A picture paints a thousand words

[Gelf VXR]

Getting some useful graphical data from excel pivit tables now.

After I lost my HD, I recovered 3 of my most recent logs 3MB in total and put them together. same day, same fuel, but altered the timing.
The graph shows average BEN for all logged values for 0.08 grams cyl over the full rpm range for the shown spark advance 40-43 dgrees advance.

43 degrees shows some big lean conditions over the stock 40 degree setting.

[Gelf VXR]

And here at 0.12 grams cylinder, 42.5 degrees is leanest, but getting leanest at 42 degrees at higher rpm

[Gelf VXR]

I'll let the graph speak for its self, I have penciled in where I think the data is missing, but you can deduce from the graph the peak efficiency over the rpm range of the differing spark values.

I had posted a different graph earlier, but I noticed I had entered some data into the sheet using copy and fill:shock:

[Gelf VXR]

So from the graph, I would say that for 0.68 grams per cylinder the optimum spark values are as follows


3000 rpm = 26 stock 25
3200 rpm = 26.5 stock 26
3400 rpm = 26 stock 25
3600 rpm = 25 stock 25
3800 rpm = 25 stock 25
4000 rpm = 23.5 stock 25
4200 rpm = 23.5 stock 24.5
4400 rpm = ??? stock 24
4800 rpm = 25.5 stock 24
5200 rpm = 25.5 stock 24.5
5600 rpm = 28 stock 26
6000 rpm = 28 stock 25
6400 rpm = 25 stock 23
6800 rpm = 25 stock 23

??

[Gelf VXR]

Running a much leaner commanded AFR than stock would explain less spark required at 4000-4400rpm, but also shows GM running spark a little retarded from optimum in areas where AFR is stoich or above peak torque

Any views?

[Gelf VXR]

Attached my stock spark map and new spark map from BENs, with some smoothing

[WHYTRYZ06]

interesting data

[Gelf VXR]

Finally decided my hi octane map based on my last 3 logs using BENs to dtermine optimum spark. Back home at the end of August and will be doing lots of data logging to record the effects, if my interpretation is correct VE table should show noticable improvements acroos the map, other wise back to the drawing board:wallbash:

Below difference table against stock hi octane map. Only mods are catback and 5" intake to cone filter

http://i5.photobucket.com/albums/y164/andysa1966/differencemap.jpg

[ssvolvo]

Gelph you ARE the man.

For your next (?) tutorial you can show us laymen how to achieve optimal part throttle timing without the use of a chassis dyno. Now that would be "fine" tuning.

Please, please, please?


Besides I'm all out of rear tires....again!:hihi:


Thanks, John

[Gelf VXR]

By all means, if succesful an update to the AutoVE tutorial to follow, as AutoVE and timing is required to be done simultaneously:cheers:

[Gelf VXR]

Forum admin advised that the torque pid I was interested may just have been a look up table based on rpm, commanded afr and spark, so I proceeded with caution.


I believe this may be correct for this pid, as the name suggests

Calculated Torque TORQUECALC_DMA Nm,Ft-Lbs E40 ECM 2

GM.TORQUECALC_DMA


The pid I am interested in and have been logging I believe is measured by the difference in rpm from one frame to the next, and as the name suggests is delivered torque



Delivered Torque TRQTRANS Nm,Ft-Lbs Performance 2

GM.TRQTRANS

I hope to prove this one way or another. If torque is calculated based on rpm, commanded AFR and spark, the resulting torque graph I would suggest should be very smooth, with no peaks and troughs. As well as parrallel to adjacent torque curves of adjacent spark values.

On the other hand if it is a I suspect, differences in the BEN value will impact on the torque curve, as the PCM has no way of knowing this difference (we use WB O2's and EfiLive for that), peaks and troughs must be as a result of the real engine condition, I think the graph below goes some way to support this theory? It is possible to corralate some of the peaks and troughs in the BEN curves with the torque curves

http://i5.photobucket.com/albums/y164/andysa1966/bentorque.jpg


Torque may not be comparable with an engine/RR dyno numbers, but it doesn't have to be, it is required to compare before and after alterations to spark and commanded AFR.

[Gelf VXR]

Gelph you ARE the man.

For your next (?) tutorial you can show us laymen how to achieve optimal part throttle timing without the use of a chassis dyno. Now that would be "fine" tuning.

Please, please, please?


Besides I'm all out of rear tires....again!:hihi:


Thanks, John

Draft Version of New Auto VE & Spark Tutorial available to download here

http://forum.efilive.com/showthread.php?p=57353#post57353

[ssvolvo]

VERY, VERY, nice work. This should be the stickiest of stickies for the E40 crowd.


:master: You ARE the man!






Thank-you, John

[Gelf VXR]

For my next project and tutorial:grin:

http://i5.photobucket.com/albums/y164/andysa1966/12228-1.jpg

[Gelf VXR]

Numpty that i am i managed to incorrectly calibrate the VE table at the beggining of my latest spark logging sessions. Not wanting to bin many hours of logged data by changing the VE table, i think i have a solution.

My problem is when i load in what i have determined to be the optimum spark values, as soon as I switch the engine off and try and start again (at full running temp) the engine fails to idle and goes into limp home mode. Iload the original tune file and it runs again.

As explained below, i find that my BENs are really lean at low apcyl / rpm with the new spark. Which is all good news because that means Im achieving what i desired by increasing the g cyl.

Note: Advise to use serial logging for AutoVE, serial logging allows you to monitor the status of the windbands and for the LC-1 there are four error states of interest that can be filtered from the log when doing AutoVE. They are 0, 2, 5 and 7 which I filter out completely, not at all possible in anologue logging! Excellent new feature.


My solution i think is that i can modify the open loop spark fueling for the under fueling in the VE table to protect the engine (and make it idle). Using the BEN values, i can multiply these with the open loop load table to raise commanded fuel in areas where its lacking (idle) until logging is complete.


Any thoughts?

[joecar]

GELF,

Please post your filters.txt file (found at My Documents\EFILive\V7.5\User Configuration), thanks.

Cheers
Joe
:gossip:

[Gelf VXR]

Another useful tip is exclude Gear = 6, which for me is neutral, this filters out all the big spikes in BEN's i have been seeing and could not explain. When you drop in or stay to long in N the fualing shuts off!!! With DFCO switched off.

My filters

[Auto VE & Spark Filters BB logging serial]
Filter=Exclusive
X0000=SAE.SPARKADV,Deg,Degrees,EQ,60.00,100,OR
X0001=SAE.ECT,C,°C,LT,85.00,100,OR
X0002=SAE.ECT,C,°C,GT,95.00,100,OR
X0003=EXT.WO2ST1,,,EQ,0.00,100,OR
X0004=EXT.WO2ST1,,,EQ,2.00,100,OR
X0005=EXT.WO2ST1,,,EQ,5.00,100,OR
X0006=EXT.WO2ST1,,,EQ,7.00,100,OR
X0007=EXT.WO2ST2,,,EQ,0.00,100,OR
X0008=EXT.WO2ST2,,,EQ,2.00,100,OR
X0009=EXT.WO2ST2,,,EQ,5.00,100,OR
X0010=EXT.WO2ST2,,,EQ,7.00,100,OR
X0011=GM.GEAR,,,EQ,6.00,100,OR
X0012=CALC.BEN1_B,factor,factor,GT,1.21,100,OR
X0013=CALC.BEN2_B,factor,factor,GT,1.21,100,

[Gelf VXR]

PS. don't worry about my spark filter, i log BEN's for each spark by excluding all other values

[ChipsByAl]

Thanks Gelf, I like the serial filters
Al

[joecar]

Thanks GELF.