95 LT-1 Idle Cell Comparison - Humidity?

Hot target idle speed set to 800rpm with 26 degrees. Idle over and underspeed tables restored to original. Also added a bit more fuel to 2, 4 and 6 trims. Bin included with log this time.

Cold start this morning was great. No problems whatsoever - sitting idle was great but I might have felt a slight amount of surge at ~20mph before pulling in the parking lot at work. This is the first 50+ mile log I can recall with zero knock events.

The idle BLM split didn't improve much but the cell 2 split is slightly better. Still significantly fewer cross counts on the right. Planning on pulling the throttle body tonight to see if anything might be clogging up the 2 and 4 idle ports.

If nothing breaks and warm / hot restarts are ok I'm going to drive this a while and see how it performs. If I change anything it will be continuing to reduce the injector constant gradually as PE seems lean.

Hoping to work on SD VE tables this weekend.

Great you are having serious progress.
The afr target and cyl cut use the same byte in mode4 message and are exclusive.
v4 has the same limitation. In v4 you can change the cyl trims on the fly.

The cyl balance test wasn`t very reliable. The fans skew the map readings and afr is moving a bit and is not too fat. The wideband is almost topped at 16.5.
A new test with constant variables will give better picture, since you also change the rpm target.

The cut test produce some good results showing some lean cylinder on the right side. Could be #6 since it is the one that fires after #3. More air left for #6 to draw. Or it could be #2 the companion firing cylinder of #3.Generally all of the right side calibrations have more fuel added. I checked the bin and the trims looks really conservative. It is not uncommon to have 5-10% difference even with stock form.
I ended with this settings in firing order.

1.070
0.984
1.031
0.938
1.039
1.008
0.977
1.047

The result of drilling the passage holes even. #1 had much smaller hole since it draws much more air and now I got it at 1.07.
I tried to get them better with wideband test but failed. The test looked better and even but the feel and sound of the engine sucked so I revert to the best feel.
I always have 1-3 problems cylinders that didn`t react to changes as expected. I am planning to play more with them if I have the time. I suspect you need much more air on the right side for some reason. It could be 1 cylinder leaning or it could be mutiple.

As I remember the passage goes to mid of the intake plug and than goes to both sides center and than goes to intake ports. 2 90* turns till it reach the ports.
#2 and #4 have the shortest path. #6 and #8 the longest. #7 fires after #5 and for #7 there is almost no air left. Also there are 2 different designed intakes. the one have a center plug the other doesn`t and there are some other visible changes in the bottom shape. Do you remember which one you have.


I noticed that the rpm and map blm boundaries are also higher in the bin. The crosscounts could be lower because of the different settings in the vette bin. I will try the settings and compare the results with some older logs.

I couldn`t resist and made a quick log with the vette CL settings. To my surprise it worked perfect with SLP midlenght headers and GM 846 cam. The right one is a little shorter and the 02 bung is the header flange, the left is longer and the 02s is down the pipe, like 8-10 inches down the flange. SO there is considerable difference in the 02s placement.

So GM had perfect CL idle settings for radical cams and headers all the time and nobody noticed. Now I will have to ditch my Open loop idle patch, but first I will do some driving and see how it goes on the road.

Later I see what settings are changed and make analysis how it works in the bin.

Quote:

---

Originally Posted by kur4o

The afr target and cyl cut use the same byte in mode4 message and are exclusive.
v4 has the same limitation.

---

So changing the Target AFR table is the only way to lock AFR for this test?

Quote:

---

Originally Posted by kur4o

In v4 you can change the cyl trims on the fly.

---

I may break out a mouse and make a serious effort to utilize it this weekend.

I have a few interesting items to report from last night's changes. First I tamed down some of my individual trims on the right bank because I was smelling fuel at closed loop idle. I also took a good bit more out of the injector constant which is now at 41.88.

The first thing that I noticed is that startup went to hell when it was starting almost perfectly yesterday afternoon prior to the individual trims change. So I think I was on the right path and removing fuel from the right was a mistake. It seems to me (and this is just a theory) that the prime pulses have to be spot-on to get the usual immediate fire and run we all know so well with this setup. I had just never considered that the individual trims were applied to the prime pulses.

Also, this log of my morning commute compared to the other three days has the lowest difference between cross counts bank to bank - 7218L, 6010R of 31811 records. I don't know if this is coincidence or a direct result of one of the things I changed - it is pretty humid and there are spot showers popping up today. I guess I need a weather station in my car.

Lastly, AFR at PE is closer to commanded than ever, so I think I'm pretty close to the right injector slope.

Quote:

---

Originally Posted by kur4o

The cyl balance ... The fans skew the map readings

---

I've been forcing the fans on low because I don't want them kicking on during the test and changing accessory load. I assume this means your method is to force them off?

Quote:

---

Originally Posted by kur4o

The cut test produce some good results showing some lean cylinder on the right side. Could be #6 since it is the one that fires after #3. More air left for #6 to draw. Or it could be #2 the companion firing cylinder of #3.

---

I'm not sure it matters but my assumption is that #4 is the one robbing fuel from #3 because it's adjacent in firing order and closer to #3 injector than #6. It will be on the intake stroke while the #3 injector is firing, whereas when #6 is on intake #3 would have just finished depleting any fuel left in it's intake port. I think #6 is what's robbing fuel from #5 because it also has the same relationship as 4 to 3.

Quote:

---

Originally Posted by kur4o

Generally all of the right side calibrations have more fuel added. I checked the bin and the trims looks really conservative.

---

I arrived at this after setting them all to 1.0. I tried using the stock trims at one point and it wasn't good, but I could have had other things way out of whack at that point so I'll revisit that this weekend. I feel like I'm going to spend 90% of my tuning time nailing down the cylinder trims.

Quote:

---

Originally Posted by kur4o

there are 2 different designed intakes ... Do you remember which one you have.

---

I think I have a picture with the oil shield off somewhere. Will update if I can find it.

Quote:

---

Originally Posted by kur4o

I noticed that the rpm and map blm boundaries are also higher in the bin.

---

I did customize the BLM boundaries when I first got it running as a 350 with this cam. I may revert this back and see if the BLMs are better distributed. It seems like 95% of normal driving falls into two or three cells. But I suspect this is just the nature of running a built engine on the street.

Quote:

---

Originally Posted by kur4o

So GM had perfect CL idle settings for radical cams and headers all the time and nobody noticed.

---

Knowing that the CL logic is controlled by a PID type feedback loop, I'd always wondered why no-one having these problems was discussing changing the gain and proportional tables.

Reverting to trims from previous tune and then adding a slight bit to 1, 2 and 4 caused noticeable improvement to starting and idle sound. Also it seems to have much more torque when I let the clutch out too fast and bog it down towards 400 RPM.

Edit: Also, surge feels almost non-existent. When I pulled in the parking lot I left it in 2nd and let it slow down until it started bucking. I didn't log it but it was perfectly smooth down to what seemed to be idle target (800).

Sorry to quote myself but I had a revelation while out getting some air.

Quote:

---

Originally Posted by spfautsch

... my assumption is that #4 is the one robbing fuel from #3 because it's adjacent in firing order ... I think #6 is what's robbing fuel from #5 because it also has the same relationship as 4 to 3.

---

It may seem counter intuitive, but what if the way to get more fuel to #4 and #6 is to also increase fuel to #3 and #5?

Here's a back of the napkin sketch of how I envision parasitic fuel flow in a gen 2 LT-1.

Code:

---

1 -> 2

3 -> 4

5 -> 6
  ^
7 ⅃  8

---

Obviously all these dynamics change when the throttle blades open, but maybe the way to get perfect balance has less to do with the size of the idle feed holes and more to do with parasitic flow.

Quote:

---

Originally Posted by spfautsch

I did customize the BLM boundaries when I first got it running as a 350 with this cam. I may revert this back and see if the BLMs are better distributed. It seems like 95% of normal driving falls into two or three cells. But I suspect this is just the nature of running a built engine on the street.

---

You should [continue to] customise the BLM RpM and MAP boundaries, based on how you'd drive if you were trying to achieve better MpGs.
The OEM boundaries are silly even for a bonestock car doing the EPA's CAFE MpG test, nevermind driven normally, nevermind modded, etc

Thanks, I'll give it some thought once my SD VE table is dialed in.

Funny you should mention MPGs and CAFE what-n-such. I filled up tonight to find I've averaged better than 27 mpg over the last four days. I've nearly worn a hole in my left forearm pinching myself. Sure enough, I'm not dreaming!

One mechanical issue that is nearly universally overlooked, is the idlespider.

The throttlebody butterflies feed the two larger main holes on top which empty into their respective cylinder runners.
The Idle Air Controller feeds the smaller hole underneath the two larger main holes, and follow paths parallel to the main runners, emptying into each respective cylinder's runner near the very end of their respective paths.

So far, I have not seen a single LT1 inlet manifold that has had its idlespider properly cleaned during a socalled cleaning. This will definitely have an impact on your idle, as it does for us all.

Hope this helps.

Clogged ports certainly crossed my mind. I did pull the throttlebody with the intention of getting an inspection camera in there, but forgot that the idle passage makes an immediate 90 degree turn so that was a bust. The best I could do was verify the holes weren't plugged by viewing them from the primary ports. I wonder if a bath in sea foam might be possible without yanking the intake.

With that said, I'm not seeing a massive BLM split at idle, and it switches sides as the throttle blades open so I'm somewhat of the mind this might be a function of airflow dynamics caused by the cam and headers.

Untill the engine run flag is set there are no cyl trims adjustments.
The run flag settings comes from 2 tables 22b4 and 22bd.

Code:

---

#1 inj - #7,#2 cyl
#8 inj - #2,#1 cyl
#4 inj - #1,#8 cyl
#3 inj - #8,#4 cyl
#6 inj - #4,#3 cyl
#5 inj - #3,#6 cyl
#7 inj - #6,#5 cyl
#2 inj - #5,#7 cyl

---

The fuel robbing theory is really interesting and explain some of the issues I have had before. Running great when cold and turns to crap slowly with heat build up.
Since the end of injection target is set to inject fuel straight into the cylinder at colder temp, there isn`t anytime left for fuel to move around ports. When the engine gets warmer the target slowly move to 180* for y and f-body and there is 180* duration that fuel is sprayed and stays in the port before the intake valve opens and it get sucked in the cylinder.

Another addon to the theory that is not mentioned is the valve overlap, under some circumstances it provides natural egr effect and can reduce the oxygen concentration in the mix.

So there is 180 degrees before the fuel get sucked and can move to other cylinders. I made a quick chart at possible fuel robbing combos the first is the injector firing and than cyl#s intake open at 0* and *90 after end of injection occured for the given injector #.

If #3 don`t fire #4 and #8 can get really lean.

I will have to do another chart for least resistance path for fuel robbing and link it to the previous chart to find the most possible combos. The diagonal path seems with the least resistance.
6,8 gets fuel from 1,3
5,7 gets fuel from 2,4
And between adjacent ports could be the other least resistance path
1<->3
5<->7
2<->4
6<->8

I'm shocked that the individual trims aren't applied during cranking because I've had it starting great and then reset the trims to 1.0 and it wants to crank forever.

I tried your eehack fork on the trims and the engine died when I clicked 'enable' in the trims group. If I clicked enable before starting it would die when I changed a trim cell. It also cut out for no apparent reason while driving it around the block to warm it up. The source I have is from a zip file named eehack_4.9_source06_nov_18.rar.zip that extracts a root folder named eehack4.7_darkmagic_06_nov_2018.

Whatever the problem is I'm pretty happy with how it idles with these trims. When it goes CL there's about a 10 point BLM split to the right, and then it moves to the left as the throttle opens more. It's running incredibly well when I get on the pedal and I'm still in utter disbelief that it made 29.4 mpg over all the driving I did between Tuesday and Friday.

Code:

---

closed / open trims:
1-82        86
8-7d        7f
4-84        81
3-80        80
6-83        81
5-80        80
7-7e        80
2-83        84

---

I tried adding a lot to the right side and it didn't seem to make any difference other than idle got choppy. I also tried the above trims but added 3 points to #3 and the split seemed to get bigger.

I looked for an exhaust leak, but the only place the header could be leaking is between the primaries in the collector and I'm not finding any soot there.

Here's a balance test at 12.8:1. I forgot to switch the fan off but IAC and spark were locked even though I noticed advance was dropping to 26 once in a while.

I'm working on VE now. The only major driveability problem left is hot restarts.

I'm probably going to take the intake off and check the idle passages out for blockage later this fall / winter. It seems like a diode is gone in the alternator too because it's having tough time keeping both fans running. It's always something...

There was some minor bug of v4 patch regarding cyl trim and maf pump shot controls. The tables didnt copy to ram properly. Actually they were copied and cleared afterwards.

Here is the corrected bin_file.cpp to compile the fork. Don`t forget to rename the file to bin_file.cpp

Now it is tested and there is no stalling of the engine. There is a workaround to manually populate the tables. When connected press the cyl trims number 1 to 8 with 1 second interval and it will copy the values to the ram table.


Another thing that I had missed in the vette calibration is the off idle cyl trims. All set to 1 in f-bodies. It looks like Gm put alot of effort to tune the vette properly and the strategy is the front cylinders get more air and the rear less.
No clue how to tune that but it will be miss or hit I guess.
Some cylinders runnig rich and other lean but the mixture equalize in the end. A real nightmare for tuners.

A play with the closed loop controls might pinpoint the blm split at idle. It is very likely to be in the tune or tiny exhaust leak. The smallest pinhole in the solder joints can lead to major problems.

Thanks for the update / patches. I'll try to test them out soon. I think I'm going to park it until the alternator comes in though. I have an electric water pump and if I lose another diode in this heat it could end badly.

Quote:

---

Originally Posted by kur4o

Another thing that I had missed in the vette calibration is the off idle cyl trims. All set to 1 in f-bodies. It looks like Gm put alot of effort to tune the vette properly and the strategy is the front cylinders get more air and the rear less.

---

Strange they'd only do that on the y body. That's where 90 percent of driving happens.

Also worth noting are the constants at 126da and 126db that apparently dictate the transition points between using the closed, open and no individual trims.

Quote:

---

Originally Posted by kur4o

Some cylinders runnig rich and other lean but the mixture equalize in the end. A real nightmare for tuners.

---

If a tunnel ram has any redeeming qualities it's airflow at WOT. The end.

I'm sure they learned a lot from the mini-ram and it's no accident just about every intake design afterwards used long runners with the inlet funnels as far apart as possible. I would imagine it also had something to do with the LS engines' different firing order. Airflow dynamics are everything...

Quote:

---

Originally Posted by kur4o

A play with the closed loop controls might pinpoint the blm split at idle. It is very likely to be in the tune or tiny exhaust leak. The smallest pinhole in the solder joints can lead to major problems.

---

I'm sort of thinking it's an actual lean condition on the right side because I'm still fighting with hot restart.

Do you suspect less fuel than needed. You mentioned that the injectors you have are reworked version of some stock gms with p/n on them. If you can read the part number it might be possible to rip some data for stock application. At least the low pulse slope. I bet it is more related to the coil than the actual hole of the injector.

I noticed that ls1 type injectors are very sensitive on lower voltage. At startup it is not uncommon for the voltage to drop below 8-10volts during cranking. So the injectors prime pulses will be at that lower voltage. It could be that the shorter open time at high coolant temp can lead to not opening the injector.

You can test cranking for half second than stop, wait 5 seconds, if it starts right on the second time it could be less fuel than needed on the first attempt.

Can you describe the exact condition when hot restart fails. I am only having problems on the first start if it had been sitting for more than 4-5 hours no matter cold or hot.