For tuners, a Holy Grail of Injector voltage Offset

[XfireZ51]

Originally Posted by efnfast

So Dom, all of this gets addressed with a custom tune?

I can't speak for other tuners although having looked at calibrations, many don't approach these parameters. Some might do "workarounds". A good number of these parameters aren't defined in the .xdfs or .tdfs generally available.
Finally, it takes time and gas. I'm pretty finicky about how I want the car to react to inputs, so I probably spend more time than necessary. But its getting easier because as we develop "standard" configurations for modified LT-5s we can at least start with a calibration that's in the ball park then fine tune from there. Don't need to constantly re-invent the wheel. As an example, I was able to use the tune on my 5.7L and translate for use in a 6.2L because of similar porting, cams etc.
Hope that answers your question.

[efnfast]

Yeah, when I have Pete do the heads, hopefully she can go over to you for a tune.

[XfireZ51]

In trying to find the right Inj. Bias for the 23# Accels I have, I think I have settled on 20% higher than what the IB is for the stock Bosch injectors on the LT-5. My observations in coming up w a more accurate IB is that the BLM split L/R banks seems to be much smaller. It still appears to diverge at idle but the spread isn't what it was. However, above that point the BLMs track very closely with a +/- of 4 from 128. Actually the avg turns out to be 133.4/125.6 or an average BLM of 129.5. This average includes idle which is about a 136/120split, so you can see that cruise is much closer. Slightly lean but I think that makes for a mellower exhaust note. Also, getting to the right tune happened much quicker. I've gone through maybe 6 iterations and now at the point where we are trading the constant .39 back and forth.Once your there, no sense in doing the seesaw.
The motor runs smoothly and doesn't have the blub blub blub note from the exhaust at low rpm/higher MAP. Summarizing it looks like chasing a more accurate IB has the advantage of:

1. Tightening the spread between the L/R bank BLMs. For me, this would also be a big argument for wanting to run C/L. In C/L, the O2s are working to bring both banks into tighter alignment in terms of fuel mixture. Without the O2s, as in O/L, you'll be working w the optimization of whichever bank you've decided to put your WB into. And someone tell me if there's a MASK ID out there that has VE or LV tables for each bank? That would be a bear. Got my hands full w 4 VE tables right now. Granted there will still be some variation due to injectors and even the O2 sensor itself, but at least you have an ECM doing self-correction. Anyway, maybe a separate thread on this topic.

2. Makes getting to the right tune considerably quicker.

My only other question is whether to bother w a bit of smoothing or just leave well enough alone.

[XfireZ51]

While going thru this, I noticed on my datalogs that the O2 signal on the R/S would drop to 0v. Sometimes just a few frames, and other times I would get a whole string of 0s. Long enough and the ECM is going to react to what it thinks is a lean condition. The 0s appeared before I started playing around w the
Injector Bias, so I suspects something electrical. I checked the ECM connEctions and pins. Also tested the O2 signal and sensor ground for continuity. They checked out.
Ran a few more diagnostic tests after reinstalling O2 sensor ground line in ECM.

1. Unplugged A/C fuse to eliminate any possible "crosstalk" from the heater circuit. Both O2s remained at .45v after engine START.

2. Reinstalled fuse, disconnected O2 and jumpered the O2 signal to ground. Engine needs to be running otherwise it will always show the ECM bias value. This dropped the R/S O2 to .03v so that was good.

3. Tried both harness extensions I had. The one I made the other day appeared to be giving me a problem. So I made one more extension and installed it. Now its working w no signal loss altho the O2 will jump in and out of Closed Loop if CTS isn't above 77C. Fans kick on and drop coolant temps in this weather. Just need to get heat into motor. Hope to take it out tomorrow and do some logs
but looking at idle the O2 voltage seemed to be tracking closely along w INT. Right was needing to be somewhat richer than left as usual.

[XfireZ51]

Done 2 runs w the new extension harness, and motor seems to be running real well. Whatever I did worked, including eliminating the O2 sensor signal drop I had on the Right Bank. No loss of signal now. However, I am getting a C/L to O/L to C/L flip flop at times. The hotter the CTS the less it happens but looking at the logs, my best guess is that I am hitting the top or bottom of the C/L O2 window.
That's what seems to be going on. Should I leave them alone or increase the spread? Right now its stock at .698v/.198v. Its seems to be more frequent w cooler MAT/IAT temps.
Maybe run less Prop Gain?

[XfireZ51]

Finally! Accel has published their Injector offsets. However, the one for the 150121 is not there. The 24# injector may be close enough. Its a bit more bias than what I "stumbled" on with my testing. I'll try it out once I get the car back on the road next spring or on a nice sunny day.


http://www.accel-ignition.com/produc...injectors.html

[tpepmeie]

Originally Posted by XfireZ51

Finally! Accel has published their Injector offsets.

Dom,
looks like its missing the low pw compensation curve, no?

That one's quite important as well, especially if you have larger injectors because you will typically be seeing smaller pulsewidths especially at idle.

Not sure this has any connection to your split BLM's. At least I can't envision such a relationship. These offsets are there to ensure the fuel delivered (by both banks) is consistent with the commanded at all conditions.

Todd

[tpepmeie]

Originally Posted by XfireZ51

However, I am getting a C/L to O/L to C/L flip flop at times. The hotter the CTS the less it happens but looking at the logs, my best guess is that I am hitting the top or bottom of the C/L O2 window.
That's what seems to be going on. Should I leave them alone or increase the spread? Right now its stock at .698v/.198v. Its seems to be more frequent w cooler MAT/IAT temps.
Maybe run less Prop Gain?

I've seen the same thing. The issue, in my case, was that the CTS was warm enough to go Closed Loop (~47C), but it wouldn't enter learn mode until even hotter. Finally determined that there was still a bit of enrichment occurring up to 68C. The learn mode only enables if commanded lambda = 1 (14.7), ie no enrichment.

Not sure if that will help, but you might look at it, and zero the OL enrichment above 56C or so.

[XfireZ51]

Originally Posted by tpepmeie

I've seen the same thing. The issue, in my case, was that the CTS was warm enough to go Closed Loop (~47C), but it wouldn't enter learn mode until even hotter. Finally determined that there was still a bit of enrichment occurring up to 68C. The learn mode only enables if commanded lambda = 1 (14.7), ie no enrichment.

Not sure if that will help, but you might look at it, and zero the OL enrichment above 56C or so.

Todd,

Correct on all counts. No the ACCEL chart does not include PW correction. I suppose for larger motors w big injectors it may be a factor. Reviewing my logs,
I didn't find any PW that approached the numbers in that table. The smallest being at idle ~ about 4ms. So I didn't worry about that table. At first I thought perhaps that table was dealing w a delta in the PW but that was incorrect.
In playing w the IB/V table, I in fact found a marked improvement in BLM split.
Less so at idle conditions but even there the delta between the two banks was
decreased. However, at cruise it appeared to really make a difference. As I posted earlier in the thread, my overall BLM is averaging 129 for L/R, w a
133/125 split as of the last tune. Frankly, I'm pretty happy w that. In some places, as in higher RPM areas of the VE table, L/R are identical. To some extent, I consider that the larger cams also contribute to the L/R differential. A small variation in cam timing could contribute to the BLM split as well. I can only tell you what I observed when correcting the IB and can't unequivocally point to a causal relationship. However, when I "overcooked" the IB, the split deteriorated. So it seems as if there's a sweet spot.
Regarding the C/L flip flop, I ran into the issue you described about 2 years ago. Playing w the O/L v CTS table, it would go C/L but no Learn no matter what the temp. Well I determined that the Commanded O/L AFR had to be = to the O/L Offset in the constants. Richer than that and it would override Learn. In my case, that's an AFR of 13.67 at 80C. Took me a while to figure it out. That would suggest having the AFR =13.67 at 68C has Learn come on sooner. Would be great to have an ECM bench setup.
The issue I am having now is that C/L will check on/off at times. Looking at the logs, it's happening at low MAP ( so trailing throttle or decel) and I can see the O2v go lean. I have lowered the O2 window at low gm/sec. to address the rich idle issue you get w larger cams. This was actually causing the Code 44 I was getting at times while idling.
I finally figured out that I was forcing the O2 below the .2v threshold long enough to set the code. Raising the window eliminated the code. So now my suspicion is that the O2v voltage is dipping below the C/L window I just raised long enough to take it out of C/L. If I hit the accelerator, it goes back into C/L.
I am thinking that if I lower the Prop Gain, I can decrease the extent of the voltage swing and keep it within the C/L O2 window. Haven't had a chance to try it out, but I usually reduce prop gains, particularly at idle. Someone suggested narrowing the C/L window but wouldn't that make the problem worse?
Prior to this, I was noticing in the logs that my O2 signal was dropping off intermittently causing the ECM to think very lean(might explain some of the BLM issue). Sometimes it was only a few frames, sometimes a whole string. And it was only on the right side, where I have the O2 harness extension because of the headers. Made a new extension and validated the O2 signal and heater circuits. No more dropout.
Beyond the BLM split, having more accurate IB compensation significantly reduced the number of Learns needed to get the tune right. Within 5 or 6 cycles I got to where the VE changes calc'd were within .39% just about anywhere in the table. So no reason to keep going. Once I incorporate these latest IB values, should be interesting to see how much change there is in the VE. At least, it's moving in the right direction.

[Marc Haibeck]

I think that a large factor in bank to bank difference in fuel trim at idle is the use of headers. With headers only two cylinders per bank fire across the oxygen sensor. Maybe because the system has only half of the normal gas samples, the average is not as accurate. The OE manifolds have the primary tubes shaped so that the gas from each of the cylinders jets across the oxygen sensor. The deflectors can be seen if the heat shields are removed from 1990 - 1992 OE manifolds. I observe bank to bank difference of up to 15 fuel trim counts on about half of the engines that have headers and no other modifications.

I have observed that when warmed up from cold, as the coolant temperature goes past 170 degrees, about half of the totally stock engines will swing in and out of learn mode at low loads. If the engine is run at full load to 7000 rpm in second gear a couple of times the system will stay in learn mode. I have not observed any effect on drivability on an engine that is switching in and out of learn mode. It may be important during an emissions test.