HyperDisorder

iTrader: (1)

With donner's help (and jscbanks original code of course) I've made a DMA rom for 93660005, as I, like many others who went to 96260709 (and its variants) have had a lot of false knock issues. I thought that this would be a fairly definitive test to help me cure any problems.

So anyway, I've attached the rom, xml and a before and after 3rd gear pull. You can see a spike or two of what is surely real knock, but you can also see the big step changes in 9366's knock base that is reasonably higher than almost all of 9626's knock filtered signals that breach 9626's knock base. I am assuming that the knock filtered signal is the same in both rom's when making the visual comparison!

So, yes there's a lot of false knock to filter out (IMO anyway)!

I use the 96262709_live xml with this rom in ELM. Only use the graphing tab - all the other tabs (fuel, timing, BDEL, etc) won't give you any meaningful info.

On a side note - I have tried det cans - both DIY ones and industrial type (work stuff that I temporarily re-purposed!) with no luck, but didn't feel convinced. I'm no expert at listening for knock!

burgers22

Thats interesting. In the 9626 trace there is no change in knock base multiplier as in the 9366, or have you used a different tracing mythology?

HyperDisorder

iTrader: (1)

I used the default knock settings in both, with the same testing technique. I used ELM with the same XML both times, and made the charts in excel.

The 9366 rom only has 3 multiplier values, which differ greatly, while 9626 has 14 values that gradually increase and decrease, which should why there's a big step change in 9366's knock base and not in 9626's.

Now that I can see the reason for my knock was this fundamental difference in roms, I should test 9626 with big step changes in the multiplier table to see the effect. I better get rid of my few remaining counts of knock first, to be safe.

french

I have also come from 9366 to 9626 and have lots of knock. I also did discover that the knock multipliers do differ between the roms.

I have yet to research further but did you get any further with this? Such as finding safe knock control for 9626.

I dont feel like reinventing the wheel.

french

Bump

french

Any information would be helpful as my knock is impeding on my sanity atm.

Assault

Did anyone get anywhere with the knocking on this Rom? Or did they just move on to a different Rom?

HyperDisorder

iTrader: (1)

So sorry for the late reply. I never did perfect my knock tables to eliminate the false knocks, but I do know that the false knocks around 2500 RPM when increasing the load are due to the ACD relay activating. I can hear it tick and see the ECU pick it up as knock.

Only seems to be an issue at higher load at 2500 RPM, so I just ignore it.

aqeelrizni

Bump.

Did you finally move to a different rom or managed to find a sweet spot to adjust the knock tables on the 96260709 to get rid of this false knock?

Assault

I eventually just bumped up all of my knock multipliers until I eliminated false knock over 1 or 2 counts. I used knock ears to confirm the false knock wasn't real knock and that actual knock was still registering as 3 counts or higher.

aqeelrizni

Thank you. I think im gonna do the same. Verify with a pair of det cans and then slowly bump the multipliers up

aqeelrizni

Dropped u a PM