[BayMoWe335]

Originally Posted by Ilma

Not related to meth in any way......but on pump gas.

Here is a procede datalog running 14+ psi on a stock exhaust system. 3rd to 4th gear pull WOT.

Notice there is zero ignition correction. It is a straight flat line all the way across, so there is no active timing being retarded.

The result is a knock event at 6K rpm with timing being pulled by the DME from 12 degrees down to 8 degrees of advance.

Thereafter, into 4th gear, the timing stays low at around 4.5 degrees.

So this directly goes against what Shiv stated earlier? Response?

[dzenno]

Originally Posted by Ilma

But if it was going to "prevent" knock why would it start at 9.5 degrees?

Wouldn't be more conservative to start at 6.5 and rise upwards....testing for knock on the way up?

This also has the classic 3 degree drop that most knock events seem to register....so I don't know.

That IS weird that its starting at 9 only to induce its own retard a few tenths of a sec later...it probably knocked then on the first pull maybe due to not enough octane post shift while on the second run it pumped a tiny bit more fuel and avoided knock..

[dzenno]

Oh wait, LOOK WHERE THE BOOST SPIKE IS right after the shift...it didn't have enough fuel to meet AFR demand in the first graph for that boost spike so it knocked..but not on the 2nd run as boost spiked earlier when timing was lower (during shift)

EDIT: We know that those boost spikes aren't seen by the DME and in that instant if boost goes high and the tune doesn't request enough fuel to meet demand DME won't do anything about fuel, tune won't do anything, so it'll knock...make sense? if the spike really isn't seen by the engine (pre-throttle blade, when it closes) then i don't know how I'd explain it..

[Ilma]

Originally Posted by dzenno

That IS weird that its starting at 9 only to induce its own retard a few tenths of a sec later...it probably knocked then on the first pull maybe due to not enough octane post shift while on the second run it pumped a tiny bit more fuel and avoided knock..

Seems weird to me too.....but I see it once in a while on my logs.

Seems to me that it may be the DME testing the waters as you inferred earlier.....but it obviously failed on that first run.

That must have registered in the timing adaptations somewhere.

The second run seems to contradict what scheherazade stated earlier on...

"Timing targets are slowly adjusted by every knock. The adaptation takes time.
This will happen until the targets are such that there is no more knock."

[Ilma]

Originally Posted by dzenno

Oh wait, LOOK WHERE THE BOOST SPIKE IS right after the shift...it didn't have enough fuel to meet AFR demand in the first graph for that boost spike so it knocked..but not on the 2nd run as boost spiked earlier when timing was lower (during shift)

EDIT: We know that those boost spikes aren't seen by the DME and in that instant if boost goes high and the tune doesn't request enough fuel to meet demand DME won't do anything about fuel, tune won't do anything, so it'll knock...make sense?

Yes....I see what you are saying.

Good eyes!

edit....The spike is hidden from the DME but the knock is not....so it takes me back again as to why the DME didn't drop timing for the second graph based on a previous knock event.

[dzenno]

Originally Posted by Ilma

Yes....I see what you are saying.

Good eyes!

I eat a ton of carrots Have any other logs to confirm this theory? If this is true it means the engine IS seeing this boost spike unlike what we've known so far...

[Ilma]

Originally Posted by dzenno

I eat a ton of carrots Have any other logs to confirm this theory? If this is true it means the engine IS seeing this boost spike unlike what we've known so far...

Don't believe the DME is seeing the spike, but if the AF ratio is going lean for a split second then I can see the logic in a knock event occuring.....especially if the DME is riding along the inception of knock.

[dzenno]

Originally Posted by Ilma

Don't believe the DME is seeing the spike, but if the AF ratio is going lean for a split second then I can see the logic in a knock event occuring.....especially if the DME is riding along the inception of knock.

If the DME saw it I'd assume it'd request more fuel there...this is why i think DME doesn't see it/expect it...i wonder with a flash how this would be taken care of...this could be pretty serious when running high boost and shifting if true...

[Ilma]

Originally Posted by dzenno

If the DME saw it I'd assume it'd request more fuel there...this is why i think DME doesn't see it/expect it...i wonder with a flash how this would be taken care of...this could be pretty serious when running high boost and shifting if true...

I just checked that log with the timing dip and boost was still spiking at 15.7 psi when timing was re-applied post-shift.

On the surface it does look like a shifting-timing issue.

On the second graph, boost spike was down to 13.7 psi when timing was re-applied by way of comparison.

[scheherazade]

I really don't know how many events it takes for the DME to adjust anything.

Shift points have sudden drastic changes in multiple parameters.
Not really the 'mid pull' sort of situation that's easily repeated (gradual, slow change in parameters, similar from pull to pull).


How information is presented to the DME by the piggy is an interesting question.

Shiv has made his system work with the tools he has at hand (eg. CPS offsetting). You may just need to use them, as that's how it's meant to be used.


It would be interesting to see how your logs would look over the course of a week, making no changes to your car's settings.

-scheherazade

[Ilma]

Originally Posted by scheherazade

I really don't know how many events it takes for the DME to adjust anything.

Shift points have sudden drastic changes in multiple parameters.
Not really the 'mid pull' sort of situation that's easily repeated (gradual, slow change in parameters, similar from pull to pull).


How information is presented to the DME by the piggy is an interesting question.

Shiv has made his system work with the tools he has at hand (eg. CPS offsetting). You may just need to use them, as that's how it's meant to be used.


It would be interesting to see how your logs would look over the course of a week, making no changes to your car's settings.

-scheherazade

I am using CPS offsetting

So I am not sure what you mean.

But anyways, I agree, there are a lot of things happening simultaneously at the shift point and Dzenno has been keen enough to catch the probable cause of my knock event.

I have not changed my settings for over a week now. Boost is set at 14 psi, Ignition correction is set to 30% and autotune is turned off so that none of those parameters get tuned out.

[scheherazade]

Originally Posted by Ilma

So I am not sure what you mean.

The CPS offsetting was a generalized mention as being 'how procede accomplishes its task'.
(eg. Change AFR vs change timing, etc. Ways to skin a cat.)


I guess the next question that comes to mind is :

A) Does running more aggressive timing make more knock
B) If so, does the knock go down over time

C) Does running less aggressive timing make less knock
D) Does any knock come back over time


I honestly don't know if the DME has the granularity or range to handle every situation a tuned car will present.
I just know it's designed to try.

-scheherazade

[OpenFlash]

Originally Posted by scheherazade

A) Does running more aggressive timing make more knock
B) If so, does the knock go down over time

Yes
No.

C) Does running less aggressive timing make less knock
D) Does any knock come back over time

Yes
No

I honestly don't know if the DME has the granularity or range to handle every situation a tuned car will present.
I just know it's designed to try.

With piggyback, the DME is still operating in the same load ranges as a stock tune. For all the DME knows, it is running stock boost/airflow. Which is why additive fuel, TIMING and boost dc adjustments are needed by the associated piggyback. It's a very simple relationship here. Some would have you believe that the DME is a magical black box that just autotunes itself. When that couldn't be further from the case. Just look back the quality of tunes from the pre-CAN days and see how better control/mapping of these additive corrections have made these cars perform/drive.

Shiv

[Ilma]

Originally Posted by scheherazade

I guess the next question that comes to mind is :

A) Does running more aggressive timing make more knock
B) If so, does the knock go down over time

C) Does running less aggressive timing make less knock
D) Does any knock come back over time


I honestly don't know if the DME has the granularity or range to handle every situation a tuned car will present.
I just know it's designed to try.

-scheherazade

Answer to A) is yes if you consider reducing timing retard to make your timing more aggressive. I have tried this and do get knock events more often. I find that generally speaking, with 30% ignition correction I get a smooth knock free curve most of the time.

Answer to B) seems to be no. But I think that the combination of passive driving in traffic versus brief periods of WOT makes things too dynamic for the DME to adapt to. I can run on 2-5 psi in daily driving for long periods, then one day go for a few WOT's at 14 psi. Perhaps that is just too much range.

Running this current combination of 98 octane, 1 degree of timing retard and 14 psi seems to give me very nice timing curves most of the time.

I pulled one earlier this morning and it looked great. Yet of the two I pulled this evening, one of them knocked......but only at the shift point. So that may be an exception.

I think my engine is pretty happy with current settings, but I do suspect you are right that the DME is always testing the curve up to it's upper threshold.

I just don't seem to see any consistent timing drop between runs when a knock event happens. It's as if the DME is always testing the stock timing curve.

[scheherazade]

I would be interested to see logs, but I believe you on this one - if at least in regards to your experience.


Your current approach of feeding back from CAN data is undoubtedly more dynamic.


Feeding forward via inclusion of pre-emptive measures into the map itself is also a workable solution - albeit more static in nature. You did it yourself in earlier versions.

-scheherazade

[Ilma]

Originally Posted by shiv@vishnu

Just look back the quality of tunes from the pre-CAN days and see how better control/mapping of these additive corrections have made these cars perform/drive.

Shiv

+1

[scheherazade]

Originally Posted by Ilma

I think my engine is pretty happy with current settings, but I do suspect you are right that the DME is always testing the curve up to it's upper threshold.

I wonder about the quality of knocks during shifts.

eg.
You lift, so fuel flow drops.
Maybe the throttle shuts a little slowly, maybe the DVs don't pop right away, who knows, but you're lean and hot.
Maybe you diesel for an instant, albeit with so little fuel in the cylinder that it's not of terrible consequence.

Sort of like starting a diesel with ether spray into the intake when the battery is low and you can't quite get it to 'pep up' enough.
It pings like crazy, but it's also such a weak explosion.

That's not a definitive, or dismissive statement.
Thinking out loud...



I mention it because when I tried a V3 CAN (later V4 beta), I also had what appeared to be knock sometimes during shifts, that I couldn't make go away.
I learned to live with it, and it appears you have learned to live with it too.

-scheherazade

[Mike@N54Tuning.com]

Originally Posted by Ilma

Your statement sat in the back of my mind and didn't seem consistent with my experience.......so I went out and did another couple of datalogs just to double check.

I am running around 14ish psi on 98 octane....no meth at all.....1 degree of ignition correction at mid rpms.

This first pull of the evening shows a knock event just as I shifted into 4th gear and timing immediately fell from 9.5 to 6.5 degrees.

So according to your thinking....the DME should remember and adjust to this knock event by adapting timing.

Drops are not always KR related. Timing is mapped to DME boost among other factors by the ECU and will always move proportionally.

Mike

[Ilma]

Originally Posted by Mike@N54Tuning.com

Drops are not always KR related. Timing is mapped to DME boost among other factors by the ECU and will always move proportionally.

Mike

OK.....but the piggy's fool the boost reading to the DME

So how does that work out?

I'm pretty sure that the DME didn't know my boost was at 15 psi when my timing dropped.

But in any case....does that then explain the typical timing drop at the vanos changeover. Is this a DME induced adjustment rather than a knock event?

These are good questions.

[Ilma]

To those who want to close this thread......it seems to be a good discussion thus far and the moderators are doing a good job keeping it civil.

If we do the same, then why can't we continue this very relevant discussion about timing?

If affects all of us.

[cstavaru]

Originally Posted by Sevak

Yesterday I wrote something now its deleted. I have contact E90post twice whitout an answer. I want it closed and the hole thread deleted.

I am really sorry about your problem, but I am not sure you have the right to demand the deletion of other people's opinions...it's like someone which has been in an accident demands all TV stations not to report about the incident.

[lyndon_h]

Originally Posted by cstavaru

I am really sorry about your problem, but I am not sure you have the right to demand the deletion of other people's opinions...it's like someone which has been in an accident demands all TV stations not to report about the incident.

Even it was common policy, I'd be pissed if I was a moderator and someone was demanding I close the thread.

Back on topic. I think this is one of the most insightful turbo discussions on this site. Too bad it had to come on the back of someone's misfortunes. However, it was probably the only thing that would strike enough views to have a discussion evolve.