Weird knock detection L98

[space387]

I have been working on the tune for my 1990 Corvette and have found that it shows a knock detection at 1600-1650 rpm. I can not seem to identify what is causing it and have tried backing off the spark table to counter it with no avail the knock retard remains at the same value. This is only at 70mph, 2mph lower or higher the spark magically disappears. I'm not sure how to combat this. Any help would be appreciated.

[NomakeWan]

If it’s at a specific speed at a specific RPM it sounds to me like it’s resonance causing false knock. Combatting that would be difficult as you would need to find the source and dampen it (may be impossible), would need to dampen the knock sensor (Teflon tape on threads; this however reduces overall sensitivity of the sensor across the board), or would need to create a new ESC profile (dunno about a ‘90, but on $EE this is currently not possible due to lack of reverse-engineering of the ESC module).

Best of luck.

[space387]

Thanks. I have seen this before where I had set the distributor wrong and the car was running with about 10* less timing than is should have but I could feel when it would pull down to about 12* actual timing leaving me with no power to pass 70mph in 6th gear. I have been very active with tuning so maybe I will try leaving the knock detection active just remove the retard ability while I tune allowing me to interpret the events.

[84Elky]

Space -- Please upload a .csv file of your log. Without knowing your ECM mask and having a matching ADX, your xdl log cannot be properly analyzed. You can see only 1 view of your attachments -- mine.

[space387]

ADX and XDF have been updated

[84Elky]

Knocks could be from transient vibration. But look at the 'Knock Count & Retard' Tab of the attached Analyzer. Analyzer tells you everything. Some things not shown because you're not using S_AUJP. See also attached .xlsx exported log.

Very high SA-TDC reported. Unless you have aluminum heads, this is way too high and even then questionable. You'll also note knocks occurring when the engine is not under any significant load other than WOT (ie - kPa generally < 70). That's not knock territory. I'd reduce entire spark table 5-6* at 1000 RPM all kPas and see what happens. Gotta start somewhere.

Other observations:
- o2 sensor randomly reporting not active causing open loop. See highlighted samples at 2880, 5166, 5457). This is due to bad connection or failing sensor. It's not due to not being heated or high enough temp because it is active during most of idle at end of log.
- Your idle TPS% needs to be reduced. Code assumes idling only when when TPS < 0.78% and MPH < 0.9. This affects IAC positioning. Voltage = 0.60% at TPS%=0 is best.
- Also note at sample 24723, TPS% randomly begins to drop to 0% while voltage does not change. These type of oddities are generally due to bad grounds or an inadequate connection somewhere.
- Additionally, you'll see that idle spark at sample 24720 drops significantly below the 20.04 degrees in your spark table.

Bottom line, a bad ground or connection somewhere could also be causing the knock retard, which may not really be occurring but is being reported as such.

All other info looks OK. But cannot rely on BLM data in Analyzer because of all the random OL samples.

HTH, Elky

PS ---
Really need to upgrade to S_AUJP. Conversion is painless, will take 10-15 minutes and you'll be using your exact calibration as now, plus will have access to every calibration item in BIN via XDF and will be reporting everything possible providing better analysis.
Here: https://www.thirdgen.org/forums/diy-prom/767540-s_aujp-version-6-released.html

[space387]

84Elky I am running S_AUJP V5, I did not see the update but will make the change tonight. I found the V5 did not have the ability to adjust the CAGS enable temperature so I manually entered it under emissions, it is located at 0x300. The grounding issue is a very real possibility and will be looked into this coming week where I have a few days off and will be replacing the clutch and flywheel for a single mass unit. This is for a 1990 Corvette with aluminum 113 heads and most of the spark table you see is stock.

One final note is that the provided tune has Highway mode enabled for spark and fuel running the engine significantly lean. From my best investigation over the past 2 weeks I do not believe this is actually spark knock but false knocks by a harmonic ( maybe the flywheel) or poor grounding as you suggested. Thank you for looking into it and I always appreciate the help.

[84Elky]

IF running S_AUJP v5, you are not using the v5 ADX. You are using the old AUJP v2 ADX which greatly limits reporting (ie - no In AE, In PE, DFCO, etc.)

[84Elky]

84Elky I am running S_AUJP V5, I did not see the update but will make the change tonight. I found the V5 did not have the ability to adjust the CAGS enable temperature so I manually entered it under emissions, it is located at 0x300. The grounding issue is a very real possibility and will be looked into this coming week where I have a few days off and will be replacing the clutch and flywheel for a single mass unit. This is for a 1990 Corvette with aluminum 113 heads and most of the spark table you see is stock.

One final note is that the provided tune has Highway mode enabled for spark and fuel running the engine significantly lean. From my best investigation over the past 2 weeks I do not believe this is actually spark knock but false knocks by a harmonic ( maybe the flywheel) or poor grounding as you suggested. Thank you for looking into it and I always appreciate the help.
Beginning with the initial version of S_AUJP (v4), 0x300-0x307 are not supported. They are in commented-out code deemed to not be necessary as it contained a lot of GM testing, so entering values for them in the cal is meaningless. Manual car will run the same with or without these items.

[space387]

IF running S_AUJP v5, you are not using the v5 ADX. You are using the old AUJP v2 ADX which greatly limits reporting (ie - no In AE, In PE, DFCO, etc.) I have attempted to use the V5 ADX when recording and result in garbage data. I will try with V6 in hopes it was just an error.


Beginning with the initial version of S_AUJP (v4), 0x300-0x307 are not supported. They are in commented-out code deemed to not be necessary as it contained a lot of GM testing, so entering values for them in the cal is meaningless. Manual car will run the same with or without these items.

Not to sound arrogant but today I was provided a tune to try for a tuner and it had the CAGS enabled. I found this was not present on the V5 options so I switched back to $8D mask and found the address. I entered a new scalar under emissions and found the stock temp of 50C was showing for a value. Changed it to 151C and it has not triggered the 1-4 shift since. If it has been coded out of the new bin then you are correct in not needing it. to this point I have just used the stock tune and edited it under the V5 mask.

[space387]

Elky I have tried following the provided instructions to carry over the calibrations from my current tune to the new bin provided but no mater what I do I get garbage out and the car will not idle. I then tried using the provided bin as sent and the car went into failure mode. For a 1990 Corvette is there a specific bin to start from and where have I gone wrong getting it to communicate?

[84Elky]

Elky I have tried following the provided instructions to carry over the calibrations from my current tune to the new bin provided but no mater what I do I get garbage out and the car will not idle. I then tried using the provided bin as sent and the car went into failure mode. For a 1990 Corvette is there a specific bin to start from and where have I gone wrong getting it to communicate?
You would start with the $8d bin you posted above in post #1 assuming that's a BIN with your latest "tune" (also called the BIN's Calibration).
Is that what you did and did you follow the instructions in the read me doc?
If so, which transfer Option did you use?
If you did something else, please describe exactly the procedure you used.
Please post the failing S_AUJP BIN and label it as such.
Tks.

[84Elky]

I have attempted to use the V5 ADX when recording and result in garbage data. I will try with V6 in hopes it was just an error.



Not to sound arrogant but today I was provided a tune to try for a tuner and it had the CAGS enabled. I found this was not present on the V5 options so I switched back to $8D mask and found the address. I entered a new scalar under emissions and found the stock temp of 50C was showing for a value. Changed it to 151C and it has not triggered the 1-4 shift since. If it has been coded out of the new bin then you are correct in not needing it. to this point I have just used the stock tune and edited it under the V5 mask.
You can put a value into any S_AUJP BIN at 0x300-0x307 and they will not be used. This is why they are not in the S_AUJP XDF files. The code is commented out. However, 0x300-0x307 are used in the Factory version of AUJP, which is pre-S_AUJP.

Factory AUJP Calibration Default Values
;------------------------------------------------------------------
; Computer Aided Ratio Selection.
;------------------------------------------------------------------
L8300: .byte 0xFF ; IF COOL GT 151c CK FOR CARS ACTIVE
L8301: .byte 0x00 ; RESET SPEED
L8302: .byte 0x00 ; ENABLE SPEED
L8303: .byte 0x00 ; DISABLE SPEED
L8304: .byte 0x00 ; RPM DISABLE
L8305: .byte 0x00 ; %TPS FOR DISABLE
L8306: .byte 0x00 ; KPa LOW BARO DISABLE
L8307: .byte 0x00 ; KPa LOW BARO FOR RE ENABLE

[space387]

You would start with the $8d bin you posted above in post #1 assuming that's a BIN with your latest "tune" (also called the BIN's Calibration).
Is that what you did and did you follow the instructions in the read me doc?
If so, which transfer Option did you use?
If you did something else, please describe exactly the procedure you used.
Please post the failing S_AUJP BIN and label it as such.
Tks.

After downloading the new zip from thirdgen.org I opened TunerPro and selected the new bin. I then loaded the new xdf and proceeded to use the difference tool to copy my latest .bin settings to the S_AUJP V6 -28. I checked the limit search to defined only and from there transferred over all the settings from 0018 through ve multiplier at 0821.

As I type I tried doint the same task with the $8D xdf loaded and noticed a significant difference in results. Is this how I should have done it under the $8D mask first?

For the ADX I simply selected the S_AUJP v6 as my definition file. Is there another step I have missed here too?

I also saw an S_AUJPN Bin file is the difference report showed nothing different, what would be the intended use of the two if different?

[84Elky]

After downloading the new zip from thirdgen.org I opened TunerPro and selected the new bin. I then loaded the new xdf and proceeded to use the difference tool to copy my latest .bin settings to the S_AUJP V6 -28. I checked the limit search to defined only and from there transferred over all the settings from 0018 through ve multiplier at 0821.

As I type I tried doint the same task with the $8D xdf loaded and noticed a significant difference in results. Is this how I should have done it under the $8D mask first?

For the ADX I simply selected the S_AUJP v6 as my definition file. Is there another step I have missed here too?

I also saw an S_AUJPN Bin file is the difference report showed nothing different, what would be the intended use of the two if different?
Tks for update. Not sure I fully understand, but please respond to these 2 items
Did you start with the BIN posted in #1 here? If not, please post the bin you used
Please post the failed S_AUJP BIN and label it as such.
Please also send me your current XDF file, not S_AUJP


Don't worry about the NVRAM 'N' version for now. You can't use it. Its calibration is the same as the regular version. Both have default values that should not be used because they may not provide the necessary "tune".

[space387]

please ignore this post and files

[space387]

Here is the Base Bin and XDF. This is using the $8D mask with an AXCN original file. I calibrated the S_AUJP while the base XDF was loaded with the file below. It runs but a little rougher than the original bin after being modified. When trying to datalog the only ADX that reads is for the $8D adx I have below. When using the S_AUJP ADX I am plagued with data errors from the stream.

[84Elky]

Space ---
Somehow we are not communicating. I want help, but you're all over the map. You started with an ARFP bin in Post #1 of your initial thread. I asked if that was the bin with the good calibration and you did not reply. Now you provide an AXCN bin here. It also appears you did not follow the calibration transfer instructions or something went major wrong because your S_AUJP calibrated bin here does not in any way agree with your "base bin" calibration. See calibration snippet below.

For me to help i need you to post exactly the following:
- The XDF for $8d that you have been using to edit a bin running in your car.
- The $8d bin that is properly running in your car.

Your base bin is on the left -- Your S_AUJP calibrated bin on the right. See any problems? I'm surprised it runs at all.

15018

[space387]

The Base Bin and Base Bin mask above are the simple answers of what I am using to run my car. The reason I started with an ARFP 3 weeks ago and now show an AXCN is because I began talking to a professional tuner and in an effort to help me learn how to make better adjustments to the bin he sent me that file. The one labeled Base bin is actually the 4th iteration of him sending a file and I dataloged. Small changes where made to get a solid base tune. I was not told it was an AXCN until yesterday. He swears they have no difference but the undefined code has a number of variances.

For the Hex difference I do not have an answer to why it is like that. when I start I have the Base min mask loaded for my XDF and load in the uncalibrated V6 bin. Next is open the difference tool and compare the uncalibrated bin to the base bin listed above. I make sure to check the defined values only box and from there select all listed changes and hit copy from right to left. Apparently I am missing something.

[84Elky]

For the Hex difference I do not have an answer to why it is like that. when I start I have the Base min mask loaded for my XDF and load in the uncalibrated V6 bin. Next is open the difference tool and compare the uncalibrated bin to the base bin listed above. I make sure to check the defined values only box and from there select all listed changes and hit copy from right to left. Apparently I am missing something.
I have no clue what happened, but attached is your calibration from base bin above now in S_AUJP v6 Base bin left -- S_AUJP right. No differences.
15023

EDIT ADDITION -----
Space --
I owe you an apology. I do now know what happened. The instructions for transfer of a Calibration using TPro are incorrect. It is not possible to accurately use TPro and an XDF to do the transfer without a lot of tedious item by item detail. A blanket copy will not work. It has to do with items that may or not be in an XDF (S_AUJP or prior). So the only way a transfer can be done quickly, easily and error-free is to use a hex editor which allows copy of the entire calibration in one step. That's why HxD is recommended for the transfer.

Instructions will be modified and updated on thirdgen.org here:
https://www.thirdgen.org/forums/diy-prom/767540-s_aujp-version-6-released.html

Again, apologies for the error. Please let me know if the BIN I sent works.
Elky

[space387]

It's very understandable the mix up. I'm doing a clutch today but once it's back together I will absolutly give it a try.

[space387]

I have always been curious how do you add commands and features to a bin? Eventually I would like to understand how the code breaks down so anything you are willing to share would be greatly appreciated.

[NomakeWan]

Depends on what you mean. If you mean adding things that you can edit in TunerPro, it's a matter of defining address ranges in the XDF with descriptions and the type of address it is (toggle, table, map, etc). You can open an XDF in your text editor of choice to inspect its structure for yourself. TunerPro itself also has an XDF editor built in to make doing so a little bit more user-friendly.

If you mean actually adding features to the BIN that were not present from the factory, such as what kur4o and steveo do to allow things like faster logging, open-loop operation, wideband data processing, etc...that requires actual understanding of the program that runs on the hardware in question, as well as knowing where free space (or useless/unnecessary code) is to place the code, and creating proper links so that the code is actually run. Far more complicated than just defining things in an XDF.

[space387]

What s I am looking to gain is how does the ECU know to do specific commands. Say it should be cycling the EGR valve, Yes there is a flag to check off and a few scalars to dictate when but where does the code say for Pin A3 check this list of scalars and flags to see if it should be enables. I understand this is very difficult and by looking over this past thread it probably seems above my ability but I do want to try and learn how these ECU's work on a coding level.

A good use might be changing the use of a pin from something like a 1-4 shift/TCC lockup to another feature like an automated cutout.

84Elky would it be possible to look up what address in the Y-Body bins controls pin A12? It is connected to the built in Up shift indicator. The owners manual says for the L98 it only tells you to shift to 6th at highway speed but for the LT5 its set to work as a shift indicator for all years when at high RPM's. If it can be found it would allow you to move the shift light to another pin if so desired for the 1227277 ecu.