Trying to tune out stumble

[MasAyinde]

Yeah, CL/OL delay table is all 0s. You're saying that you still get knock there no matter what you changed, and you just don't want your ECU to learn to pull timing to try to fix it?

[rtbrjason]

Yeah, CL/OL delay table is all 0s. You're saying that you still get knock there no matter what you changed, and you just don't want your ECU to learn to pull timing to try to fix it?

 

In the 2800 rpm and 1.3 load yes. I've pulled more timing from there than the ECU is trying to pull and it still didn't go away. It did run poorly though with that low of timing.

 

I already run fairly low timing and AVCS in that area (lower than stock) and don't feel like it's a real knock event or in a high enough load to destroy the engine.

 

Every other knock event has been easily taken care of with a touch of fuel or slight reduction in timing. I don't get any real knock above 1.3 with the exception of that one instance at 2800.

[MasAyinde]

I'm thinking the issue might not be with timing, but rather fueling. In looking at all of my logs, the only knock that I have is when throttle is quickly applied. I see it when I pass cars, and it's there as soon as I give it gas after shifting during spirited driving. It seems to not be at a specific load/rpm.

[sjwelna]

I'm thinking the issue might not be with timing, but rather fueling. In looking at all of my logs, the only knock that I have is when throttle is quickly applied. I see it when I pass cars, and it's there as soon as I give it gas after shifting during spirited driving. It seems to not be at a specific load/rpm.

 

That sounds like it could be a tip-in enrichment issue, is it going lean?

 

-Steve

[MasAyinde]

Yeah, the A/F ratios that I see when I hit the gas do look a little leaner than the target OL A/F ratios.

[MasAyinde]

So I recently flashed to stage 2 and I'm not seeing any knock at steady WOT pulls, but there is still knock when throttle is quickly applied. I was reading up on tip-in enrichment and I saw that you should adjust the values based on how much A/F correction + A/F learning there is. The only thing is, there is none when I see knock. The knock I see is around 2.4 load @ 43-4400 RPM. I've hit those cells a bunch of times in the log attached, and there is no knock except when the throttle is applied quickly. Instead of looking at A/F correction, should I look at the few cells before and compare the A/F ratio to the target A/F ratio and use that percent difference to calculate the necessary change for tip-in?

 

Also, A/F is bottoming out at 11.14 when target is 10.7. Is this just a result of me not having a WBO2 that can read down that far?

romraiderlog_20110811_061702.csv

[sjwelna]

Without a true wideband like a PLX or Innovate/etc you really shouldn't play with the fueling tables IMO. The stock "wideband" will only display down to 11.1, i.e. anything below that and it still reads 11.1. You're probably running about 17-18 psi from that log? I know I'm seeing around 2.5x load on my current tune at 15 psi. Can you switch out Base Ignition Timing for Total Ignition timing? Have you rescaled your knock correction advance table so it reads all the same number (i.e. 6) for easy tuning yet? It definitely looks like you're getting knock when you tromp the throttle, but when you look at feedback knock correction it's adding timing back until you shut the throttle and whip it back open again. It could be a fueling issue or it could be just that you might be targeting too much timing in that rpm/load range after your shift or when you punch it. The ECU is basically in "oh sh*t" mode going from 0.4-2.4 load in 3 lines (984-986). That's a huge swing through the tables it has to interpolate, have you tried just pulling out a couple degrees timing in that range? What happens if you roll on the gas slow?

 

-Steve

[MasAyinde]

The entire knock correction advance table isn't at 5.98, just a large area on the right. I don't believe that timing is the sole cause of the knock; I feel like I would see less severe knock around those same cells at full throttle if that was the case, but I see none. Assuming that the stock O2 sensor is accurate between 12 and 14, it looks like it's staying way too lean when I get on the throttle (06:20:13.765 and 06:20:13.953). Being that lean, I'm not surprised that it knocked at all.

 

I can try rolling on the gas slow, but I doubt that I will see any knock. I hit those same cells at 06:20:55 and there is no knock. Since there is no issue with timing when I'm not stomping on the gas, I think I may just avoid doing that until I can get a WBO2 and can properly tune the fueling tables.

[sjwelna]

The entire knock correction advance table isn't at 5.98, just a large area on the right.

 

Ok good.

 

I don't believe that timing is the sole cause of the knock; I feel like I would see less severe knock around those same cells at full throttle if that was the case, but I see none. Assuming that the stock O2 sensor is accurate between 12 and 14, it looks like it's staying way too lean when I get on the throttle (06:20:13.765 and 06:20:13.953). Being that lean, I'm not surprised that it knocked at all.

 

Have you changed your OL/CL delays to zero? If you haven't I would zero them out and see if that helps. The o2 reading is saying it's lean (12:1 or higher) for the few cells as you hit the throttle because it's likely still in closed loop and targeting 14.7 for a/f ratio.

 

-Steve

[MasAyinde]

OL/CL delays are all 0. I'm pretty sure that the solution to this problem is with tip-in enrichment; is it really that risky to play around with if I don't have a wideband?

[NSFW]

It's not risky to mess with it, but it is very hard to get right, even with a wideband. I've tinkered with it a few times and I'm still not happy with mine.

 

But on the other hand, tip-in is only active for a brief period after a rapid throttle change, so it doesn't come into play very often. I could see tip-in causing a single brief stumble if/when you stab the throttle, but that's all. In the log you posted, the throttle opens gradually, so I'd bet that the tip-in code probably didn't even execute.

[MasAyinde]

Where do you see it opening gradually? I'm looking at line 983-984 where it opens 83%, and that's where I'm seeing the knock. I think I'll get my math on and see how much tip-in needs to be enriched to get those A/F ratios where they need to be at big throttle changes.

[NSFW]

My bad, I was looking at the previous log. Yes, a movement that abrupt will trigger tip-in.

 

For tip-in, you really should get the latest RomRaider, as it has a new "Fast Polling" checkbox that gives about 4 times as many rows per second. That's pretty cool in general, but especially for transient stuff like tip-in.

 

If your ROM ID is A2WC522N, I can give you the XML for an couple extra parameters that will make it a little easier to see how much tip-in is being used.

[MasAyinde]

I have 521N. I heard something about Romraider 1.29 being the last to support the VAG-COM cable.

[MasAyinde]

Nevermind, I'm thinking EcuFlash. I'm going to DL the new RomRaider and get to logging

[NSFW]

I have 521N. I heard something about Romraider 1.29 being the last to support the VAG-COM cable.

 

I think you're thinking of EcuFlash. RomRaider is still at 0.59 or something.

 

521N and 522N use the same table definitions so chances are pretty good that these logger parameters will work for you:

 

               <ecuparam id="E815" name="NSFW Tip-In Applied Counter" desc="Number of iterations that tip-in has been active.">
                   <ecu id="2F12785606">
                       <address length="1">0xFF59F0</address>
                   </ecu>
                   <conversions>
                       <conversion units="Iterations" expr="x" format="0" />
                   </conversions>
               </ecuparam>


               <ecuparam id="E817" name="NSFW Tip-In Cumulative" desc="Total amount of tip-in enrichment added by all iterations.">
                   <ecu id="2F12785606">
                       <address length="4">0xFF59EC</address>
                   </ecu>
                   <conversions>
                       <conversion units="Iterations" storagetype="float"  expr="x" format="0.0000" />
                   </conversions>
               </ecuparam>

 

Worst-case, you might just see garbage numbers in your logs.

 

In theory, the "tip-in applied" counter will increase, pause at 5, then drop to zero. In practice it's usually 0 or 5 because it moves so quickly, even with fast polling. The cumulative value behaves similarly.