nevernamed

Hi everyone,

I recently got my IROC put back together, but haven't had much luck getting it to run yet. Engine details: 383 TPI, still running stock injectors from the 305 it came with, SLP intake runners, TPIS ZZ409 cam, iron eagle 200 cc heads, hooker 2460's and Borla cat back.

I had it running ok last summer with that configuration, made a couple of changes over the winter: high flow intake base, ported plenum.

So far I haven't really even gotten it to idle right, it stays running and sounds ok but smells really rich. The confusing thing to me is that the O2 sensor says it is lean and so it keeps dumping more fuel but never gets it to come off the lean condition. Here are some plots of the datalog from cold start, I reset the ECM prior to this run. The first plot y axis is labeled mV by mistake, it's just the bit status for the indicated values, I offset them from 1 a little for clarity. It goes into closed loop at about sample 1250.



After the O2 sensor comes ready, it reads lean and rich as the idle goes up and down, but then after it goes into closed loop it just stays on almost 0 volts. The block learn goes up to 148 and it pulls the integrator back to 128, but the O2 values never moves? The other interesting thing is when I held the throttle at 1500 RPM, initially it went rich and the block learn and integrator started to remove fuel and it evetually started switching the O2. When I went back to idle it went back to 0 volts at the old settings. But then at 2000 - 2400 RPM it never got it to start switching.

When it's at idle, it smells rich and the exhaust is a bit hazy. If the throttle is rapped, it throws a puff of smoke. I did try driving it for awhile to see if it would come around but it eventually threw a flag for lean too long and I don't like driving it when it seems so rich.



This is a new O2 sensor, single wire, in the standard location for these headers. I also verified that the compression is good.

Here is a shot of the plugs after the above run.

http://nevernamed.home.mchsi.com/plugs.jpg

Any ideas would be welcomed.

Thanks

rgarcia63

First install a universal heated O2, the headers can't keep it warm enough at idle. I'd go with 30lb injectors regardless of the actual hp you'll get, but I wouldn't make it a priority right now.
Approximate HP for flow rate @ 45% BSFC, and 80% duty cycle (8 injectors):
270hp/19lbs
338hp/24lbs
419hp/30lbs

Rich/lean is good, but cross counts are more important.
What is the injector pulse width, and IAC counts at idle?

Hey, can you email me a sound clip of your exhaust. I need to make up my mind on what to get very soon.

nevernamed

Thanks for the reply. I did think about the O2 sensor not getting hot enough, but it seemed to work ok last summer with the same header configuration. And if the RPM at 1500 is evidence that the sensor got hot enough there to work, then why didn't it also work when held at 2000 - 2400 RPM? I will try a 4 wire sensor in the near future just to be sure.

Here is the IAC and pulse width for the same run as above.



Here is a link to an exhaust .wav file, but it's not exactly what the car normally sounds like. If you are farmiliar with the Borla exhaust, it has a diverter that can send the exhaust through the muffler or straight out the other side of pipes with no muffler. The only .wav file I have for it right now is bypassing the muffler at idle. This is with 2460's and the Hooker y-pipe but modified to have a 3 inch outlet and 3 inch pipe all the way back with no cat (I have a '74 block in my car so I got rid of the cat). This ended up being too loud to avoid harrassment so I am currently running with the open side blocked off and all of the exhaust sent through the muffler. I can get a clip through the muffler side if you are interested in the next couple of days.

http://nevernamed.home.mchsi.com/exh.wav

nevernamed

Ok, here it is with a 4 wire O2 sensor. Acts pretty much the same. I don't understand how the integrator can be happy at 128 and the block learn set at 139, but the O2 sensor is reading basically 0 volts all the time at idle? Then at 2000 RPM the integrator goes to 100 and stays there, the block learn goes to 108 and stays there and the O2 basically stays at 900 mV the whole time? I've gotten something really wrong here but I don't have a clue what to do next. I'm afraid my car is garage bound until I can get some help with this

rgarcia63

Yea, that's what I thought too, once I had the 4-wire installed the difference was like night and day.
Before I bought a new one I got one used at a salvage yard.

dimented24x7

iTrader: (2)

It looks like there is something really wrong with the O2 or the signal from it. You can tell just from the log that the closed loop corrections are swinging so wildly that its dragging the idle speed along with it. After that it looks like the CL routine reaches the limits of its authority and stops providing corrections. Are all the grounds good? Using an ACDelco heated sensor, or just a generic one? Is the heater element working (IOW, does it get hot outside of the car at KOEO)?

RBob

iTrader: (1)

It may also be that the AIR valve(s) are not working properly. If they start dumping air into the exhaust manifolds when they are not supposed to the O2 will see it as a lean condition.

RBob.

nevernamed

The AIR system isn't connected to the headers. Actually the entire system is gone but not disabled in the computer. I was tracking when the air was diverted to see if it coincided with any events, but it doesn't seem to.

The main problem seems to be that at idle, it reads lean and the computer can't correct it (is 139 the limit for BLM? why is INT at 128 and still reading 0 volts?) and at RPM it reads rich but the computer can't correct it, and it always smells rich and the exhaust is hazy.

What the heck is going on with this?

pandin

This is how my 383 was with a stock chip, idle showed lean smelled rich, cruise was way rich.

The "cure" was to raise the idle and just off idle maf numbers (table 1).

If you are running the stock 19# injectors they will be limited (static) at peak torque, I have 24# (50# FP) and 30# injectors would be better.

nevernamed

Hey pandin, that makes me feel a little better. And I finally found a reference to the BLM/INT limits, so the response at RPM makes sense to me now. It smells rich, looks rich, O2 says rich, computer tries to adjust but can't and ends up at the limit. So that part is fine and I will hopefully be able to tune my way out of it.

The idle part still really confuses me though. How can the computer think it is fine, based on the INT getting adjusted to 128 and the BLM apparently not at the limit, but the O2 just stays at 0 volts? Not understanding that is making me a little worried to start changing things. And the fact that it still smells rich to me at that point. I have found some references to the O2 starting to read lean if it get's rich enough? Is there an explanation to how that can happen?

I suppose if I get the response at RPM fixed, the idle part can potentially be avoided by raising it as you suggested, but I would really like to understand what is going on there.

Street Lethal

iTrader: (16)

Did you check for any possible vacuum leaks at the bottom of the china wall when the new high flow intake was installed? I would inject some propane into the PCV port on the valve cover when at idle just to be sure....

pandin

I have also seen that O2 pattern cold O2, Highway mode, or anything that adds "extra" oxygen to the exhaust (vacuum or exhaust leaks).

Remember unburnt fuel and air, looks lean to the O2/ECM but is really rich.

dimented24x7

iTrader: (2)

If there is any issue that causes misfires or poor combustion, the O2 will falsely compensate.

The other thing that stands out from the datalogs is that the O2 momentarily goes to ZERO volts at times. Have you checked to be sure that the O2 singal is not inadvertantly getting grounded? If the O2 grounds, it will do exactly what your seeing. The BLMs soar and it goes WAY rich, and then the ECM stops and locks the INT when the ECM sees that the O2 volts are stuck. Im fairly certain that this is the issue. Check the O2 wiring carefully all the way back to the ECM. You can do this by unplugging the O2 and measuring with a DVM. The ECM should provide a steady .450 volts at the signal lead in the O2 connector with the ign. on.

pandin

dimented24x7 also wrote:
Its possible that cam overlap is the cause of the rich condition with the O2. The O2
rich/lean thresholds can be changed to values below stoich to help out the fueling.
You will also need to adjust the prop. gains. I remember having to do this due to a
false lean condition from the cam overlap, but its really more of a problem at low
intake vacuum when the exhaust scavanging can pull fresh air into the headers.
Although, you stated that it worked OK with the current cam and stock intake,
correct?

pandin

Too break the O2 loop temporarily set the max BLM to 130-140. This will stop the max rich and allow you to see what else is pushing the rich condition.

The intergrate can be at 128 and the blm at 150 or 100. That just means that the system is in a steady state but adding/subtracting a consistent amount of fuel.

Low or high intergrate numbers means the system is dynamic and looking to still adjust the blm. This is an unsteady state, any intergrate number 5 numbers above or below 128 will adjust the blm up or down 1 number. The up and down 5 numbers helps to smooth out (average) the process.

Also remember that different cells can be at different BLM's

nevernamed

I had checked with a propane cylinder for vacuum leaks but just to be sure today I tried with starter fluid. Sadly the drivers side has a leak based on the starter fluid method, not sure why propane didn't show anything. Now just have to hope that it's the runner and not the manifold. I also sprayed into the valve cover and didn't get any response there.

dimented24x7

iTrader: (2)

Sorry about that. Yeah, edited it out after a second look at the graphs. Thought it might muddle the situation at this point in time.

dimented24x7

iTrader: (2)

As I said in my earlier post, there is something that is causing the O2 volts to go all the way to ZERO. Its clear that when the O2 signal returns, that its correctly indicating that the engine is running rich, and immediatly begins reducing the BLMs and INT. But, it looks like the signal keeps cutting out, causing the O2 loop to dramatically add fuel and run the engine real rich. The problem is not with the O2, or a vacuum leak, but the O2 SIGNAL PATH! The lean code is more likely to be from the loss of signal from the O2. Find out why the O2 signal keeps going to zero volts. The most likely cause is a grounded signal wire. The ECM provides .450 volts on the O2 signal wire to allow a neutral voltage in the event of a loss of signal from the O2. When its grounded though, it pulls the O2 signal wire to zero volts. From the rest of the log, there are other issues at work as well, but you need to first restore the O2 signal before you can attack any of those other problems. Find out where the signal wire is grounding. I suspect that its likely a chafed wire near the header.

Street Lethal

iTrader: (16)

Wanted you to check the obvious first, being that everything took place after the new parts were installed. Had a similar issue last week, as I was waiting for a stop light here in New Jersey, and the car immediately, out of nowehere, starting running rough and stalling. Had to power brake it for a mile to get it home. No codes were thrown, but I immediately started thinking tune, as we all apparently do when the AFR is being effected by something. Then it dawned on me that I recently installed a new set of plug wires, and sure enough, when I traced them down, about four of them were burned right through from lying on the exhaust manifold. Changed the wires, and voila. This isn't to say that, like Dimented pointed out, the lead to the O2 sensor isn't damaged, it probably is. Just don't rule out the obvious, because it ran fine before the mods....

pandin

because it ran fine before......

My "dog" ran/was fine before.... now he is not the same. What could be wrong?

In the world we live in this statement is funny to me.

I worked in repairing things my whole life. Most things including all animals and machines can "work fine with 25% of their parts bad. You can still function with the flue, cold, and cancer. Anything can go wrong at any time, car wreck, heart attack, and it just breaks. So the trick is to find that one thing that changed to move you from "it was running fine" to "I have a problem".

When you "fix" something first you have to sort out what is likely bad and then make it better. Why this takes so much time and thinking is, A) there are so many total parts (replace them all $$$), and B) many different "bad" parts can cause the same "problem" (my brain hurts from thinking).

A lot of responses are "It runs better but that wasn't the "problem" you just changed a 25% are broken but runs part.

Something is "fixed" only when you think it is fixed. This is when you stop changing the, 25% are bad but it runs parts.

And back to about the "Dog", he was hit by a car on the road, I just don't understand the limp "He was fine yesterday"

nevernamed

Are you ok pandin?

rgarcia63

He was fine yesterday.

pandin

I had to leave for work so the proof reading was very lacking in the post of concern, my bad.

I'm fine, it is just that the statement, "it worked fine yesterday", has no being on weather your car will or will not work today. People seem to put a lot of faith into these statements, like it has any thing to do with what just changed in their car.

It may have been a "little" out in left field for an answer but I was trying to get people to think about all the many things that can change in the engine control system and not be important. Other things happen that are very obvious but go unnoticed. Only one thing that changes is the "real" problem but is hidden behind all the other not so broken things.

I have seen cars that the owner said "Isn't she a wonderful car and it runs great", but what I heard and saw was knocking rods and a little smoke. A few days later, the guy told me his "perfect" car had just died and "it was running fine yesterday!". I didn't have the heart to say a word. Just look at him and said "Wow".

Street Lethal

iTrader: (16)

I can appreciate the analogy with your dog, however, if you bothered to read what I said you would have realized that I said it worked right before the mods, not that "it worked fine yesterday". Those are two entirely different concepts....

nevernamed

Went through and redid the runners, no change in the condition and the same vaccuum leak plus a new one.

It was at one specific place on the intake that the starter fluid caused the RPM to jump, right at the 1&3 intake ports. But now it is there and also at the 6&8 intake ports (although it's a little hard to tell exactly where the fluid is spraying, but it wasn't anywhere on the passenger side before). Maybe it's the injectors? They are pretty old. Maybe it's time to upgrade the injectors, but then I'll have another variable that has changed.

Has anyone else had problems finding vacuum leaks with the propane method? Starter fluid works better I guess but the spray goes everywhere and it's hard to narrow things down.

Oh well, next weekend is planned I guess. It'll sure be nice to get this thing back on the road.

dimented24x7

iTrader: (2)

Have you checked the leads for the O2 yet? As Ive stated, this is not a vacuum leak or faulty injectors. Its a problem with the O2 wiring. From the datalog, the O2 is working fine, but there is a disconnect between the O2 and the ECM. Most likely a grounded signal wire. Ive illustrated what Im saying in the pic below. Take a look at the areas Ive highlighted. They clearly tell the whole story of what is going wrong. The true cause may not be what I stated, but the problem is simply a loss of signal from the O2.

pandin

From the left field, I agree with dimented24x7.

One of my rules of repair, is to fix the know problems and then reverify the symptoms.

It is possible to have a vacuum leak and a shorted O2 wire.

A vacuum leak with extra fuel (BLM 108) will cause a very high idle speed.

If I read your charts correctly, the RPM is up and down, not just up.

Six_Shooter

iTrader: (2)

I have had absolutly ZERO luck trying to use propane to find a vacuum leak.

Starter fluid works best for me.

nevernamed

I did take a look at the O2 signal path. When the O2 sensor is disconnected, the ECM floats around 0.45 volts, it doesn't show 0 volts. It acted the same way with a standard and 4 wire O2 sensor, so I don't think it is a problem with the sensor itself. The signal comes back as you indicated in the plot above because I held the RPM's up for a while to see what it would do. And when the RPM's are up it does read rich correctly and tries to correct for it.

I thought a vaccuum leak could cause the O2 sensor to read lean because of the unmetered air in the exhaust? This should be worse at idle because of the higher vaccuum? And I thought a vaccuum leak was a problem in and of itself, and should be fixed before trying to correct anything with the tune. My comment about the injectors was that they are quite old and the o rings are getting a little sketchy, perhaps that is where the vaccuum leak is coming from.

Not saying I know what's going on here at all, just trying to clarify what I was thinking. I figured I should fix the vaccuum leak before doing anything else. And it's not a bad leak, the idle is actually fine other than being really rich. The idle going up and down was while the coolant was warming up and the IAC was doing something (that I admit I don't totally understand either ) I didn't attribute that to a vaccuum leak. Maybe I am over reacting to it, but starter fluid should not be able to cause the idle to increase correct? This needs to be fixed?

pandin

Your logic is sound. Pick one " obvious" problem and "cure it" then recheck the symptom.

One way to sort out your rich at idle issue is to limit the max blm and intergrate to 130 - 140 and see how it runs.

If the is the O2 or any associated cause (shorted wire), the richness (smell and black plugs) will be less.

If it is a vacuum leak, it will be lean and an injector bad (leaking fuel), it will still be rich.

What you are doing is restricting the feedback correction, looking for a hidden cause.

For an example of an unexpected issue cause picture this, mouse got into car while in storage and chewed some of the wires causing an intermittent short when motor torque over.

It took 6 mouths and a few parts to sort out, I will explain the details to anyone that wants to learn the logic of this type of repair. A little tape was the "real cure" and lots of D-con when car is stored.

dimented24x7

iTrader: (2)

With a vacuum leak, one cylinder (or more) will be lean, with the others around it rich. At the O2, it will usually show up as a somewhat normal O2 reading with increased closed loop corrections. In the above log, the change in signal is extremly abrupt, which would be very unusual for a vacuum leak. The leak shows as a somewhat more gradual change towards lean when the vacuum in the intake manifold rises, since it takes a finite ammount of time for the air to enter the cylinder, combust, and travel out into the exhaust.

As for the IAC, there are probably other issues as well. It could be feedback between the O2 and idle control routines due to too much gain in the O2 corrections. This will cause the idle to surge, and the IAC to try and correct for it, which leads to a circular feedback that causes the idle to hunt. It could also be in the IAC calibration having incorrect gains, or faulty/dirty IAC, to name a few other things.

nevernamed

Finally narrowed down the vacuum leak. I think the injectors were part of the problem, those have now been replaced. I don't think the lower manifold was part of the problem, but it has been redone. The main problem ended up being the cold start injector, it was leaking badly around the o-ring. So now no idle increase with starter fluid anywhere, which is nice to have solved, still having other issues though and no time to work on it until next weekend again

nevernamed

After getting the vaccuum leaks under control, apparently the cam is still causing enough of a problem to make the O2 sensor to read false lean at idle. The only way I could solve it was by raising the idle RPM, even after adjusting the rich/lean threshold and slow o2. But now with idle at 1100 it gets good crosses and wide band verifies the AFR is good. The good news is the part throttle mixture is looking good pretty much everywhere, we'll see what wide open looks like after a getting a few more miles on the motor

dimented24x7

iTrader: (2)

Wow, so it was the overlap causing it? Its amazing how abruptly the O2 readings change from that from the logs. Never would have guessed it could be that bad. I guess it must be due to resonance allowing excess scavanging at lower idle speeds. I know my car used to do the same thing, but it wasn't anywhere near this severe with the shorty headers. Looks like a big cam, long tubes and O2s dont mix well...

I dont know much about the $8D, but if its possible, you may want to just do an open loop idle, or possibly adjust the O2 gains to block corrections below 1400 RPM. From the logs, below 1000 RPM the O2 signal drops off the side of a cliff, so you could run into surging even with teh RPMs set at 1100, especially if you have a stick.

pandin

Attached is my latest data log graph, notice the similarity to the lost O2 mv. Mine is because of Highway mode being activated. This is from a cold overnite start, idle to Closed loop, and then my drive to home.

This graph is plotted against time, total is 600 sec of run time.

dimented24x7

iTrader: (2)

Yeah, lots of air can make them read near zero volts. The thing about his, though, is how ABRUPTLY it changes. My car did the same thing at low speeds, but it was more gradual. As the idle dropped, or I let the clutch out without using the throttle, the aparent AFRs on the wideband would drop to around 18-20:1 from scavanging, and the O2 readings would exponentially decay off on the datalogs. The headers combined with dropping manifold vacuum must allow his engine to start scavanging excessivly around 1000 RPM. Learn something new every day.

Ronny

iTrader: (1)

Help me out here. Could OL idle be a good idea? Take the 02 feedback out of the issue at least at idle off idle. I believe mine has a MPH qualifier(set at OL <3 mph) .

pandin

On mine, I had the air injection hoses disconnected but still had the check valves installed on the manifolds. I had to cap the check valve inputs because the exhaust scavenging was so great, that you could hear the air sucking in. A really good exhaust system will pull the air/fuel right on through the engine. Add a big cam over lap and down the exhaust pipe goes, a bunch of unburned fuel/air, which will read lean on the O2 sensors, even if it is rich.

dimented24x7

iTrader: (2)

I used limited CL corrections at idle with good success due to the fact that there was some areas that the O2 would cause a false rich condition. You can also use the rich/lean thresholds as well to correct for the false rich condition, provided its not too severe.

dimented24x7

iTrader: (2)

Lol, I did the same thing early on with the valves. Found out the same thing that they allow lots of air into the headers. I've hooked up vacuum/manifold pressure gauges to my headers, and they will pull about an average of -1/2 PSI of pressure when the exhaust flow is low. With a wideband hooked up, as the RPMs fall and the manifold pressure rises, the wideband will gradually read leaner and leaner. Its obvious that its a false lean condition due to the fact that the mass airflow sensor is indicating a normal ammount of airflow into the motor, and the motor still ran well. Unlike my experience, though, it looks like the transition to where the most scavanging can take place is much more abrupt with long tubes and larger cams.

JP86SS

iTrader: (1)

I thought this was a MAF car so I didn't post before, The $8D code runs with the 100 Kpa spark setting from the main table until the temp comes up. Not the closed throttle table.
Check that the spark you want when cold is setup there.
Could be adding to your problem.

nevernamed

Initially I tried to keep it in open loop for idle, but couldn't figure out how to make that happen. I'm using the $32B program. I just checked the hack again and didn't see anything about a MPH limit, that would be a cool fix. I also tried to disable the BLM using the LV8 limit, but then I didn't want to keep it from block learn in other driving conditions where the LV8 is low. So I begrudgingly increased the idle, so far that has been working good, but it doesn't sound as mean.

pandin

To rebalance the BLM's from the idle to cruise, raise the idle MAF numbers (mine idles at 7-8 grm/sec) in table one. Do this at and above your idle Grm/sec. Mine was leaner at idle then at cruise so I added to maf table one (LC5B5 5BD Arap 6E)

volts change new old
1.46 0.00 22.28 22.28
1.28 0.00 17.70 17.70
1.10 0.00 13.93 13.93
0.91 0.54 11.41 10.87
0.73 0.63 08.98 08.36
0.55 0.90 07.10 06.20
0.37 0.63 05.12 04.49
0.18 0.45 03.68 03.23
0.00 0.00 08.36 08.36

Try this and see if you can lower your idle RPM again. Now all my BLM's are the same from idle, cell 0, to cell 15.

dimented24x7

iTrader: (2)

My bad. for some reason I thought you had switched over to the $8D.

If your using MAF, than open loop might actually not be that bad. The $6E is probably the best out of all the MAF codes. I had a modified version of the $6E code running with an LS1 MAF, and that ran quite well in full time open loop. If you use a MAF thats in good condition with properly calibrated tables (using a wideband), than there isnt as much need for closed loop. With the $6E code, the fueling is strictly based on the mass airflow rate. Because of this, the MAF actually provides ok accuracy over a wide range of temperatures and flows.

Brighton

I don't think anyone answered your question above so I will try. I think the fuel trim was disabled. When fuel trim is disabled, the integrator is reset and locked to 128 while the BLM is frozen at its' last value when fuel trim transitioned to the disabled state. Having O2 droppout to 0 for and extened time probably caused fuel trim to be disabled--it can't trim the fuel with the integrator/BLM if the O2 sensor is not responsive.

ddahlgren

Have you pulled the plugs lately to see if they are all the same? In a previous post you had them and they were wildly different from each other some very rich and fouled and a couple that were fouled and might have burned themselves cleaner. I know it is autoshop 101 but you must of had the distributor out to upgrade the manifold base. Could any of those parts been damaged? The cap or wires? I bought a new set of Delco wires and they were 16000 ohms and the car ran poorly, old used ones were 3000 and car ran fine. I was changing only due to age rather than a real problem while doing a tuneup. I would get it running and if seems to run ok above aidle go take it for a short ride and pull the plugs right after with minimal idleing before hand. See if they at least all the same or equally wrong, if not fix what is wrong with that cylinder or cylinders first. If it was just the O2 sensor the plugs right or wrong would all be the same.