Occasional Parking Lot "Buck" After Startup

[CCS86]

I have an occasional issue that is hard to reproduce, but I got a good log of it today.

It seems to happen when the car has been running less than a minute, and generally when it is warm to hot. I will be cruising through a parking lot, decelerate, then when I pick the pedal back up (barely), I'll get a big "buck".

I was thinking it was an overactive transient accel gain, but this log makes it look like something else:

Buck.jpg


Just prior to the ETC spike:

• ETC TQ is steady at -12 to -10 ft lbs
• Scheduled TQ is stable at 40 ft lb
• Desired load spikes to 0.57 out of nowhere, so ETC follows
• MP0 Inverse table calls for ~0.1 load at this RPM and a 37 ft lb scheduled TQ
• Cold start emissions reduction has dropped spark to -1*

My only guess at this point is that the desired load spike is in response to the spark reduction. But that seems crazy as a commanded 0.57 load corresponds to a scheduled TQ of over 220 ft lbs, over 5 times the amount shown in scheduled torque, not a modest increase because of efficiency compensation.

[MRRPMBRP]

Not sure it matters, but in that screenshot you have over 6 degrees of throttle angle error. On mine, that's way out, I try to stay under 2 degrees, and most of the time I'm under .5 degrees.

[CCS86]

Good eye, but I think in this case the ETC angle error is a result of this bizarre spike in desired load / MAF, not the cause of it.

In similar conditions (ETC vacuum, effective area, ETC angle, etc) my steady state ETC error is near 0. It always tends to spike in transient conditions, which I attribute to lag between "desired" type parameters and measurable parameters.

[superman07]

So I assume a vanilla torque problem would be more temperature agnostic. So is it a combination of a lambda modifier / spark correction both going in the wrong direction. Virtual VE for that area? It seems like a couple areas would need to be off in the same direction to give you a spike that large.

Patiently waiting for Murfie to chime in

[CCS86]

So I assume a vanilla torque problem would be more temperature agnostic. So is it a combination of a lambda modifier / spark correction both going in the wrong direction. Virtual VE for that area? It seems like a couple areas would need to be off in the same direction to give you a spike that large.

Patiently waiting for Murfie to chime in

Yeah, this seems like a huge miscalculation.

Temp doesn't seem to drive it so much as engine runtime < 1 min. It did it again to my today at an ECT of 168*F

I seem to have only one table related to VE: [ECM] 32044 - VE Correction vs. ACT vs. ECT. It doesn't have a runtime component to it, and given the ACT and ECT when it occurs, the correction factor is like 2%.

Maybe there is a secret in one of the new tables about to drop for us Copperheads. The cars seem WAY more likely to throw IPC TQ errors when cold / low runtime. Maybe there is something downstream of ETC TQ Req that is scaling up desired load/MAF based on ERT.

[MRRPMBRP]

Your spark went super low in almost no time at all, then maybe it decelerated and your foot gave it a little gas, this would account for the load increase, since throttle opened, and maf went up. I think the cold start emissions reduction caused the immediate timing decrease, starting the whole thing. Can you see what emission reduction timing is on the table?

[GapRider]

I had the same issue playing with tunes. What I saw was cold start emissions timing jumping in (which was a lot lower) causing eng brk tq to dive as spark dived. throttle would jump open to compensate and the buck.
since i'm a newb, I just increased the cold start spark tables to 25 which made the buck go away. the downside is idle hovers pretty high for several seconds after startup because i don't get the normal super low emissions cold start spark retard when startup idle is expecting it. I think the high idle speed/air and low spark coincide for stock tune during emissions reduction on startup and for several seconds.

[CCS86]

Your spark went super low in almost no time at all, then maybe it decelerated and your foot gave it a little gas, this would account for the load increase, since throttle opened, and maf went up. I think the cold start emissions reduction caused the immediate timing decrease, starting the whole thing. Can you see what emission reduction timing is on the table?

I have done everything in my power to disable that mode, but it still shows up. Again, hoping for some fixes in the upcoming update.

I'm not sure whether it is solely the cause of the spark dropping though. I can repeat that behavior by blipping the throttle slightly from idle:

spark drop.jpg

I tried changing the Spark Only Torque Ratio vs. Requestor: for driver demand to 0.5 from 0.0, but it still drops way down like this.




spark drop.hpl

[GapRider]

I have done everything in my power to disable that mode, but it still shows up. Again, hoping for some fixes in the upcoming update.

I'm not sure whether it is solely the cause of the spark dropping though. I can repeat that behavior by blipping the throttle slightly from idle:

spark drop.jpg

I tried changing the Spark Only Torque Ratio vs. Requestor: for driver demand to 0.5 from 0.0, but it still drops way down like this.




spark drop.hpl

the first log cold start emission spark jumps in and spark dives, tq dives and throttle snaps open (I made the whole table 25 to fix)

The second log blip is like the blip required in some cam tunes - causes tq src to go from driver demand to Target N and spark src from idle feedback to torque control. maf goes way above desmaf in the excursion. My cam tune use to do that until I adjusted dashpot values to actually work (even though tq decel is enabled in my cam tune). I don't know what to do there
hope there is some help in there

[MRRPMBRP]

Does yours have 44600? I keep mine zeroed out, it will also go into emission reduction, but nothing happens, spark stays the same.

emiss red.jpg

[CCS86]

Does yours have 44600? I keep mine zeroed out, it will also go into emission reduction, but nothing happens, spark stays the same.

emiss red.jpg

I do and mine is also zeroed.

I still think DD TQ control could be causing the spark dip.

The spark dip could just be circumstantial, and not the cause of the giant "desired load" spike.

[Thatwhite5.0]

Spark Advance is following the cold start max retard for Drive. It's in the negatives. Mapped Point 10 that would be the distance for emission reduction, even if it wasn't zeroed out, it doesn't have timing that retarded.

I think they changed the name of the table. It's Emission Reduction Base Drive. Mine used to do the jump with the same torque error. I increased the max retard during cold start so the cells that had -5 now had 12 or 13, increased some driver demand. I think you have a formula for A/D counts. You have that huge angle error just because Engine Brake Torque is -45 and your Torque Request is -5, that gave you -7 throttle angle error. Then after the rev your Engine Brake Torque goes to + to 60 and your Torque Request stays at -5, then your throttle angle error ends up around +7. I use angle error to help tell me which way to go. - means DD is more than Engine Brake Torque, + means DD is less than Engine Brake Torque. Adjust Torque Tables, Driver Demand, or timing, whichever.

[CCS86]

Spark Advance is following the cold start max retard for Drive. It's in the negatives. Mapped Point 10 that would be the distance for emission reduction, even if it wasn't zeroed out, it doesn't have timing that retarded.

I think they changed the name of the table. It's Emission Reduction Base Drive. Mine used to do the jump with the same torque error. I increased the max retard during cold start so the cells that had -5 now had 12 or 13, increased some driver demand. I think you have a formula for A/D counts. You have that huge angle error just because Engine Brake Torque is -45 and your Torque Request is -5, that gave you -7 throttle angle error. Then after the rev your Engine Brake Torque goes to + to 60 and your Torque Request stays at -5, then your throttle angle error ends up around +7. I use angle error to help tell me which way to go. - means DD is more than Engine Brake Torque, + means DD is less than Engine Brake Torque. Adjust Torque Tables, Driver Demand, or timing, whichever.

Here are my "emission reduction base ..." tables:


emissions red base.jpg

[CCS86]

My current theory is that this behavior is caused/governed by a table we can't yet access.

The presence of a table which added frictional compensation based on ERT or ECT would explain this huge jump in desired load/MAF which causes my buck. It would also explain why cars are much more likely to throw IPC TQ errors when cold.

If my car has been running for even 30 seconds, I don't get any TQ errors. But, if I pull away say 10 seconds after startup, I can trigger an IPC error every time.

Eric should be kicking out this new definition update for my strategy very soon, and I will be digging into the new stuff heavily searching for this mythical table.

[MRRPMBRP]

Keep in mind, that the ecu will jump to the lowest timing setting in the timing tables.

[CCS86]

Keep in mind, that the ecu will jump to the lowest timing setting in the timing tables.

No timing tables have negative spark advance though.

[dtownmikebrown]

My current theory is that this behavior is caused/governed by a table we can't yet access.

The presence of a table which added frictional compensation based on ERT or ECT would explain this huge jump in desired load/MAF which causes my buck. It would also explain why cars are much more likely to throw IPC TQ errors when cold.

If my car has been running for even 30 seconds, I don't get any TQ errors. But, if I pull away say 10 seconds after startup, I can trigger an IPC error every time.

Eric should be kicking out this new definition update for my strategy very soon, and I will be digging into the new stuff heavily searching for this mythical table.

Ever make any progress on this? I'm having an similar problem which seems related to clutch engagement. It happens consistently from around ten seconds to a little over a minute after startup. Throttle input doesn't seem to make much of a difference, nor does engine temperature. As soon as there's the slightest bit of friction between the clutch and the flywheel the throttle snaps open. If I try to engage the clutch too fast it bogs instead. I changed the cold start emission reduction spark tables but that didn't make any difference. It's mildly inconvenient that I have to wait at least a full minute before driving away to avoid potential embarrassment.

AnnoyingBlip.hpl

[CCS86]

No progress.

I think there is some sort of a run-time based torque loss table currently hidden from us. This issue has never occured after this brief time window after startup.