I'm a little frustrated, after spending the better part of the afternoon trying to get a stable idle on my Camaro, but I think I need some guidance from those with more experience.

I've read and read and read, and I understand the vast majority of the theory behind this stuff, but implementing it in the real world is quite different (for me, anyway).

The calibration is ARAP, and the mask is $6E.

This engine is a freshly-rebuilt L98. I believe most of my trouble with tuning idle is coming from the ZZ4 cam. Specifically, I cannot get a stable idle -- it surges from 500 RPM to 1500 RPM, often stalling itself out. This occurs in open loop; I haven't been able to keep it running long enough to enter closed loop. I put a brand new MAF sensor on it today (unrelated issue) and I checked for vacuum leaks about a hundred times.

Based upon what I have read, I've tried the following things (always resetting the previous change, because I could never tell a difference):

-- Added timing at idle RPM/LV8 ranges
-- Reduced lower entries on "O2 Error Reduction Gain vs. Airflow" table
-- Reduced lower entries for MAF Table #1

I adjusted minimum air (pegged then disconnected IAC, etc.) before doing any tuning, but I had to buy a longer screw because the throttle blades needed opened farther than the stock screw would let me. With the IAC disconnected, the engine started to surge when the RPMs reached lower than 1000, but it would usually just fall to about 500 RPM and die. So, in essence, my minimum idle speed is somewhere around 1000-1250 RPM. A little ridiculous, huh?

Idle speed in the chip is set around 800-900 RPM.

I have logs from after each trial, which I can post if the need arises.

I just can't figure out why it's doing this... I've seen recommendations for adjusting things (and I've tried them), but hardly do I ever see why things should be adjusted. I guess if I knew where to look I could make something of this problem. Can someone point me in the right direction?

 

Big cams like at least 25* at idle. Look at your scan data to confirm you are getting the commanded spark advance. It really sounds like a vaccum leak or loose air tube after the maf....

 

If you disconnect the iac and the rpms still vary then it's most likely a vacuum leak..

 

I have a choppy old hydraulic cam thats around a .480/.480 w/ tons of duration in my 1967 327. I had a steady idle that does not wander at all w/ a stock aypy bin w/ my own timing table (idles around 24 degrees). I bumped up the MAF values in the first table to richen the idle up, making the O2 sensor happy, but this wasnt necessary just to achieve a stable idle. The car has the muscle car heavy lope but the 6E MAF keeps it steady and under control. It's definitely not the 6E that's your problem

 

I wouldn't call the ZZ4 that big of a cam. I, too, am running the ZZ4. Mine started up and ran (including idle) fine on the first fire-up attempt ('89 6E, etc), and I have aluminum heads, headers, etc.

It shouldn't be surging. I would have to agree with 11sOrBust on the leak. What are the IAC counts doing during the surge? Have you checked your MAF? A bad MAF can also cause a very poor idle.

 

to find out if timing is an issue, a quick/simple way to test is to unplug the ecm from the distributor (it's the single wire that you unplug to set the base timing), then grab a timing light and twist the distributor to play with the timing between 15* and 25* or so and see if a particular area helps smooth things out (finding what the engine wants). this can help point you whether you have a timing issue or a fuel issue. be sure to set it back to 6* when you are done testing and before hooking the computer back up!!

if it still surges, then you're having a fueling issue (assuming no vacuum leak).

you mentioned that you altered the maf table, i'm assuming leaning it out on the bottom end. try going the other way, richen it up. it could be lean surging?

I don't know much about 6e mask but I think I remember that you have to do more than just adjusting the values in the scalar tables to get them to take effect - can anyone correct/verify that?

beyond that... pull the plugs, see if it looks lean. how's your fuel pressure? is the fuel pump staying on and running smoothly? zz4 cam is not a big deal, shoudln't be having this much issue, something else is up.

you didnt advance or retard the cam, did you? did you check that it was installed correctly and not one tooth off etc.? if the cam timing events were off, it would certainly act weird (i'm sure someone with firsthand experience could jump in and share observations).

If you have good fuel pressure and richening it up doesn't help, then i'd start looking into making sure the cam was installed corectly and not severly advanced or retarded.

 

Attached is a graph of the IAC steps in relation to the RPM. The RPM and step count are proportional, so I guess that's a good thing.

I decided to try setting the minimum air again. Grounded A&B, waited 30 seconds, pulled the plug, etc... I slowly adjusted the throttle blades until I could get the engine to start, which occurred around 1500 RPM -- it was running very smoothly with no surging at all. Then, I slowly backed off the screw to get the engine speed as low as possible. I was literally able to get it down to about 500 RPM with NO surging or unusual things happening (it even had that throaty idle!). During this time, the engine ran long enough for the ECM to enter closed loop, and I saw the BLMs were pegged at 108 with the Integrator close behind at 110-114. Alright, so I'm rich at idle. Anyway...

I turned the ignition off, hooked up the IAC, restarted the car, and again things went haywire. I think the problem lies somewhere in the decay rate of the IAC (my other computer is off right now, so I can't check TunerPro), causing the RPMs to drop too much, too fast... which, in turn, causes the ECM to open the IAC to prevent the engine from stalling. The process repeats. Overcompensation? It's a brand new IAC, too.

Upon startup, I'd ideally like to see the engine hit ~1500 RPM and take its good old time getting down to about 800 RPM -- if it wants to take twenty seconds, that's fine by me. I just don't see any way to set this inital speed, nor a way to reduce the erraticity of the IAC control.

By the way, my logs show 15-18º timing advance between 550 and 700 RPM. It's labeled "Timing Advance" in the log (Moates' software), so I guess I should add my 6º base timing to this for a total timing number? If so, that's right on par.

EDIT: 91L98Z28, I didn't see your post until I already sent mine. That's a good idea about the "mechanical" way of finding a timing number that the engine likes. Fuel pressure's set at 53psi; just replaced the regulator yesterday (I bumped it up a bit). Cam was installed correctly, no advance or retard.

 

what injectors are you using? 53psi is a bit on the high side. my personal preference is to stay with stock pressures and tune in the ecm... dropping back down to the stock 42-45psi or so would probably help the richness (if you really are rich) for starters and get you closer than you are now.

You mentioned you have a new IAC. compare the pintle shape to the old/original IAC. GM has different pintle shapes, and this can affect the amount of air let in at various positions and thus how much of an airflow difference the IAC makes for each step change.

Personally, here's how I'd set it up.. at least to get to a workable starting point:

1) set commanded idle speed to 800rpm
2) completely close the IAC and unplug
3) set up the throttle blades so that the engine idles at 800rpm.

that way, the engine can't die because it will always have enough air to run. this "de-sensitizes" the IAC (it doesn't have as drastic a range of effect on the engine, it can completely close and there's still enough air to run). This is how I have my two speed density cars set up and I really prefer it.

beyond that, hmmm. there is three parts of your graph where the IAC is essentially not changing for several frames at a time, yet the engine RPM is still falling. If you're running really rich at idle, the engine could be loading up. then when the IAC pops open, not only does it raise the RPM, but it leans the mixture which gives you that 1500rpm surge.

I think I'd start with lowering the fuel pressure and see what happens.

 

the darkened areas is what i'm referring to.

no IAC movement yet RPM is falling.

you're either losing timing, or the engine is loading up on fuel, is my guess.

 

Yeah, not having a WB makes it a little hard to tell if the computer is telling the truth about the richness. I'm using stock 22-lb/hr injectors. I will try lowering the FP... my reason for raising it was a lean condition I noticed while running a $32b mask when I first got the engine running (I really didn't tune anything, though).

I basically tried what you suggested, about making the minimum air speed equal to the commanded idle speed (with a closed IAC), but I can't even get the engine to start until the throttle is opened enough to allow for about 1500 RPM -- any other setting it'll fire once and die. Could a rich condition at idle/startup prevent it from starting up at a lower RPM (i.e. "flooding" the engine to some degree)?

Here's the same graph as before, but this time with timing advance included. It does seem like I'm losing some timing during those inactive IAC moments.

 

a rich condition at startup could be causing the hard starting, it takes that much of throttle opening to lean the AFR out that the engine will fire. I am mostly familiar with the 8D, but I remember reading that when you switch from 32/32b to 6e, you have to disable the cold start injector, because 6e adds more cranking fuel on it's own without the assistance of the cold start injector. if your cold start injector was still hooked up, that could certainly be giving you extra fuel as well at startup. combine that with the high fuel pressure and you could be really running rich at startup.

Try unplugging the CSI (if it's still hooked up) and going back to 44psi and see what happens.

your new graph is interesting. There is still a few spots where the timing and IAC are relatively static, yet the engine is losing RPM.

What's your plugs look like? how's the exhaust smell?

finally, as for how i suggested setting up the IAC steps/minimum idle speed/throttle position. I suppose if things were really rich it woudln't work for you. It's how I set my two SD cars up though, based partly on input from this forum (to take the IAC out of the picture as much as possible) and also based on experience (things just run better for me that way!). I like to see about 5 steps on the IAC on a fully warmed up, no load idle (no load meaning, all electrical equipment off, no load on the alternator, fans off, trans in neutral if an automatic, etc.)

 

I'll take a look at the logs and see if I can't find some other variable that might be causing the loss of RPM.

I already unplugged my cold start injector, so that's probably not an issue here (although, I forgot to the first few times and that caused some interesting problems!).

The exhaust smells rich, no doubt about it, but at least it's not pouring black smoke . Seriously though, when I drove it the other day I used a quarter tank of gas in twenty miles.

The plugs are a little tough to pull without removing the headers, but I think there's one or two that I can access.

It's been raining off-and-on here, so I don't know if I'll be able to try anything new tonight. I'll have to see how the weather looks when I'm done with work.

 

for the section that you graphed above, was the o2 sensor active yet? it might be interesting to see that output, if it was warmed up and working and not just stuck around 500mv (too cold to work).

i'm interested to hear what the results of turning the fuel pressure back down will be.

 

Well, I didn't have a chance to do anything this evening... ran into a few issues with one of my clients and I had to deal with them. Hopefully tomorrow...

That graph is from data obtained while in open loop. I couldn't get the thing to run long enough to reach closed loop with the IAC enabled.

 

I was able to spend about an hour or so trying to nail down this idling issue. I reset the FP back to factory (about 42 psi verified by my pressure gauge), burned a virgin ARAP bin (with timing tables from AUJM), and ensured that the cold start injector was unplugged.

It idles MUCH, much better now. I can actually get it to start with the IAC disconnected and the minimum air set to about 800 RPM. The speed still tends to fluctuate by about a hundred or so revs, but that could be related to the lower-than-ideal timing I'm seeing at idle (about 18° total).

However, it was extremely lean the first time I tested it -- BLM and INT values were mostly pegged at 160 or close to it. I increased the injector constant up from 22-lb/hr to 25-lb/hr, which seems to have helped a little, but the leanness is still there. I don't know how far I should take this value if I want to adjust fuel this way.

Just for kicks, I held the throttle at about 2000 RPM for a few seconds -- immediately, I set off a Code 34 (MAF low). After turning the car off and on again, the code went away. This makes sense, because if the MAF is reading less air than is actually being passed across it, I won't be receiving the extra fuel to compensate. I'm thinking about resetting the injector constant back to what it was and tinkering with the MAF tables and scalars. I have a good idea of how they work in terms of the source code, so manipulation can't be that hard.

There's also some surging during coasting; I haven't looked at the logs closely, but the IAC isn't moving when this is happening. I need to examine them and figure out what's going on. Deceleration enleanment isn't enabled -- at least that's what TunerPro is telling me.

Also, I pulled one of the plugs before I turned down the fuel pressure. Naw, she ain't rich!

 

Your code 34 and surging could very well be related. Have you tried the MAF tap test? Tap it lighly with a screwdriver handle while the car is running and see if the car reacts. If it does, the MAF is bad.

 

Mangus, will the tapping test work with a chip-based MAF? I just replaced the old MAF with a new one (free lifetime replacement), probably the day before I started tuning the idle. The car threw a Code 34 with the old sensor too, but not conditonally like it does now (i.e. increased RPM illuminates SES light).

 

I'm not sure what you mean by chip-based MAF. If the MAF is new, it is probably fine. Its worth checking though.

It's quite unusual to get a code 33 or 34 under normal circumstances.

 

go the other direction with the injector constant if you want to richen it up a bit. the smaller the injector the more on time the ecu thinks it needs and vice versa.

 

Mangus, the "chip-based" MAF isn't really an exact description, but that's just the general term I've seen used on the boards. Instead of a thin wire like the OEM pieces, an aftermarket company makes MAF sensors that use another means of measuring air -- it almost looks like a thin piece of film that's likely conductive. There's a big microchip right next to it, so that's probably why many people refer to them as a "chip-based MAF." From my understanding, there is no burnoff required, but the electronics mimic a "fake" burnoff signal so that the ECM doesn't trip a code.

You know, I thought I was doing something wrong! Raising the constant to add fuel just didn't make any sense. I'll have to try that.

 

Alright, I went through some of my data and took a hard look at what might be causing the surging during coasting (which eventually leads to surging at idle when I come to a stop). I put together three graphs, trying to rule out certain things. I'm going to include my comments with each graph as a separate post. I included a portion of data at the beginning before coasting for transitional purposes. The dotted vertical line shows where coasting begins. The target RPM for this timeframe is 800 RPM. The engine is also running lean throughout all of this (BLMs around 140).

I made sure that I was looking at a segment of data where the throttle was closed. This was verified by ensuring that the TPS voltage was at its minimum setting.

The first thing I thought to check was air consumption of the engine. If the throttle blades are closed, the only place that air could be entering the engine is through the IAC (assuming there are no vac leaks, and I haven't found any the hundred times I've looked). A plot of RPM against MAF airflow and IAC steps can show whether or not the IAC is doing what it should. I used a mathematical modifier on the MAF airflow to better illustrate the curve (and to get it to fit on the plot).

It seems like something is causing the IAC to randomly shoot open. However, it corrects for this by properly reducing its step count. When the process repeats, the engine surges.

Now, what can cause the IAC to open like that? Below are a few more plots that might shed some light on this (hopefully).

 

I figured that just for the heck of it, I'd see what the timing was doing. Seems like everything is alright here.

 

This last one is sorta fun. Plotted against RPM is BLM, INT, and base pulsewidth. Again, I had to modify a data range (base pulsewidth) so that it would fit on the plot. I've also illustrated when the BLM cell changes.

I see three ways that this graph can be interpreted:[list=1][*]When the INT resets itself, there is a surge.[*]When the BLM cell changes, there is a surge.[*]When the IAC shoots open, the INT has to compensate for a lean condition, which increases its value and the BPW, meaning that the fuel system is working just fine and the problem still lies with the IAC.[/list=1]
Here are my arguments for each case:[list=1][*]This coincides directly with the graph. However, we mustn't forget what the IAC was doing in the first plot.[*]This also seems plausible, though the BLM cell didn't change during the first surge. Also, it might be possible that the influx of air coming from the IAC is changing the cell.[*]This logic seems to make the most sense. See below.[/list=1]
From all of this, this how I think the graphs can be explained:

Something triggers the IAC to open. Extra air is released into the engine. The RPMs increase enough to change the BLM cell. The INT resets itself at this point. Since more fuel is needed to mix with the extra air, base pulsewidth increases. The IAC realizes that the RPMs do not coincide with the target idle RPM, so it closes to reduce engine speed. Before you know it, the whole process happens again.

Hmm, why do I get the impression that I overanalyzed? If everything seems logical, then I'm truly curious -- what could be causing the IAC to open for (seemingly) no reason?

 

The IAC is shooting open because the RPM is falling. it's stall saver.

I chopped up your picture and resized it to make it more obvious. Note that at the first red dot, RPM is around 1000. At the next red dot, the RPM is down around 750. At this point, the IAC *has not opened yet at all*. The ECM is seeing a stall occuring and snapping open the IAC. In every case on your graph, a sudden rapid RPM drop is *preceding* the IAC snapping open.

So, the question is - what is making the engine suddenly "stop running"? I don't think a single cylinder miss would cause this (but maybe someone who has actually looked at a single cylinder miss on a data stream output could comment).

if you could get the car in closed loop and log the results, then we could see what the o2 sensor is doing (such as if a long lean period or long rich period precedes the stumble/saved stall)...

I'd also say that the MAF appears to be working correctly based on the graph results.

.

 

Yikes! The IAC opens in response to falling RPM. The RPM is falling in response to the injector pulsewidth getting cut in half. The injector pulsewidth is getting cut in half everytime the INT is reset to 128.

If I didn't know any better, I'd say it's some kind of bug in the stock code. But since so many people run this code, I'd think it would have shown up before...

Does anyone know what conditions cause the INT to reset to 128 for this mask?

 

So, do you think that the RPMs increase to such a high number because the IAC is trying to continually compensate for a falling RPM, and when the speed finally does start increasing again, there's enough "inertia" (figurative, not mechanical) behind this process that the IAC can't control it, no matter how much it closes?

I see what you're saying -- and the IAC saves the motor from stalling by adding air, which adds fuel (to indirectly compensate for the pulsewidth reduction).

I looked through the $6E hack and I found the following about INT resetting:

• There is a flag that forces the INT to reset every time the BLM cell changes. This flag is unchecked in my BIN.
• There is a flag that resets the INT when a canister purge takes place. This flag is checked in my BIN.
• There are settings for the minimum and maximum difference between BLM and INT values. This is set at 5, but that's a little odd since I've seen as much as 30 points between the two. Maybe this works over a specified time interval?

That's all that I was able to find. I remember reading somewhere that the INTs periodically reset as the BLMs change (since they really are the BLM "calibrator") -- I just don't know when that happens. It might have something to do with the BLM/INT difference thing. Or, maybe the "force INT to reset when BLM cell changes" flag might be inversed, similar to the way the "normally-open/normally-closed" fan circuitry changed in $6E -- so it's possible that having it unchecked would really enable it. I dunno, I'm just taking stabs in the dark.

Of course, there's always the possibility that the INT resetting is merely a response from another, completely different, action.

 

I thought of something else. 91L98Z28, you predicated the pulsewidth reduction on the fact that the INT is being reset -- were you implying that the PW randomly got smaller when the INT reset, or that the PW responded to the INT being reset? Here's what I mean:

As you know, the INT is a modifier to the pulsewidth, just like the BLM. It doesn't have as much of an effect on the final correction factor, but it still matters.

Assume that the INT is initially equal to the BLM. The INT keeps increasing, and we reach the final split-microsecond before it updates the BLM and resets itself. At this point, the current INT value and the "old" BLM value are the two things determining the pulsewidth. Then, the INT drops but doesn't add enough to the BLM. The fuel correction is lost because the INT, the main thing correcting fuel at that moment, was reset to its base value.

So, this is another prediction: Because the engine is running lean (BLMs around 140), the INT really has to keep up with increasing the pulsewidth. Perhaps, it's changing too much from its base value of 128, too quickly, and can't physically keep the BLMs in check. Overall, the mixture needs richened anyway, which would bring about a flatter INT curve (it won't have to reach such high values to compensate for the lack of fuel). Then, when the INT resets, the drop won't be as large and there won't be as much of an effect on the pulsewidth or RPM -- and maybe the BLMs would even update like they're supposed to, keeping everything stable.

If you think about it, this same theory could be reproduced on an engine that's running too rich -- in this case, when the INT reset, it would be increasing the pulsewidth and adding fuel to the engine.

What do you think?

 

I found a parallel that doesn't make much sense -- it appears that the INT is almost in direct relation with IAC steps. Why would the INT increase (lean) as the IAC closes (less air)?

 

stick it in open loop and see what happens. also have you checked your MAF to TB hose to see if its got a cack in it ? seen that on a few tpi cars. as soon as the engine shifts from surging itll will open then close the crack. considering the age of the rubber in the hose seems almost to likely.

 

funstick, when you talk about running it in open loop for testing, do you mean turn on the cold engine and see what happens (open loop), or temporarily disable CL in the calibration?

I made sure to check the MAF-to-TB hose when I replaced the MAF sensor. It's surprisingly intact and flexible, and doesn't appear to have any cracks. The opening/closing of a crack makes sense, though.

 

crank up your closed loop temp and see what happens. be cursious to see what going on. you said fuel trim was reading fuel lean. could be a screwed up o2 dumpng fuel on the plugs.

 

I had the broken hose thing happen to me . It was fixed after I replaced it. I run the zz4 camshaft and, the $6e . I had mine fire right up with the stock 1989 305 5 speed bin. According to the guage it sits at 700 rpm ....no surge . Hopefully you gt it figured out and , post it here .

 

see if you can include you maf voltage or grm/sec reading in your next datalog. i gotta curiosity.

 

Here are my notes from today's testing:

Changed: Reduced injector constant from 22.36- to 19.01- lb/hr to aid in leanness.
Observations: Still no surging at idle, but tends to jump around when brake pedal is released or accelerator is let off. Set Code 34 (MAF low). INT remains at 128 more frequently, but BLM is still high (150-160) for most cells, especially 0 and 1 at idle.

Changed: Unplugged and capped power brake booster vacuum hose to rule out leaky booster.
Observations: Seemingly no difference, aside from the fact that the brakes suck . Code 34 again.

Changed: Increased fuel pressure (I know, I know...) to ~50 psi due to indication of lean situation across almost all BLM cells. This was just a temporary test; I'm turning it back down after today.
Observations: BLMs are equal to 150 or 160 for nearly all driving. INT reads around 124-132 in cells 0 and 1, but very low (80-100) under load in cell 15 (how can BLMs be above 128 and INTs be below 128 for an extended period of time? Why doesn't it update?). I think this shows that I at least need fuel at idle. I also noticed a lack of power -- RPMs drop significantly, almost turning the engine off, when the accelerator is pressed 1/3 or more (flooded?). Codes 33 and 34.

I don't know if I'm burying myself deeper or not. I was reading about the Code 33, and I think it might just be tripping because of the cam -- some people have set the "MAF High Diag Error Threshold" constant a little higher than stock to alleviate this problem. I might try that someday.

Overall, I believe that we're still dealing with the same problem as before -- it's just that today's efforts were a little skewed due to my interest in correcting the fuel mixture... I guess I got a little too rich. I didn't have time to finish up because I had to be someplace.

funstick, it sounds like you have a hunch about the MAF tables. That's probably the sole thing that I should be adjusting right now, huh? I don't have the capability to log MAF voltage (yet), so I tried to equate it from the g/sec readings -- but that just seems too time-consuming, since the ECM interpolates between cells. I'd be here forever. I hope that just the g/sec values are enough. Here is a link to the entire log of the data in the graphs:

http://home.alltel.net/usaaf/etc/4-8-05_1811.xls

 

I'm really having a hard time buying into the fact that the cam itself is causing all these problems, it's not that radical of a cam by any measure. In fact, as others have posted, the cam should run fairly decently with no other changes.

There's a number of issues here (and I'm trying to go back to basics).

1) you raised the fuel pressure AND lowered the injector constant, but your BLM's had no response to these changes. Since you've had the fuel pressure high before and the plugs turned black - this is telling you, the motor is NOT needing more fuel. In other words (in my opinion): the BLM is *wrong*. It's reading artificially high when no lean condition actually exists (at least, not for the spark plugs you pulled!!)

-> pull ALL the plugs. are there any that are NOT rich? perhaps a cylinder that is not firing and is pumping lots of raw air past the O2 sensor causing it to read lean all the time? Be sure to clean the plugs so you can at least roughly read them again later (they are already wasted for fine tuning, but at least if you get the black crud off, you can know if it's gross rich if you pull them again in the future). Or just put in 8 fresh autolites.

-> swap the o2 sensor for a KNOWN good one (or a brand new GM one). if it's contaminated, defective (bosch?) or otherwise not reading right, then you'll drive yourself around in circles trying to get closed loop to work right. The spreadsheet you posted looks like a dead O2 sensor to me. You've got a fully up to temperature engine that's been running for 7+ minutes. If that's not enough to wake the O2 sensor up, it's dead (or it's reading a significant lean condition, which assuming all your plugs check out as rich, is NOT the case).

2) set the injector constant back to 22# and the AFPR back to stock. The motor really should not need drastic enrichment across the board. The plugs agree, so does your brake experiment (after lowering the injector constant, the brake booster had more effect on RPM when you used the brakes, because it was leaning out a rich mixture.

3) ohm out all the injectors. multecs are known to fail (the windings short out) which is what I assume you have (stock 22#). You should see 14-16 ohms on each injector. Anything less and it's bad, the windings are shorting. If one is messing up internally, you would be pumping raw air out that cylinder and messing up the o2 sensor reading (there should also be obvious evidence on the plug for the cylinder with a defective injector).

4) disable closed loop altogether (set closed loop enable to 254*).

4) reset the computer (30 seconds without power) to reset ALL the learned BLM data back to 128.

If all 8 plugs look the same, and the injectors check out OK, ....

5) take the car for a drive and see how it responds. get it up to normal temperature and log the o2 sensor data and the MAF sensor data and also the TPS sensor data...post that stuff. With closed loop disabled, this takes the o2 sensor entirely out of the picture as far as fuel mixture is concerned, and your BLM's will stay at 128 for all cells, and you'll get a stock unaltered fuel curve. Also, if we had a data log of when the actual error is set, that might be quite helpful.

As for the MAF errors, ...I'm glad I don't have one of those on my car. But that doesn't help you.

There is a code 33 troubleshooting chart here:

https://www.thirdgen.org/techbb2/sho...hreadid=245300

in short:

Code 33 indicates: ECM has seen flow in excess of 45 grams per second (above 2.2 volts) for one second when:

* Engine is first started
OR
*TPS < 50% AND RPM < 2000

ANd a code 44 chart here:

https://www.thirdgen.org/techbb2/sho...hreadid=281516

In short:

code 34 indicates: ECM has seen low air flow less than 2.5 gm/sec (low voltage) for one second when:
*Engine is first started
OR
*RPM > 600 AND TPS > 6%

Other thoughts:

* is the motor basically a bone stock L98 with a zz4 cam? For example, does it have headers on it, and the o2 sensor is not coming up to temperature properly (headers can do this especially on an older dying sensor)? Some people with headers have needed heated o2 sensors.

* how is the wiring harness on the car, and the connectors? are any loose/problematic/broken/ nasty/ corroded? Your spreadsheet looks like it has all good data in it, any chance you could post data when the CEL actually comes on? that way we could see everything that's happening when the code is set.

 

See previous post. basically, I don't think it's actually running lean at all, either there is raw air getting into the exhaust (missing/dead cylinder, exhaust leak) or the o2 sensor is dead.

1) Are you sure you are starting with virgin ARAP code?

2) you might also consider grabbing another ECM from the junkard to swap in. Do you trust the ECM you have?

pulsewidth is supposed to be the calculated result of BLM/INT/other values, not a driving factor, thus I don't see the pulsewidth causing the INT to change (especially with a dead o2 sensor which isn't giving any feedback anyways). The INT is a short term minor trim - in no way should it have such a significant effect on the pulsewidth. There has to be something occuring which is cutting the pulsewidth so drastically - the INT alone shoudln't do it, even if it does drop from 173 to 128 (frame 2703 to 2704 in your spreadsheet). Nothing else between those two rows indicates why the pulsewidth changed so drastically. LV8, MAF, EGR, evap purge, RPM, CTS, TPS, all the same. Are you sure all the decel fuel enlean/cutoff stuff is disabled?

 

final post for the night, I guess I'm pretty much in agreement with funstick (besides, he has far more experience on MAF than I do) on pretty much everything i said.

you mentioned that holding the car at 2000rpm sets code 34 (MAF low) immediately. can you replicate that, data log it, and post the results. code 34 is set when < 2.5 gm/sec is measured (essentially NO air flow)...but the engine is getting it's air somewhere?! I would triple check the MAF to intake bellows (take it off, flex and bend it at every ridge and look for cracks, leaks, airholes...heck even fill it with water to see if it leaks (and dry out before use of course). that air is coming in somewhere, either through the MAF and it's not reading it (or it's reading it but the signal isn't getting to the ECM), or somewhere else and it's not going by the MAF in the first place.

look forward to seeing more data logs, especially ones where the check engine light goes from being off to being on, so that we can see the conditions that generated it.

 

It does make sense... the plugs show I'm rich, the ECM says I'm lean, so either extra air is coming in from somewhere or the ECM or a sensor is fried. I don't think it's a sensor, though, because everything related to fuel is just about brand-new:

I installed a new AC Delco oxygen sensor about a week ago, because the old Bosch sensor I had was garbage -- that one really was defective. The MAF sensor is also brand new; I got a free replacement from Advance Auto Parts. The MAF is made by Micro-Tech, and I believe is exactly the same as the Wells MAFs. The fuel injectors were cleaned, tested, and calibrated by Cruzin Performance sometime in late '03. My car didn't run at all last year, so they might have 500 miles on them. Still, I will be sure to pull the plugs and check the impedance of the injectors.

The engine is a rebuilt L98 with a ZZ4 cam, '601 heads (~10.2:1 compression), and headers. I've also done some very mild TPI porting, but probably nothing that even matters.

The wiring harness is good. When I rebuilt the engine, I went through the entire harness loooking for frayed ends, broken wires, and anything else amiss. I found no problems. My friends thought I was crazy, but I was just being me.

I'm using a virgin ARAP bin with the following adjustments for my car (since I have an '86): VATS disabled, fan flag changed, using all AUJM timing tables (main, PE, etc.), and increased Idle RPM vs. Coolant table by about 200 points at the higher temperatures (nominal idle speed around 800 RPM). The cold start injector is also unplugged.

The ECM could be bad -- I've had premonitions about it before. Good luck finding a '165 in a junkyard around here, though.

The datalogging section of TunerPro RT tells me that DFCO is disabled, but I don't see any on/off settings in the BIN-editing section.

This is a log from yesterday (the last one I made, see my post above). I believe I found a source of the Code 33/34 errors. TunerPro RT and Craig Moates' software doesn't log when a code gets set, at least from what I've found. What program does?

http://home.alltel.net/usaaf/etc/4-11-05_1657.xls

Frame 777 (Code 34): The g/sec reading randomly drops to 0.51 for seemingly no reason. Lines before and after 777 look pretty much normal from that standpoint. This is confusing, since I don't even see how the ECM can interpolate to a value of less than 2.5 g/sec. Look at MAF Table #1, you'll see what I mean. The lowest value possible is 3.23 g/sec.



Frame 1441 (Code 33): This is during a period of acceleration. The RPMs drop significantly, the engine "catches" itself, and things return to semi-normal. This would trip a Code 33, since the RPM is low and airflow is still high (but TPS isn't really high).

Frame 2114/2115 (Code 33): RPMs drop more than 500 points and TPS stays the same. Again, it seems like it "catches" and recovers -- but check out the MAF flow in frame 2117. It goes from 43 to 11 in one frame (?), and TPS hardly decreases at all. I remember when this happened in the car, I was going up the hill to my house (hence, the high LV8 numbers) when I thought the engine was just going to stop running because the RPM dropped so fast. Once I reached the crest of the hill, it was fine.

Well, I just wanted to provide some background to the points you made, 91. I'm going to play around with this some more this afternoon and check for failed parts, so I should have some observations (and hopefully data) by tonight. Wish me luck.

 

Just to add ,

The intake I am useing is an edlebrock base ( smoothed out a bit) and runners ( siamesed in 4" ) the plenum was also siamesed . Finally I am useing the stock 083 heads. It ran on the stock 305 5 speed prom . Like said above idled at 700 ....no surge.


Just thought I should toss it in for ya .

 

sounds like a maf drop out. could be a bad ground. try regrounding the meter directly to the battery. thats why i wanted the MAF data. ground the MAF diectly to the battery.

 

What do you mean when you say TunerPro/Moates software will not log if a code gets set? The only way that would be the case is if something critical fails and the car goes into limp home mode, in which case you cannot connect with ALDL.

I've logged a number of sessions with codes set. =)

M

 

D's89IROCZ, I'd think that your setup would cause the computer more trouble due to the extra intake work -- and if yours idled without a problem, then there must be something else going on with mine.

I was able to do some non-EPROM work tonight before I had to go and fix a friend's DSL Internet connection. Mechanically, everything still seems sound:

I checked (again...) for vacuum leaks, and there were none. The intake tube is as solid as ever; the clamps are tight.

I reset the fuel pressure back to stock.

Injector impedances were all either 16.4 or 16.5 ohms. The reading would start out at around 18.0, but after a few seconds would drop to the final value. They all did this, so I'm assuming that's normal.

funstick, I'll try grounding the MAF to the battery. I'm going to do another log, before I try that, and see if I can reproduce the same surging and "lean" issues. I might try another BIN or mask altogether to see if things get better or worse.

Sorry 'bout that... I meant to say that it won't log the instance that a code gets set. For example, I can't see what data frame trips a code.

 

I think i started the issue about logging the codes and when they occur. my 8D/730 does, maybe '165 doesn't?

then again, I use diacom.

 

I am having a simalar issue. I would like to try the suggestion of grounding the MAF to the battery.

Where do I hook up the ground wire to the MAF?

 

Sorry if this is'n't appropriate. But I had a big surgeing issue when the car was stock ..... I was getting a MAF burnoff relay code ...while driveing but .... I had a connection issue to the relay ( exposed wires ....yanked out somehow) . My car surged while going down the road ....bad enough to really require alot of brake preasure. This occured once and a while. I have seen it on another board aswell..... was the burnoff relay .....but have NO clue why it would make it surge

Just thought I 'd give ya a reaching possibility

 

Thanks for the relay suggestion, D's89IROCZ. All of my MAF relays are less than three years old, but it won't hurt to check them again. I think I'm going to run through the MAF Code 33/34 flowcharts in the GM manual, using the links that 91L98Z28 posted a few days ago.

jackballs529, check out the links I'm referring to (above). They include a diagram of the wiring, so finding the grounds shouldn't be hard. I think they're pins A and B, but better check again just to be sure.

Oh, I forgot to mention that I was able to pull five plugs without having to remove the headers... they were clearly indicating a rich condition, but weren't all caked up like before. I dunno...

Time to go play in the garage .

 

This turned out to be a rather interesting evening. Before firing the engine up again, I went through all of the wiring for the MAF sensor and fuel pump -- basically, the little block of relays next to the power brake booster.

I wasn't able to follow GM's flowchart/checklist exactly for some reason -- certain steps just didn't work out. Instead, I simply looked at several wiring diagrams and invented my own "troubleshooting checklist." None of the wires are shorted and every component gets power when it's supposed to, with the correct voltage. Interestingly, I found that the fuel pump connector was wired different from how the relay operated: pins E (hot) and D (ALDL G) are supposed to be normally-closed, but on my car the normally-closed pins were E and A (fuel pump). I don't think that this really mattered, since the relay was still connecting the right wires when energized, but I fixed the mis-match anyway. I attached a diagram of the circuit just for effect.

I made two runs tonight. The first one, using my baseline ARAP code, was not very good. It drove alright, but got very rich (BLMs at 108) under light throttle and at idle. The motor would randomly "fall on its face" without me even pushing the gas. After I came home and put the car in Park, the RPMs went up to about 1800 and stayed there. IAC steps were at zero, so I have no idea where it was getting the extra air from. Code 33 would intermittently cause the SES light to come on and off while driving.

The second run I just don't understand. I reset the minimum air to about 650-700 RPM (a little lower than before), and turned on the engine without even touching the chip. It fired right up, went to its idle of 750-800 RPM, and hung there as stable as ever. Driving felt great, with the only hiccup being a very minor jump under deceleration, but I hardly even noticed it. BLMs were much better, with the usual value between cells being 135-140. It set Code 33 and 34, this time constantly.

So, I ask myself, what could have caused such a drastic difference between runs? I don't think that the minimum air really changed much in terms of air, but I realized after the second run that I forgot to reset the TPS voltage -- it was now at 0.43 volts. Could the reduction in TPS voltage have compensated for something else? My other thought was that maybe just driving it for a little bit allowed the ECM to stabilize itself between runs. Is that a possibility?

On a side note, a small child (probably about seven or eight) saw me driving up the hill to my house. "Your car sucks!" he exclaimed in a tinny kid-voice as I drove past. I just laughed, threw it in neutral and gave him a hearty rev. Hmm, I guess his rollerblades were just too cool for me .

 

I'm starting to think this is the never-ending story!

I want to make sure that I understand this right:

Prior to first run: car idles fine (sort of) at 800rpm (give or take)

You make no changes.

You make the first run. You observe BLM's typically at 108, and an intermittent code 33.

AFTER the first run the car for unknown reasons idles at 1800rpm, whereas it didn't before.

You adjust the min idle speed lower.

You make NO other changes.

You drive the car again, and observe near perfect manners, and BLM's typically in 135-140 territory with a constant code 33 and 34.

Have I summarized that right? If so, I submit that at least one (or more) of the following must be true:

1) your TPS is toast in ways that I can't logically explain (code 33 and 34 can both be related to a TPS problem)

2) your ECM is toast. maybe it's not supplying a proper stable +5V reference voltage to the sensors and you're randomly getting crap bad data (like your MAF dropouts). We still haven't explained why you get a huge reduction in pulsewidth when the INT resets to 128 for no obvious reason, either. If the ECM lost control of the IAC motor, this could potentially explain your high idle when the ECM thought the IAC was at 0 (and it actually wasn't, perhaps).

3) you have a sizeable, variable vacuum leak which started leaking more after the first run (raising the idle speed), and also leaning the engine out some (resulting in better operation/drivability on the second run). Can you post data logs from both runs?

4) the new MAF is bad. (have access to another car you can swap with?)

5) as of yet unidentified wiring/electrical gremlins. (Are you using good, shielded, RF blocking plug wires? no solid cores, right?? Are all the engine grounding straps in place? etc..)

6) your car is posessed by evil demons who seek to make your life difficult (or at least convince you to convert to '730 /$8D).

Anyhow. I would concentrate on fixing the MAF codes. you can't possibily diagnose/tune drivability issues while you're throwing code left and right.

Also, next time you get a high idle problem, leave it running(!) at the high idle, open the hood, unplug the IAC, kill the engine, and restart... If the idle is different, then you have a vacuum leak somehow, somewhere. (the IAC was unplugged so it can't have moved to give you the idle speed difference). The idea is to see if you have a vacuum leak, or the ECM is losing control/track of the IAC position and thinks it's at 0 when it's really hanging open letting air in. I doubt the IAC is sticking/not working since it does such a good job of helping to save your stalls when you are idle surging.

Are you SURE you have no local junk yard access to a '165 (or it's service replacements 16198259/88999194?)? They came in a wide variety of vehicles, including:

87-91 s10/sonoma/s15 2.5 pickups
87-91 postal vehicles (post office LLV's)
87-89 astro
87-90 safari
86 sunbirds, firenza's, cavalier, cimarron, skyhawks (possibly manual trans cars only?)

reference also:

https://www.thirdgen.org/newdesign/t...erchange.shtml

 

Dunno if this would be of any help or not, but I have a ZZ4 TPI car that's wired so that I can switch back and forth between MAF and MAP.

FWIW, mine's really happy with about 24-26° of timing at idle, and it's not that much worse off with 20-22°

Twiddling around with spark and fuel on the romulator, I've gotten it to idle fine even down to 550 RPM when warm. At 700 RPM, which is my normal target warm idle, it's very smooth.

So, what I'm saying is that the ZZ4 cam probably isn't the source of your headaches.

I haven't run the MAF setup in a long time, but with the '165, it would idle OK on a stock AUJM (only changes being VATS disable, and inejctor constant set to what I have). Wasn't perfect, but it would run well enough that you could drive it around. Also mine will idle stable, although a little high, with the MAF completely disconnected, just running off the backup routines in the PROM. I never tried driving it like that, though.

Like others, I'm also suspicious of a vacuum leak, or perhaps a bad MAF.

Good luck

 

Does anybody want to buy an '86 IROC in good condition? The only stipulation is that she doesn't run...

91L98Z28, your summarization is completely correct.

Today, in an effort to rule out practically every location there could be a vacuum leak, I removed the throttle body and installed new gaskets on the "TPI plate" and the IAC manifold. I tested the IAC motor while it was out, and it worked just fine. I couldn't find gaskets at the auto parts store, so I made some out of a sheet of intake material I had laying around. I coated both sides of each gasket with copper spray. These things should be as SOLID as ever now.

3:28pm
Could not get idle correct. Felt like it was firing on seven cylinders. Removed each plug wire, one at a time, and the motor seemed to run better in all cases. Checked all electrical connections.

Replaced brand new aftermarket MAF with old OEM piece (once thought to be bad, but figured it worthwhile anyway). Drove alright, but still "fell" upon deceleration (RPMs would drop and then return to normal). RPM ran up to 1500 just before turning off the engine. Pumped accelerator a few times and returned to 800 RPM.

Started engine once more after setting minimum air again. Required depression of accelerator to fire, but after a few seconds idle was normal.

Letting cool down to test cold start operation.

4:51pm
Died immediately upon startup, but second try worked fine. RPMs still drop upon deceleration. BLMs look normal, but at 160 in cell 0.

Code 33 was set within the first 90 frames, but MAF airflow was no higher than 10 g/sec .

5:14pm
Set total timing at idle to approximately 22°. Seemed to start up easier. While sitting at red light, target idle went to 1120 RPM and stayed there (it set itself to that before I stopped, but never lowered upon MPH=0 or BLM cell=0/1).

Frame 41, MAF signal dropped out to 0.12 g/sec. Normal after that.

Code 33 set again within first 112 frames, but no g/sec value was above 45.

5:32pm
Set MAF High Diag. Error Threshold to 255 to circumvent Code 33 (for testing only). Would not idle smoothly, blew black smoke from tailpipes. Tried previous chip and was fine. Tried this chip again and it wouldn't idle. Dunno...

5:44pm
Raised MAF Table #1 intermediate values to help richen idle. BLMs still at 160 in cell 0. RPM drop every time INT resets at idle, but does not surge uncontrollably.

5:55pm
Decreased fuel injector constant to 19.01 lb/hr. BLMs lowered to 156 in cell 0 (doesn't seem like much for decreasing FI size by three points). High idle returned before shutdown.

I also tried lowering the fuel injector constant further, but the engine would not even run (fired once, stumbled, died). I reinserted my chip from 5:44 and it required slight accelerator pulsation to get it to start -- once I stabilized it, (I presume) the IAC took over and lowered the RPMs to idle speed. Chip 5:44 never required the accelerator to stop before.

The following is a summary based upon what I've learned today:

Surging: While driving, it seems to occur because the pulsewidth becomes equal to ZERO -- then the stall saver comes in and saves the day. DFCO? While idling, it occurs when the INT gets reset to 128 (just like the original problem!).

Mixture: BLMs are only at extreme values (160, in this case) in cell 0, at idle or while coasting.

High idle: IAC seems to become "stuck," never reaching zero to correct for the increased RPM. Target idle is always 775 RPM under these circumstances.

On another note, I should mention that at the end of the 5:55 log, I got pissed and ran the engine up to 4500 RPM in my driveway (real good for a new motor, huh? ). I figured I should investigate the log when I witnessed black smoke pouring out of my exhaust. The load variable was 160 and 208 for the two frames (awfully high?), and the total spark advance was 24° (awfully low?). Interestingly, the spark advance relative to the base timing (in other words, what the ECM adds) was 33°. How can that be higher than the total timing advance?

Another day of frustration, it seems. I dunno, but I'm really getting the impression that the ECM is at fault here. The random "dropping out" of sensor data, the unusual IAC behavior at high idle, and irrational feedback (I lower the injector constant and nothing happens) leads me to believe this. Plus, some of these symptoms just seem so intermittent. I'm going to the junkyard tomorrow to find an ECM; thanks for the crossreference.

I have data logs from every run, but didn't want to upload them since they total about four megabytes. Let me know and I can post them if anyone is curious.