How to tackle suspected real knock post-shift only?

[bradsm87]

I know a lot of cars do have a tendency to have a nasty single knock on one or two cylinders when grabbing the next gear suddenly. I don't want to take any timing out there because it doesn't knock there when in steady state.

I'm tempted to lower spark retard intensity 1 & 2 as we all know a -10 degree timing pull (ALL of the timing advance) is excessive for ANY knock event. I do like the feeling of having a huge safety margin though.

Are there any other/better tricks?

http://datazap.me/u/bradsm87/rev25?l...6&zoom=428-466

[CarAbuser]

Bank1 trims are going crazy during shift, looks like the bank goes lean.

Maybe a fuelling issue in cyl1 and a genuine knock event as a result? I've never seen post shift timing corrections that severe before.

[CKI@Motiv]

That split in trims though... probably leaking injector

[bradsm87]

Dammit. Earlier logs didn't have that stft split. The front 3 injectors are 2 year old index 12s. The rear 3 are 2 weeks old! I'll be broke if I have to buy any more. I'll lower the boost till I put the LPFP in and see if that helps.

[jt0407]

Is that 87 octane? I've never seen timing corrections like that on such low boost.

[bradsm87]

Is that 87 octane? I've never seen timing corrections like that on such low boost.

It never has timing corrections in steady state. Only from the hard shift. It's E21 and it takes WAY more timing than it did with 98 RON. I'm extremely happy with the knock resistance of the fuel. As the others mentioned, there is an issue at play.

[jt0407]

It never has timing corrections in steady state. Only from the hard shift. It's E21 and it takes WAY more timing than it did with 98 RON. I'm extremely happy with the knock resistance of the fuel. As the others mentioned, there is an issue at play.

Odd thing is, the corrections never clear up. Normally, they minimize a few hundred rpm, post shift.

[The Ghost]

I know a lot of cars do have a tendency to have a nasty single knock on one or two cylinders when grabbing the next gear suddenly. I don't want to take any timing out there because it doesn't knock there when in steady state.

I'm tempted to lower spark retard intensity 1 & 2 as we all know a -10 degree timing pull (ALL of the timing advance) is excessive for ANY knock event. I do like the feeling of having a huge safety margin though.

Are there any other/better tricks?

http://datazap.me/u/bradsm87/rev25?l...6&zoom=428-466

I was able to tune out shift knock by reducing timing in the high rpm mid/high load cells by 3-4 degrees (those cells now have 3-4 degrees below peak timing, mich may in fact be more than a 4 degree reduction from the values you have now). I watched the logs and found the cells which are hit during this shift.


YMMV

[bradsm87]

Odd thing is, the corrections never clear up. Normally, they minimize a few hundred rpm, post shift.

If it's a bad knock, it recovers slower.

[martial@mhd]

rail pressure is crashing.

[bradsm87]

rail pressure is crashing.

Yeah I've pushed it too hard. I've lowered the load target at high RPM until the fuel is nearly empty, then I'll put my new LPFP in.

[bradsm87]

Here's a log from today with boost tapering a little from 6000rpm and gear change a little earlier. This keeps it within the limits of the LPFP and rail pressure is OK. Lambda is fine throughout the run, including the gear change. Still two big timing drops.

Yes there is a difference between bank 1 and bank 2 STFT. I'm not concerned about it. It could be the age difference between the two banks' injectors. I don't know but the STFTs are doing their jobs. They both follow the same trend and hold their AFR target correctly. Other than post-shift, it accepts a lot of timing without any drops at all. I don't believe any individual cylinder to be lean. This time, cylinder 4 had a worse timing drop than cylinder 1 so i don't believe there to be an issue with any one cylinder.

Any other tricks for post-shift timing drops other than to reduce timing in that area? I don't like the idea of doing that because I'll lose performance in that area, regardless of if it's a constant run or post-shift.

http://datazap.me/u/bradsm87/rev28-p...2&zoom=476-553

[CarAbuser]

Here's a log from today with boost tapering a little from 6000rpm and gear change a little earlier. This keeps it within the limits of the LPFP and rail pressure is OK. Lambda is fine throughout the run, including the gear change. Still two big timing drops.

Yes there is a difference between bank 1 and bank 2 STFT. I'm not concerned about it. It could be the age difference between the two banks' injectors. I don't know but the STFTs are doing their jobs. They both follow the same trend and hold their AFR target correctly. Other than post-shift, it accepts a lot of timing without any drops at all. I don't believe any individual cylinder to be lean. This time, cylinder 4 had a worse timing drop than cylinder 1 so i don't believe there to be an issue with any one cylinder.

Any other tricks for post-shift timing drops other than to reduce timing in that area? I don't like the idea of doing that because I'll lose performance in that area, regardless of if it's a constant run or post-shift.

http://datazap.me/u/bradsm87/rev28-p...2&zoom=476-553

The boost mean spikes in that area quite considerably and you have a throttle closure post shift.

I had similar problems with throttle closures during shifts on my DCT. I just changed my load curve to taper more aggressively towards the line so when it spikes during shift it is matching the new boost target. Not an ideal solution but seems to be the only reason the factory tune doesn't suffer the same problem.

Try increase the boost target post shift and see if it changes the corrections.

[CarAbuser]

Also try running a little richer up top. That helped my post shift corrections. I'm typically at 11.3 toward the line. You are around 12.

[bradsm87]

Thanks for the ideas. I think you're on to something with the AFR target. I'll try going richer right at the end which should keep cylinder temps a little cooler for the shift.

There are a couple of undefined items of interest I'm looking at too. The first one I'll try is torque ramp rate after shift engaging. If it is what it sounds like, it might allow control of how quickly timing is re-introduced after a shift. Decreasing this rate should help.

[bradsm87]

I've finally fitted an upgraded LPFP and now my rail pressure and AFR is solid. Still big timing drops post-shift only and it looks like repeated knocks detected because the cylinders that have timing drops post-shift continue to further retard over a few logging points.

I'm really sure that the best and proper fix for this is to re-introduce timing after a shift a little more gradually, not instantly like it does. Is this what having having the non-stock value in "Disable Tq Reduction by Ign" does? I don't want to do anything silly to the transmission but what would re-setting this back to 00 like stock do?

[Sered]

It may not mean much but I keep that value as stock on my dct Z4. There is a post $#@! timing drop on 2-3 and it takes a good 1k rpm or so to fully recover. It does so gradually. I wonder if those shift bog tables also cover this stuff too

[bradsm87]

Setting "Disable Tq Reduction by Ign" back to factory settings did the trick! You can see in the log that the ignition is re-introduced over a couple of logging points, rather than instantly:

http://datazap.me/u/bradsm87/rev36?l...6&zoom=226-309

Ignore my other issues. The torque limit active through the midrange I'll fix in the next revision. I NEED to fix the boost spike during shift though. The stock charge pipe will NOT like the 20psi spike!

[Snelson0330]

Wait hold on, there is a FIX to this issue?! I thought it was normal?!

So if i am running the E85 BEF and want to correc the post shift timing drop for the 6at i can edit the flash and change this variable??

[bradsm87]

Wait hold on, there is a FIX to this issue?! I thought it was normal?!

So if i am running the E85 BEF and want to correc the post shift timing drop for the 6at i can edit the flash and change this variable??

I think I got lucky with that particular log. Subsequent logs still do have drops.