LSJ Tuning Woes

[Gtstorey]

I always point out there is a lot more in the ECM that can’t be touched or seen with HPT. On some of the ECM’s you sometimes just have to live with some bugs. It probably is a cat or o2 test being ran.

 

[Rauq]

I don't have any more solid or useful input than to agree with the above, that the random commanded AFR dips are just an LSJ thing that we can't see and/or can't control in HPT. The P12 is not very well supported.

I will add, I'd come across clues in various LSJ tuning forums that linked commanded AFR dips to misfires, and to maxing out the COT threshold table. I was having some idle misfires due to an idle RPM target of 800 which was just a little too low for 80# injectors, even with a BRFPS. Bumped target idle up to 900 and the idle misfires went down as did the commanded AFR dips. I will soon be trying some changes to the COT threshold table that I've seen elsewhere, where the adders are zeroed out but the thresholds are not maxed out (maxed out shown, other tunes I've seen have all the threholds at 2200).

One more point regarding living with the bugs, it does not appear to me that super rich commanded AFRs throw off fuel trims, if you're doing fuel trim tuning, and it shouldn't throw off WB tuning either. If it's commanding 10.10 AFR like mine did and the WB is reading 11.11, that's still 10% rich for that point on the VE or MAF chart, and that can be dialed in.

Where are you seeing whack commanded AFRs, at idle or elsewhere?

 

[Chubbs]

Where are you seeing whack commanded AFRs, at idle or elsewhere?

Yes to both. The specific case I was able to reproduce was while cruising under light load around 3000rpm. As the tach swept up and down across 3000rpm, it would often go full rich. Normally, I would think it was transitioning between the VE and MAF tables, but I had VE disabled due to the tune. I was not able to find anything odd in the tune around 3k rpm.

I have also seen it while coming to a stop light when the engine is hot after cruising for a while. Sometimes it will just go full rich while sitting at a stoplight. It seems to spit and sputter a bit when taking off after it happens, but otherwise runs ok once moving again.

I've been researching a lot about the IVT gain tables, and I suspect my problem lies in there somewhere. A lot of the guys that tune LSJ's set that table to all "1" while tuning so it doesn't interfere with the commanded mixture. Goat Rope Garage didn't mention it, but he wasn't specifically tuning an LSJ either.

 

[Chubbs]

Getting way, way deep into things here....I haven't seen injector offset discussed on the forums before, but it sounds pretty darn important. From the tune I got from the HP tuners repository, I found this:

Look at that weird spike. The discontinuities. I have never seen a graph look like that, and there is no way it's correct.

Obviously, I could fix it, or use someone else's table and copy it over to mine. But those will always be someone else's values, and not my values. Where do I get the "right" values? If I buy injectors, will they come with documentation about their offset at various voltages and pressures? I'm pretty sure if you get flow matched injectors they come with flow documentation (also important), but I have not seen offset documentation anywhere.

By the way, I have 60# Siemens injectors, which I am reading also isn't great for our engines. The spray pattern is all wrong.

 

[Gtstorey]

There is always the chance that you have an issue with the o2 sensors that "confuses" the ECM in certain circumstances and causes it to act weird. Maybe the o2 heater is starting to fail and the sensor isn't quite warm enough. Maybe not bad enough to throw a code but enough that the computer reacts.

Don't buy injectors that that don't come with good GM data or are not confirmed to be GM injectors that you can pull data from a stock tune. Both of the FIC's have GM data available or depending on flow rate needed, pick injectors from another GM application that will work. I guess that ZZP can supply GM data for what they sell.

Anything currently in the HPT repository should be stock, but there may be some pieces that have slipped through. Or did you pull a file from before they removed the modified tune? I wouldn't worry to much about the spike at 7v since you really will probably never be there. I'm not sure about the dip at 110kPa, but it's pretty minor. What does your original tune show for this table? You might want to start your own tuning thread with more details about where you are in the tuning process and more details about your car modifications.

 

[Gtstorey]

And also welcome to the world of real tuning that doesn't show up in the youtube videos. You can only partially go by videos for GM v8 LS engines since all of the ECMs have there own peculiar differences.

 

[ctuinstra]

That spike is in the stock tune. I chalked it up to a boo boo at the factory. Also never worried about it since the stock setting, and my current setting, is to use the VAC chart and not that MAP chart. So it never comes into play. If I felt that I needed to use that chart, I would have smoothed that spike out. I’ve heard of boo-boos in stock files

 

[Gtstorey]

Yeah, they don't worry much about the sections it doesn't actually use.

 

[ctuinstra]

This is only happening when you are forcing it to stay in Open Loop? So it’s only when trying to tune it? It may be simply that it’s not happy about something being forced in Open Loop, which is not its normal state. I wouldn’t worry too much about it unless it’s happening in Closed Loop Also and causing drivability problems.

as other have stated, during tuning, if the ECM commands 10, and the system is able to produce a 10 AFR, all is good. I understand you are asking why should it be commanded 10. That does seem odd, but I doubt it has much to do with your normal tune.

 

[Chubbs]

This is only happening when you are forcing it to stay in Open Loop? So it’s only when trying to tune it?

No, sorry, I should have clarified.

I first noticed it while I was tuning (as I had my laptop plugged in). But it also happens when I'm "done" tuning, have turned on all the sensors and am using the VE and MAF tables, like the stock car does. It runs well most of the time, but every once in a while it goes pig-rich, and i've never been able to figure out why.

It also seems to be somewhat inconsistent from day-to-day. I'll do some VE tuning on monday, and get it "close". Then I'll try again on wednesday, and it's way off. I understand the weather changes, and that's normal....but the ecu should compensate for temp differences, right? or at least get it close again. I'm seeing 10% swings depending on when I tune.

That's why I started poking around with the injectors. I suspect something is wrong there that is making the calculations off, so the ECU can't compensate correctly.

Along those lines, could anyone share a working tune with 60# injectors? I have found a couple on the forums (one was Ghostknifes) but it appeared to be early in the tuning process before it was finished. I'm not sure I've ever seen a completed tune with 60# injectors.

 

[ctuinstra]

How long of a drive are you doing for the data gathering during tuning? I was doing really short drives, sometimes just a minute or two and having problems with my tune bouncing all around. Then I started making drives for 30+ minutes or long and got a lot better results. The more cell hits, the better the overall averages.
Also are you smoothing the VE tables?

 

[Gtstorey]

Are you using any data filters when tuning?

 

[Chubbs]

How long of a drive are you doing for the data gathering during tuning?

I usually let the car warm up for a minute or two (not fully warmed up), then start loging, and drive for between 10 and 20 minutes. Maybe it isn't long enough?

Also are you smoothing the VE tables?

I usually do two or three drives without smoothing, end up with some spikes, then smooth it. So yes, my VE table is smoothed.

Are you using any data filters when tuning?

No, I'm not really sure how to do that. My scans are set to require 10 cell hits before registering anything.

 

[Gtstorey]

You should have it fully up to temp before logging. Then you need to see what is driving the swings. Is it going rich or lean during deceleration or does it go into DFCO that gives you bad data? Maybe post a couple of logs showing it getting close followed by it being farther off.

 

[Rauq]

for reference, my VE scanning has the following filter:

[50091.155]>0.01&-0.1<[50090.156.slope(250)])<0.25&[50030.90]<215

accelerator pedal greater than 0% (filters out idle or coasting in gear)
accelerator pedal slope (rate of change) between -10% and +25% (filters out quick pedal transitions)
MAP less than 215 kPa (that's where the VE table stops)

If you're logging coolant temps and want to exclude under 140, you can add the below to the filter. I just don't hit record until I'm up to temp.
&[50010.242]>140

 

[Chubbs]

I always point out there is a lot more in the ECM that can’t be touched or seen with HPT. On some of the ECM’s you sometimes just have to live with some bugs. It probably is a cat or o2 test being ran.

I drove quite a bit over the 4th, and I think you got the right answer. There is something odd going on that is perhaps just "normal".

Periodically, my o2 sensor status will go from closed loop to "OL - FAULT". The ECM then commands a 10.1 AFR for a few seconds, then it goes back to CL mode. I have no idea why, but the idea that the ECM is running some sort of test seems reasonable. I occasionally get a CEL for P0300 (random misfire), but the code does not seem to coincide with the OL - FAULT condition.

It seems like the ECM is forcing a fault condition to do....something....then goes back to normal.

Since it doesn't seem to affect drivability, I am going to choose to ignore it for now. It's weird, but also don't seem harmful, as without the AFR gauge and an active scan, I never would have noticed in the first place.

 

[Rauq]

Are you logging misfires?

 

[Chubbs]

Are you logging misfires?

Not that I know of, I've never tried before. Initially I was convinced my misfires were because of screwy mixture, and would go away once I got it dialed in. As I've dialed my mixture in better and better, the misfires persist. Probably time to troubleshoot them.

 

[Rauq]

Not that I know of, I've never tried before. Initially I was convinced my misfires were because of screwy mixture, and would go away once I got it dialed in. As I've dialed my mixture in better and better, the misfires persist. Probably time to troubleshoot them.

Nearly all of my misfires have been due to low idle with big injectors, or coughs on quick shifts. Neither of those, obviously, point to issues in the power-making portions of the tune, but I mention them because the idle misfires seemed to cause whack AFRs from time to time.

 

[Gtstorey]

Misfires will pretty much always cause wacky AFR. It will show as lean since the o2 is measuring oxygen and there is excess since none of it got burned during the misfire.

If you continue to tune with misfires, you need to add a filter to remove that data.