Just went from OLSD to CLSD, some questions

[Kevin Doe]

EDIT: Sorry for the wide pics, I have a widescreen monitor.

For the first time I switched from open loop speed density to closed loop speed density. I figured I might pick up a mpg or two by doing so.

I during my last fine tune of my VE talbe I did some minor tweaking. I did probably an 1.5 hour drive hitting a lot of the cells. The corrections were small.

Here you can see my BENs from my logging process. In this screen shot I'm hiding cells with lower than a cell count of 50, and filtering the data. Most cells have 100 counts or more. Many of the cruising cells have 1000 data points.
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/autoVEBENs.jpg


Then I took that BEN data, applied it, and did some smoothing. I also used some WOT pulls and hard engine deceleration logs to get the cells you cannot sit at steady state.

This is a screen shot of the % change from the tune while logging, to the new tune I just loaded. As you can see, in the main part of the map where you'd be cruising the % changes are very small, and very close to the BENs from the logging. Positive numbers mean that I corrected it in the leaner direction, just as the BENs told me to do.
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/autoVEcorrections.jpg


So I took that new map, put it into a new tune. Then I turned on the closed loop fueling, and turned DFCO back on. Went out and drove maybe 5 seperate times, for a total of 2 or so hours of driving. Then I took a screen shot of my long term fuel trims for bank 1 and bank 2. I am very very confused by how far positive they are.

Here are the LTFTs
Bank 1
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/LTFT1.jpg

Bank 2
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/LTFT2.jpg

When I look at some logs to see what the wb02 says its doing in closed loop I see that the BENs come out to about 0.98. I'm really confused on what is going on here. When cruising I see that the FUELSYS says that I'm in CL-Fault mode, and switches to OL when I get on the gas heavy. Is CL-Fault mode correct for a speed density setup? I'm thinking it is, but not sure.

My setup consists of basically a big cam only motor. I have long tubes, 3" exhaust w/o cats, and a dual 3.5" intake. Pretty basic setup. I have a texas speed MS4 cam.

I'm not sure what is going on these trims, and why they say my maps are sooooo lean. Any help or input here would be greatly appreciated. Also, my WOT is spot on now, and I don't want any stupid trims to be added to them making it overly rich.



I have attached many files. Here is a description of the files I've attached.

5-3-2008_002.tun is the baseline tune I was using when I did the VE BEN logging.
5-4-2008.tun is the new tune that I created from the logged BENs. I also switched to CLSD, and turned DFCO back on.
5-4-08 VE BEN main.efi is the long log from my VE BEN logging.
5-7-08 Work Commute 2.efi is a ~20 minute drive to work. Following this last drive is when I took the screen shots of the LTFTs.

[SSpdDmon]

Assuming you have headers that move the O2's back down the exhaust path...

Do you have a seperate bung for your WBO2 so you can run it along with the NBO2s? If so, you may have to reduce the O2 switchpoints until closed loop maintains a stoich AFR. Try dropping them 8% for starters and see where it puts ya. :)

[Kevin Doe]

I have 1 7/8" long tube headers. The NB02 bungs are pretty far back. Probalby about 1/3 back on the transmission, or more.

Here is a pic of my headers before I installed them.
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/RX-7LS1Headers2.jpg

8% on the switchpoint? I'm not on a computer where I have EFI Live Tune tool. If I remember correctly the stock switch point is 450. So you propose trying 414 for those values. Its a curve if I recall correctly. Change all of the values to 414 and try that. Any explanaion of why the switch point needs to be different based on being further downstream.

I do have seperate bungs for the NB02s and WB02. I have the WB02 after the X pipe to try and get an average AFR.

[SSpdDmon]

Yeah, try setting them to 415 or so. It should lean out the point the fuel trims hang around. Think of the STFTs as a sin wav. The STFT's go + and - during closed loop. The average of the STFT's can be found by drawing a line down the middle of that sin wave (green lines below). If the line you just drew is above what the PCM thinks is stoich (i.e. the switch point - red lines), it assumes you're running rich and the LTFTs go in the negative direction. If the average line is below that switch point, it assumes you're running lean and the LTFTs add fuel.

In your case, the LTFTs are adding fuel. So, you need to lower the switch point line. Make sense?

[Kevin Doe]

A few more questions.

1. Why try 8% less (415)? Why did you pick that number.

2. I see that you can adjust the LFTF boundaries on for both the MAP and RPM boundaries. Since I have a large cam, does it make sense to update the boundaries to more reasonable values that I see with my big cam. If this is commonly done, can you suggest some boundaries to try. In the meantime I'll research and see what I can find on this subject.

Thanks!

[SSpdDmon]

I suggested 8~10% because of someone else's recent post in another thread (that I can't seem to find right now). Also, because it's not that big of a change and you'll be monitoring the results with your WBO2. My intentions and suggestions are rarely 100% accurate....but rather more in line to get you going in the right direction. :)

As for adjusting the boundaries, if the tune is spot on for all of the MAP/RPM's you drive in, why would you need to move these? If you do, I'd bump them up a little - but not a whole lot. Just my $0.02...

[Kevin Doe]

In open loop I was spot on, so I don't see why closing the loop would change things otherwise. I'll report back with my findings!

[5.7ute]

The problem stems from the narrowband being sensitive to exhaust gas temperature. Moving the sensor further away from the cylinder head places the sensor in much cooler exhaust gases, changing the voltage output. In the attached chart I got from the innovate site you will see that sensor output can be between 150mv to 650mv at stoich. This will give ample room to set the switchpoints to suit your new sensor position.
By using the highlight function in the scan tool you should be able to average the O2 sensor voltage to give an idea of the switchpoint. The same highlighted area will give you an average AFR off your wideband, telling you if you need to adjust the switchpoint up or down. Logging GM.DYNCYLAIR.DMA will tell you which CL mode you are using so that you adjust the correct switchpoint.
Hope this helps,
Mick

[TAQuickness]

Wouldn't it stand to reason that if you disabled CL fueling the NB's voltage should draw a straight(er) line for a given AFR? (Kinda like when NB's peg at 800+ mv at WOT)

If so, you could monitor AFR with the WB and associate stoich to the respective NB voltage. At that point you would know which way to move your NB switch points.

[SSpdDmon]

Wouldn't it stand to reason that if you disabled CL fueling the NB's voltage should draw a straight(er) line for a given AFR? (Kinda like when NB's peg at 800+ mv at WOT)

If so, you could monitor AFR with the WB and associate stoich to the respective NB voltage. At that point you would know which way to move your NB switch points.
http://www.motownmuscle.com/forums/images/smilies/mmm.gif Assuming you monitor based on which CL Mode you're in, I don't see why not.

[TAQuickness]

Enlighten me on the CL mode. I've been NBless for 3 years now and am drawing a blank

[SSpdDmon]

Essentially, you can set different switch points for different grams/cyl. It's all under the Engine>Fuel>O2 folder IIRC.

[TAQuickness]

muchas gracias. i feel much less a tard now

[5.7ute]

Wouldn't it stand to reason that if you disabled CL fueling the NB's voltage should draw a straight(er) line for a given AFR? (Kinda like when NB's peg at 800+ mv at WOT)

If so, you could monitor AFR with the WB and associate stoich to the respective NB voltage. At that point you would know which way to move your NB switch points.

Trust me to overly complicate matters, but you are dead right. That would be much easier. Like you TAquickness I havent had a narrowband in my car for a few years now so I am basically running on theory.
Cheers Mick

[Kevin Doe]

Ok, I moved my switch point from 450 across all CL modes to 425.

My trims did move in the correct direction (down). They all look pretty good besides a few cells in the matrix. The cells that look wrong to me are in the 1400-2200 column, and in 37-57 kPa, and 57-77 kPa cells. I took these screen shots after ~45 minutes of driving, during 2 ignition cycles. Not sure thats really enough time to settle out or not, or hit every operating condition required. But here they are.

Bank 1
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/5-8-08LTFT1.jpg

Bank 2
http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/5-8-08LTFT2.jpg

The cells that look off to me are large positive values, meaning that they are adding fuel. However if I look at a log of driving at those conditions I see that the AFR is still slightly rich, and the BEN is of average 98% at those conditions (14.68 is the command at those conditions that I inspected in the log).

Perviously (before I changed from OL to CL, I saw BENs of 0.995 in those cells in the map. If this is the case, then when I switch to CL, and the long term trim becomes approximately +10%, I would expect the mixture to be VERY rich, about a BEN of 0.90, but that is not the case. I see a BEN of 0.98 at thos conditions in CL. At this point I don't really understand what is going on.

[5.7ute]

Can you post a log with both 02 sensor voltage pids, wideband pid & dyncylair.dma as well as the usual rpm,ect,iat etc to look at. Make sure you are in open loop though as TAQuicknesses way should be quicker.

[SSpdDmon]

What is the grams/cyl range for those cells? You need to log actual AFR averages based on grams/cyl. Then, you can translate those areas that are performing incorrectly into what CL mode it's in. Then, you'll know where to adjust your switch point. Make sense?

[Kevin Doe]

I switched my tune back to OLSD and logged WB02, NB02 voltage, short term fuel trim, load (g/cyl), and many other standard parameters. Then I took the data and sorted it so I could put it in the LTFT matrix.

I am currently running 425 mV as the 02 switch point across the board.

You'll see 5 values in each cell in the matrix.
1. GM.DYNCYLAIR_DMA (g/cyl)
2. WB02 (AFR)
3. Short Term Fuel Trim (%) - NOT ACTIVE
4. Narrow Band 02 Voltage Average (V)
5. Long Term Fuel Trim (%) - NOT ACTIVE This was logged yesterday from when I was in CLSD.

I'm not sure why the AFR data isn't dead on. I did some filtering, but I guess not enough, and I wasn't as steady state as I was when I did the VE tuning. This was taken from a ~25 minute long work commute. I filted out low coolant temp data, and data with throttle movements greater than a threshold.

My real issue was with the parameters in the 1400-2200 rpm cells with 37-57 and 57-77 cells. The LTFT was very large in those cells. If I updated the switch points for the loads (CL mode) that correspond with those cells I'll be changing other cells as well, right? Also, table B4104 looks weird. The x-axis uses GM.DYNCYLAIR_DMA and calls the units in (grams/second). I was under the assumption that the units are in (g/cyl). The same parameter is used in the spark table and it calls the units out in (g/cyl). The range is 0-320 in the CL Mode table, which isn't anywhere near the actual DYNCYLAIR_DMA range ~ 0-1.20. What is going on there?

http://i93.photobucket.com/albums/l55/Kevin_Doe/RX-7/5-9-08LTFTMatrix.jpg

From this data, where do I go from here?

I have also attached my tune and log file that I created this data from.

[405HP_Z06]

Kevin,
What did you use to create the matrix above?

[Kevin Doe]

I created that in excel, however I just did it manually. It was actually pretty much a pain in the ass since the cells aren't the same size.

[SSpdDmon]

If the STFT's/LTFT's weren't active, why are there values there? When I turned mine off, they registered 0%.

[Kevin Doe]

If the STFT's/LTFT's weren't active, why are there values there? When I turned mine off, they registered 0%.

I honestly have no idea. Looking back at the data it looks like I was a little confused between SAE.SHRTFT1 and SAE.SHRTFT. They both say they are short term fuel trim for bank 1. The numbers are different, but only by 0.002 or something. Basically the same number. The numbers in the chart were from from the 02 parameter, not the Fuel parameter. The difference is insignificant.

Either way, I looked again. My long term trims were zero during the log. The short term trims still had values. Also, the long term trims that I put in those boxes were from a pervious log where I was in CLSD.

Can you look at my tune file to make sure I"m properly in open loop? I figured the short term trims were calculated but not used.

[SSpdDmon]

Closed loop temp enable should be maxed out. It's under one of the fueling folders. If that was done, you shouldn't have had STFTs. Hmmm....

[Kevin Doe]

I have that table set to 122 deg C (251 deg F). I definately never touch that temp. I have no idea.

[stigmundfreud]

did anything ever come of this?

\did you check b4108 which will return stft in OL or b4206? I've just been asking a question on whether stock os will work happily in semi-open loop mode by only running stft in CL