly accelerating from 1k to 2k) knock count holds @ 9. Seems to only be 0 when maintaining load/rpm or in decel. Keep in mind the knock sensor isnt in the block and my car isnt very loud. Does everyone see this in their setups? My lifters are a little noisy now and then (noisiest @ idle), but other than that there isnt any noise/vibration I can find.

 

[assault187]

So your still reading Knock with the knock sensor off the block and wrapped in a cloth????? and its the same outcome with both sensors? Maybe its a bad ECU, Leaky caps possibly screwing up the board. Check it out, It wouldnt hurt to pop the ECU out and crack it open to see if anything is visable

 

[cjridert1]

So your still reading Knock with the knock sensor off the block and wrapped in a cloth????? and its the same outcome with both sensors? Maybe its a bad ECU, Leaky caps possibly screwing up the board. Check it out, It wouldnt hurt to pop the ECU out and crack it open to see if anything is visable

yes, still knock with it our and in a cloth - but always 9, no fluctuation (maybe it needs grounded to block? - this is my 5 minute project today). Same outcome with new and old sensor (old one was replaced ~5 yr ago). ECU is only a year or so rebuilt, but I thought the same thing. Its still under warranty, but if I open it I void it. ehhh, ideas? Thanks much for the reply.

 

[cjridert1]

Grounded the sensor today and success - 0 knock thru rpms, but timing adv. seems too high - mid to high 20s all the way thru.

Car runs much better than before (go figure - 43 knock count), but timadv is near 30, this seems a bit too high right (although, only 10psi)?

I then sanded down where sensor goes into block (I see a little paint) then put the old sensor back in (with the JB weld) and did a run with only 8lbs.

RPM knock Timadv
2100 0 27
2500 0 27
2800 7 24
3200 20 19
3600 23 18
4000 22 18
4500 39 12
5100 37 14
5500 35 15
6000 33 18
6300 33 18

Is phantom knock this high for you guys? It seems I should @ least be able to hear it, no? That and I dont have any knock @ idle from lifters, etc... I logged knock of 24 in second gear with very little load (going into 3rd gear pull). O2's are good (~.74 @ redline). Car still throws knock spikes when revving @ idle.

 

[cjridert1]

I have 8.5:1 pistons - could the higher compression be the problem? I dont have a fmic, but it doesnt seem that slightly hotter air would cause this much knock - somebody take an educated guess?

 

[cjridert1]

Nobody has taken a stab at this? I was thinking about permanently mounting the knock sensor out of the block (since there is no audible pinging) and using my ear if there ever is knock, like the domestic guys do. My one buddy suggested it may be a wristpin/piston slap - what do you guys think? Car has lifter tap, but nothing else I can hear. Pull the knock sensor and run it as is? I cant diagnose the RIDICULOUS knock (via sensor) and am ready to not care. The engine was rebuilt ~60k ago and has done some autocross stuff, low # of passes and has been daily driven for years with pretty low boost (15). I know somebody wants to take a guess or make a suggestion

 

[spyderturbo007]

O2's are good (~.74 @ redline). Car still throws knock spikes when revving @ idle.

Keep in mind that 0.74v is considered lean. Although the O2 sensor voltage is just a guideline it can still be used as a generalization.

 

[123212321]

Nobody has taken a stab at this? I was thinking about permanently mounting the knock sensor out of the block (since there is no audible pinging) and using my ear if there ever is knock, like the domestic guys do. My one buddy suggested it may be a wristpin/piston slap - what do you guys think? Car has lifter tap, but nothing else I can hear. Pull the knock sensor and run it as is? I cant diagnose the RIDICULOUS knock (via sensor) and am ready to not care. The engine was rebuilt ~60k ago and has done some autocross stuff, low # of passes and has been daily driven for years with pretty low boost (15). I know somebody wants to take a guess or make a suggestion

Don't pull the knock sensor. That is never a good idea - it is your last line of defense. Just get DSMLink.

 

[2gGSX]

Just an observation, but I've noticed that the 1g front O2 signal is much less reliable than the 2g front O2 signal for some reason.

 

[cjridert1]

Keep in mind that 0.74v is considered lean. Although the O2 sensor voltage is just a guideline it can still be used as a generalization.

.74 is the lambda value - sorry, so this is ~11.5:1 or so. I didnt log the voltage off the O2. I should have said that.

Thanks for the replies guys

 

[spyderturbo007]

.74 is the lambda value

Thanks for the replies guys

Ah, that would cause the confusion.

 

[cjridert1]

was on the phone today with the SBR mechanic and he guessed crank bearing (seems this would make a noise though, or not necessarily?) - what do you guys think?

 

[2gGSX]

That's almost like telling someone with mild chest pain that they're about to have a MI, and then perform an emergency angiography.

Bad analogies aside, I think that that is a fairly weak diagnosis which would result in a PITA test to verify (e.g. pull the oil pan and crank girdle, and then inspect the bearings for failure). With knock only occurring during free rev and low RPM operation, it is more likely to just be phantom knock.

Do you have any data from a pull?

 

[cjridert1]

That's almost like telling someone with mild chest pain that they're about to have a MI, and then perform an emergency angiography.

With knock only occurring during free rev and low RPM operation, it is more likely to just be phantom knock.

hahaha! Tommorow I will pull the oil pan and check the bearings for failure (rebuilt engine has only 60k). 43 count for phantom knock seems too high for phantom knock (or no, Ive only worked on THIS dsm)? I get knock spikes of 43 when revving in neutral and during pulls the knock plateaus at whatever number (35-43); knock sensor is sending signals when free revving and when accelerating (even lightly), but knock occurs most @ high rpms/load. At idle there is no knock.

Also, timing advance does seem high comparing to most other peoples ~20*. I checked timing again today and made sure it was at 5*. What would be causing this?

Do you have any data from a pull?

heres a pull I took the other day @ 6lbs in third

RPM knock Timadv
2100 0 27
2500 0 27
2800 7 24
3200 20 19
3600 23 18
4000 22 18
4500 39 12
5100 37 14
5500 35 15
6000 33 18
6300 33 18

Thanks!

 

[2gGSX]

Wow, well with that sort of timing retardation I can't really tell you to just write it off as phantom knock. If this were real knock, then it could do some real damage.

Is this a tuning issue perhaps? I know that the wideband is logged in lambda, but what does it say on the gauge? Have you calibrated it lately, if that's an option?

 

[cjridert1]

Wow, well with that sort of timing retardation I can't really tell you to just write it off as phantom knock. If this were real knock, then it could do some real damage.

Is this a tuning issue perhaps? I know that the wideband is logged in lambda, but what does it say on the gauge? Have you calibrated it lately, if that's an option?

Im looking around to find if I can calibrate it, but it is most likely good (I used it with stock injectors until recently - so the wideband confirmed what the ECU/O2 readings should have been. I got the WB to tell me that this was or wasnt an a/f tuning problem.

I have had the car for probably 6mo now and it has ran like this since Ive had it best I can remember (which makes me think its not a bad crank bearing, as it would have died already).

When I unhooked the knock sensor and the timadv was ~30* - is this an uncommonly high #? Thanks for the help

 

[2gGSX]

In closed loop operation (low load cruise), that's completely normal.

 

[cjridert1]

In closed loop operation (low load cruise), that's completely normal.

When I took the sensor out and grounded it so I got 0 knock, I got 30* tim adv. - so this is under load with knock = 0. Sorry, should have already had this posted.

RPM TimA
2000 32
2500 26
3000 25
3700 24
4100 24
4900 25
5300 26
5800 28
6400 29

 

[2gGSX]

Assuming you're staying in closed loop while driving and revving around like that, then yeah, that should be fine.

 

[cjridert1]

Assuming you're staying in closed loop while driving and revving around like that, then yeah, that should be fine.

I would imagine it is open loop - the O2 flatlines (stock O2 sensor) and this is a WOT pull - it would be open loop right? (assuming Im not switching meanings again).

 

[Raph]

I would imagine it is open loop - the O2 flatlines (stock O2 sensor) and this is a WOT pull - it would be open loop right? (assuming Im not switching meanings again).

A 6psi pull (you posted that earlier) isnt WOT ,if you are lean spooling the turbo,then you look fine.


Edit: do you have an exhaust leak or cracked manifold?

 

[cjridert1]

Edit: do you have an exhaust leak or cracked manifold?

No leaks - Manifold and o2 housing has a few heat surface cracks in it if you take it apart, but is not cracked thru.

 

[2gGSX]

I would imagine it is open loop - the O2 flatlines (stock O2 sensor) and this is a WOT pull - it would be open loop right? (assuming Im not switching meanings again).

When I took the sensor out and grounded it so I got 0 knock, I got 30* tim adv. - so this is under load with knock = 0. Sorry, should have already had this posted.

RPM TimA
2000 32
2500 26
3000 25
3700 24
4100 24
4900 25
5300 26
5800 28
6400 29

I thought this pull was not at WOT? Those timing numbers look similar to what a 1g ECU shoots for at 1.0 g/rev load, which is why I assumed that was taken while you were driving around and easing the car up the RPM band.

If that WAS at WOT, then you're flowing very little air, as you would be around ~14 lbs/min at 6400 RPM then. I'm also not very familiar with the VPC you have listed in your mods list.

 

[cjridert1]

I thought this pull was not at WOT? Those timing numbers look similar to what a 1g ECU shoots for at 1.0 g/rev load, which is why I assumed that was taken while you were driving around and easing the car up the RPM band.

If that WAS at WOT, then you're flowing very little air, as you would be around ~14 lbs/min at 6400 RPM then. I'm also not very familiar with the VPC you have listed in your mods list.

Yea, this is a pull in third - I dont know if I can trust my airflow #'s since I have SAFC changing it to run the 550s can I? Or is there a like +16% conversion I can do maybe? The VPC just takes the signal from the MAP sensor and converts it into karmen (I think), so with VPC zeroed, it is stock ecu maps (I think) and the mass air flow is the same. Ill make a pull tomorrow (with 3* base timing) and post it.

 

[2gGSX]

I would think that the SAFC still functions the same way, which means the airflow numbers your ECU is seeing are deflated by the SAFC adjustment. If you do make a pull, I would suggest putting the knock sensor back on and enabling it. Perhaps I'm more cautious by nature, but if it were me, I still wouldn't make pulls with the knock sensor saying that there's knock.