dbrunner2

I'm having an interesting problem with my LT1 (1994, installed in an '89 Camaro chassis with a T5, everything is basically stock including tune and emissions, except that the smog pump is not hooked up at present and the VSS has never worked with the PCM, LT4 knock module installed, MSD opti, running on 89 octane). My long term fuel trims (BLMs) have gone quite low, but well balanced left to right (cell 16 = 108/109, cell 17 = 115/114, cell 18=109/109). The PCM never leaves the range of cells 16-18 for some reason (at least FreeScan thinks so.)

Now this would ordinarily make me think bad MAF/EVAP dumping in too much fuel/bad fuel pressure regulator/etc., but the PCM has also started detecting knock off and on, almost always between 1100-2000 rpm, and under mid to high manifold pressure (0.5 bar or more). (This range appears to coincide with the configured range for enabling EGR in the PCM's constants table...interesting...) This causes it to pull timing (if the engine is fully warmed up), with the effect that power delivery is poor when I get on it below 2000. When I climb above 2000 the car takes off and delivers good power.

Consequently, I'm wondering if the problem is that the PCM thinks it is running richer than it actually is, and is pulling fuel, making the engine more prone to knock. However, I can't figure out how this could be, since the O2 sensors are both going from rich to lean and back to rich around 1.6 times a second when the car is in closed loop. It seems like a working narrow band O2 sensor has to cross over at right about stoichiometric AFR no matter what, and a bad sensor would manifest itself as low cross counts or stuck at high/low voltage. Also, with the problem affecting both sides of the engine equally, both sensors would have to have gone bad in the same way.

Does anybody have ideas on what I might look for? I have the ability to reprogram the PCM (TunerCat) and I can use FreeScan to take data logs while driving. I'm open to any suggestions.

[EDIT] Forgot to mention, I'm using shorty headers, not any kind of stock manifold. [/EDIT]

MikeGyver

First check the obvious things. Look for any vacuum leaks or exhaust leaks (before or near the o2 sensors). Also check for misfires (plugs, wires, distributor, etc.), check/clean the 3 elements on the MAF sensor, with a Q-tip and rubbing alcohol. Check fuel rail pressure with the regulator vac hose removed and plugged; should be within about 4psi of 43.5.

Injuneer

Are you sure its in closed loop? It will stay in Cells 16-18 until it goes into closed loop. Do you possibly have a code for the O2 sensor heaters?

1.6 switches per second is too low. May be a sign of contaminated sensors. That might also explain faulty data from the O2 sensors.

dbrunner2

It doesn't set any codes; the only reason I say it's in closed loop is that I can see the O2 voltages and short term fuel corrections oscillating back and forth with a roughly constant period, at steady engine load conditions. The whole cycle takes about 0.6 second on the left side, slightly longer for the right (so the sensor will switch over, either from rich to lean or vice versa, around 3.2 times a second). What conditions are necessary for the computer to go to closed loop? I know the coolant temperature condition is being met, since this is happening with coolant at 88 degC.

The sensors have been in service for 4 years now, so I can't rule out the possibility that they're fouled; I may just slap new ones in and see what happens. I also have a new MAF coming today, so I'll see if that has any effect. I'll also see if I can find someone with a fuel pressure gauge to make sure my regulator is OK.

There aren't any apparent ignition problems (ignition wires are all intact, and I can't feel any cylinders actually misfiring). There also aren't any vacuum leaks that I can see, or significant exhaust leaks (there's a tiny exhaust leak at the flange past the right O2 sensor), but I would expect either of those to produce high BLMs, not low.

WS6T3RROR

Majority of the time, clean the maf and the trims will go back where they belong. Especially likely if any sort of k&n filter is in use. You would not believe what kind of havoc a dirty maf can cause. Freaking out the knock sensor etc.

dookie454

How does it start and Idle and generally run with very little load? A 2000 Chevy truck I looked at had some rough light load and off idle symptoms and slight surging, felt like it was running very lean down low.

What we did was shut the truck off, waited a while then held the gas pedal to the floor and cranked it.. what we heard was one cylinder firing over and over, the only way that could happen is if it had fuel, and with the pedal at WOT during crank shuts the injectors off, so we proved a leaking injector without doing any real work.

The leaking injector was causing that side to read rich, and the PCM was pulling fuel to compensate but was making the other 3 cylinders too lean.

Other than that, my CAM makes my 02's read to rich, and the PCM leans it out.. I cant think of anything else except I did have water leak in my 02 extensions and when I would wash it the 02's would go way lean and I couldnt even drive it uniless I was in PE mode or unplugged them.

dbrunner2

OK, so I tried a bunch of different things this weekend:
* Retarding timing ~2.5 deg across the board using the adjustment screw on my MSD opti. No apparent effect here, so I restored it to normal.
* Cleaned the MAF elements using alcohol and a Q-tip as MikeGyver and WS6T3RROR suggested. (I couldn't put in the new MAF since it appears to be a reprogrammed 5-wire MAF from the newer engines, which came with an adapter harness, and my current intake tract relies on my custom sensor ends, which keep the screen and the original throat size of the sensor but have slightly smaller OD and ID on the outlets. This will change soon; I don't know that this is the issue though, since putting the car into speed density by disconnecting the MAF doesn't seem to help anything as far as the knock retard. For info, the GM part number for the sensor kit I got is 19207203.)
* New O2 sensors, since Injuneer indicated the cross counts were too low. The old ones had accumulated some carbon. The BLMs improved after putting in the new sensors and cleaning the MAF, but not enormously (the lowest I saw was 111, instead of 108; this might be attributable to the nonstandard intake tract). However, the cross counts didn't change too much; my rough calculation gives around 3.5 crosses per second from the data log I took today.
* Checked fuel pressure, both at normal idle and with FPR vacuum hose disconnected; it read 35 psi at idle and 45 psi with the hose off, both of which seem right. The gauge didn't have a long enough hose to allow me to drive around with it, unfortunately, but I don't think the fuel pump is the issue since the car runs well above 2000 rpm, including at WOT.
* Replaced fuel filter, just since it's cheap and it's probably been in service too long.
* Filled up with 93 octane (the tank was probably 92 overall, since I didn't run it dry).

Unfortunately...none of these seemed to have any effect on the knock retard. I'm not even entirely sure if the knock is real or false; if there is knock going on, it's too subtle for me to pick it up audially, which makes me wonder if this is just header noise that the LT4 KM isn't screening out. I have the log file in CSV format from today's drive if anybody would care to look at it; it looks like I can't put attachments up on here, so I've posted it on my personal web page.

Dookie454, the car runs just fine starting, idling, and under light load, except for a little rough running (feels like one cylinder misfiring) idling after a cold start. I don't see split BLMs or short term fuel trims like I'd expect if I had a bad injector, but since there's no way to really prove that all the injectors are good short of getting them all flowed, I ordered a set of Ford Racing 24# injectors that should arrive tomorrow. Consensus on here seems to be that the SVO injectors are good quality, and these are supposed to be a flow matched set. At $215 versus at least $80 to get the injectors flowed and cleaned, plus downtime, it seems like a good idea to just change them. (If you're wondering why I got the 24# and not 30#, I don't have any power upgrades planned, and I'd like to keep idle fuel metering as good as I can.)

Another conundrum is the matter of the BLM cell numbers; I'm at a loss as to why they would never leave the range of 16-18. FreeScan's status bits display does show the PCM going into closed loop when it's supposed to, and there are no faults set. The fault display is working, since I could see and reset the fault thrown when one of my new O2s died prematurely and started reading lean all the time. I'm thinking the problem is the BIN I used to reflash my PCM; unfortunately my PCM started life with an automatic, and the best I could find for a 1994 manual transmission file is the one from http://www.bfranker.badz28.com/LT1Edit/LT1files.htm, which says that HPP+ was once uploaded to the computer and it was then returned to stock. Perhaps the return to stock was not a complete one? If anyone has a virgin 1994 manual BIN, or if you know whether the 1995 BIN file can be used despite the different cam, that would be a big help to me.

Thanks, everyone, for all your help so far!

Injuneer

If I read this correctly, you have a stock tune, but some sort of replacement 5-wire MAF sensor, with an adapter harness. Does that sensor have the same output calibration as the stocker? Although you claim to have a stock tune, later you talk about using a .bin from another vehicle, and seem a bit uncertain regarding its viability.

You typically can not hear knock, because the knock sensor picks it up before you can hear it, and the PCM pulls timing to eliminate the knock. The fact you can't hear it doesn't mean it isn't there.

The SVO injectors will flow 5% more than they are rated at, when used in the LT1's 43.5psi fuel system.

dbrunner2

Sorry for the confusion; until I did some more digging recently, I didn't remember about the source of the BIN I had used to program the PCM. I actually just found what claims to be an unmolested 1994 manual BIN on the same site--major duh moment, unless it was added since I checked a couple days ago! I just dumped that into my spare PCM (with appropriate changes like turning VATS and CARS off), along with the LT4 knock module, and I'll be testing that out either tonight or tomorrow.

Regarding the MAF, I'm still using the stock sensor, although not in a normal LT1 intake tract (it's using the "lid" style air intake that came with the car), since the new sensor will not fit my current installation. The new one (Delco p/n 213-4657, GM p/n 19207203) is sold (by GMPartsDirect and RockAuto, possibly others) as a stock replacement, but it is a decidedly different design as I mentioned; I need to confirm that it is programmed to behave like the stock sensor before I install it.

wrd1972

Can a dirty MAF in any way cause a split BLM?
Is there a left and a right bank thingy in the MAF taking a reading?

dbrunner2

To clarify: the replacement sensor I got has 3.5" OD ends like the stock unit, but my intake currently requires that the ends be reduced in size, which I did with the current stock sensor while keeping the bore at the sensing element cross section the same and maintaining the screen. I know that this will throw off the sensor's calibration to some degree, but the car was able to run, when I first assembled this setup 4 years ago, without any knock retard issues, and I haven't changed the intake since then.

As far as the issue of inaudible knock goes, is there a way of determining whether sensed knock is real knock, or just normal noise from the headers? The only thing that occurs to me at the moment is to do something like reprogram a PCM to deliver extra fuel and back off timing (and/or fill the tank with race gas), and go driving to see if knock retard still occurs, which seems like it might not necessarily eliminate all forms of real knock.

WS6T3RROR

No.

dbrunner2

Results from today:
I obtained a new, supposedly unmodified 1994 M6 tune from http://www.bfranker.badz28.com/LT1Edit/LT1files.htm, opened it in TunerCat, disabled VATS, traction control, and CARS, and turned off codes 24, 46, 72, and 97 (since I have no VATS, ASR, or skip shift solenoid, and my VSS circuit doesn't work with the PCM), and flashed the PCM with the resulting file. The only difference I noticed was that the PCM obeyed the fan turn-on temperatures set in the constants table, which it did not do with the previous BIN. So, the old BIN was indeed not a stock program. The car still doesn't leave BLM cells 16-18, however, even when all conditions for closed loop are satisfied (O2 sensors reading away from 0.45 V, coolant temperature reporting above configured minimum for closed loop, and 205-second timer elapsed) and the status bits indicate it is in closed loop. It even shows the ability to learn (the BLMs change over time). At the moment I have no idea what else to try to get it to spend time in cells 0-15.

As far as knock retard, the jury's still out; I'm waiting to let it run a few days. If I understand Injuneer's posts in other threads correctly, the PCM learns over time when knock is likely to occur, and thereafter applies knock retard at those points (perhaps this is the low octane knock retard function?). The new injectors are in, perhaps that'll help.

Hasz28

Spark plug is loose