cardo0

Ok i will admit i personally dont know how many samples are averaged into each "snapshot" data point the ECM collects then displays. But a lean test with a controlled leak and then a rich test condition using propane should verify the data one way or another. Whether its lean or rich to begin with doesnt matter as we are looking for comparison of before and after - relative direction of sensor output. If u cant see and understand that i cannot help u either.

As for narrow band accuracy using the stock control band around stoichiometric of 14.7 A/F the narrow band O2 sensor is much more accurate than the Wide Band sensor. The Narrow Band sensor is limited to a very narrow band between 14.55 to 14.85 A/F. Correct, this is no way a ratio for performance. But we are not going there on this car. We want it near stoichiometric (lambda = 1) for the stock tune in the ECM. What im saying is buying a quality $300-$400 Wide Band will buy u nothing for a stock tune and is actually less accurate near lambda = 1 as the Narrow Band sensor will have a 200mV spread/band the Wide band will only show <10mV for that same spread/band. But the expensive Wide Band sensor has its own possessor so it will display current A/F ratio, lambda or maybe even Phi. Also get ready to calculate temperature and press errors into your A/F ratio reading for the Wide Band sensor. Finally your stock ECM wont know what to do with Wide Band sensor output so dont expect it will replace your stock O2 sensor.


BTW tuning for performance using a wide band HEGO your gonna need a entire new set of VE tables and ignition maps and u may need to tune in the open loop mode full time - be ready for some advanced tuning. Wide band tuning is best done open loop - not the direction OP is going now. In that case for performance a Wide Band HEGO would be an requirement.

My advice dont buy a Wide Band sensor until u know what it will do for u. If u can borrow or rent an Wide Band for checking your ECM/Narrow Band data then it would be useful.

Red1990VT

Hello Everyone,

Great to see the continuing discussion, with lots of good ideas...

I did know what TCC stood for, thanks, and thought that maybe the ECM was (through all the various sensor inputs) controlling the lockup. I recall the ECM controlled shifting across all gears didn't come until later, but in the 82 the ECM does seem to decide when to lock and unlock the transmission clutch. And it seems to do it at very low speeds, consistent with the GM early 1980s desire to get into top gear as soon as possible for gas mileage. I could unplug it, but I'm not sure how that would help as far as the hesitation is concerned. Eventually I want the car to use the lockup.

The behavior of the IACs does have me a bit stumped. From my testing, it does seem the IACs open up enough to keep the car moving at 35-40MPH on level ground, I'm just not sure why it is happening. Plugging the IAC ports stops that from happening, so for some reason the ECM is commanding the IACs to open at 35 MPH with the throttle plates closed. My thought was that the ECM was trying to correct an overly rich mixture, but again, why is it doing that? IF the mixture is too rich, either there is an ignition miss, or something else is at play.

One hint might be that using an IR temperature gun, I found that the CAT has apparently equal input and output temperatures. Yesterday I bought an IR gun, and measured around 439 degrees F at the CAT input flange, and around the same at the output flange. The shield, which really covers the CAT well, was around 190. The equal input and output temperatures would seem to indicate a nonworking CAT - I don't know if that means it is clogged. From what I've read, the output should be at least 20 degrees or more hotter than the input, and it isn't. At times it seemed like the output was actually lower than the input. To get ready for this, I let the car idle up to operating temperature (recently installed 195 degree thermostat) drove the car for about ten minutes, and then measured with it in the driveway and idling. And again, roughly equal input and output temperatures, or even lower output readings.

It certainly seems there may be something wrong with the CAT after all, even though the vacuum test (3000 RPM, steady vacuum, not a slow decrease) doesn't indicate a clog.

So while the O2 sensor does show what appear to be correct voltage swings in the data (for the most part) the readings may not be good data given the CAT input and output temperatures are so equal.

I don't know if I have strong enough data from the IR test to replace the CAT at this point or not. And if I do, have I still got another problem that will simply ruin a new CAT anyway? I still have to check the ignition module and pickup, and if there is an ignition induced misfire, the unburned air fuel mixture (rich) would likely cook a new CAT as it may well have done to the current one. (I should mention the exhaust on this car looks original, and it may just be 32-33 years of use have used it up).

What a puzzle. Any additional thoughts and advice are appreciated. I'm not sure a Wideband O2 sensor would help at this point, given the CAT readings, happy to hear thoughts on that too...

cardo0

FYI u can buy cat eliminator pipes that would allow u to do your testing. Or maybe even just gut the cat u have now - i think the early cats like yours are pellets that can be removed w/o to much effort.

Something u need to do is read the instructions on your IR temp gun. U will find IR sensors dont work well with shiny metal surfaces - a common mistake with uninformed user. Its got to do with the emissivity of IR but dont lose sleep over it now. I think u just want to prove the cat good or bad for now.

Myself i would run a straight pipe until testing and repair complete then upgrade to a new high flow replacement cat - much less restrictive than the old stock cats.

If u find its not the cat causing the bog/stall/high idle driving then u may need a good ECM. I believe Howell performance FI still supports these. Howell also sells their own TBI service manual that includes troubleshooting for $12: http://howellefi.com/tbi-service-manual.html.

Hope this helps.

lionelhutz

Do you understand what the other sensors do and what the data represents?

I doubt the cat is plugged but you can try replacing it if you want. I would expect the BLM data to drop below 128 if the cat was plugged since values below 128 indicate the ECM is removing fuel because it is running too rich.

There is no averaging of any kind on the ALDL data. The O2 voltage you recorded on each line of your log is the instantaneous voltage when that log data was captured and transmitted from the ECM.

cardo0 - Your assumptions about the ECM averaging the O2 data are wrong. There is no averaging on the O2 data internally or externally. Your test would prove the O2 shows a change in AFR but it's rather pointless since the log shows the O2 voltage widely varying so the O2 is already measuring the changing AFR.

The O2 voltage constantly swinging all over the place between 0V and 1V is the reason I suggested using a wideband O2 to verify the AFR. It has nothing to do with the accuracy of the sensor. A wideband will give a nice linear voltage (or even a display) which is easy to read. You could also replace the O2 sensor or even install a second narrowband O2 sensor to verify the current one. Either way would work, but the wideband 02 tells you the AFR while you have to interpret what the narrowband O2 sensor is telling you.

Also, there are NO O2 sensors which measure the AFR, say 12.5:1 for example. ALL O2 sensors measure lambda. A O2 sensor will only tell you a ratio like 14.7:1 if you take the lambda reading and multiply it by some theoretical AFR such as 14.7:1. The reason is that not all fuel has a perfect mixture with a 14.7:1 AFR but all fuel is at the perfect mixture when lambda = 1. Same as power enrichment. 12.5:1 may be perfect for one fuel while 12.1:1 may be perfect for another fuel. But, lambda = 0.85 is perfect for both fuels.

The above makes a complete mess when you "calibrate" your wideband O2 for 14.7:1 fuel and then start tuning with E10 blends and expect to reach some best AFR that you read online, such as 14.1:1 in closed loop or 11.5:1 in power enrichment. See how using lambda and tuning for 1 in closed loop or 0.85 in power enrichment eliminates that mess?

A narrowband O2 (which is the one used in your car) just tells you if the engine is richer or leaner than lambda = 1. That is all it does. Typically, about 0.45V is lambda = 1. Any value above 0.45V is rich and any value below 0.45V is lean. A narrowband O2 is completely useless for any type of lambda measurement other than 1.0.

alconk

Whew.. this is getting confusing. But interesting..

cardo0

U seem to ignore the wide band will display what ever u select it to if it has that feature/capability and many wide bands can display A/F. The result is the owner operator can read a direct value from the display - something he cant do with the stock ECM. Yes it will display A/F ratio if selected.
Like i posted previously the ECM has to trend the variying O2 output or it cant control A/F in closed loop. In that case then sensor output voltage band/sweep has to be representative of A/F or excuse me lambda (where 450mV is "stoich" regardless of fuel). If it only used an instantaneous value it would be totally random as the output varies. And in Red's spreadsheet i saw consistent high values for O2 above 2000rpm.

Your "stoich" value for blended fuel is lower and sure lambda still equals 1 when all the oxy molecules are used up and no excess fuel left. But again the ECM wont let u tune to lambda 0.85 in closed loop. So why do u need a Wide Band? U cant tune the stock ECM closed loop with it. Go open loop tune? Again u better be ready to change all your timing maps let alone fuel corrections - depends on your tuning software how it gets done. To much fuel and to much timing at high rpm then get ready for loud noises.

And u totally ignore the Wide Bands need a temp and press correction. On top of that where u gonn'a install the probe??? Not in the tailpipe after the cat! Pull the stock O2 sensor then u cant run closed loop to watch it. Drill the pipe and weld in a new bung? Really the only way so be ready.

If your ready to tune creating new timing and fueling maps then buy a Wide Band. Use the stock tune then stay stock.
Enough said.

Red1990VT

Hi Everyone,

I have been attacking this problem with the thought that at one time the car worked well from the factory, with the factory PROM and stock tuning installed. And that whatever is going on is a failure or breakage of something. Remapping or retuning the car now would I think be tuning a car to work properly with some kind of issue - an issue that might continue to shift or change over time, and negate the new maps.

I kept wondering why the car would act like it is on cruise control and do 35-40MPH on level ground with the throttle plates closed. I had previously shown myself that didn't happen if the IAC passages were blocked, and then theorized that the ECM must for some reason be wanting more air and commanding the IACs to open up. What I overlooked was that despite having previously cleaned the IAC passages, the IACs might simply be sticking. So I pulled both, cleaned and carefully lubricated them, cleaned the passages in the throttle bodies again, and then observed both IACs operating properly (connected but uninstalled in the TB bases, ignition on, start, run, ignition off, ignition on, start, run, ignition off) and verified the pintles do move in and out properly under ECM command.

I put fresh gaskets on both IACs, carefully reinstalled them, reassembled the car, and test drove it. The cruise control behavior appears to be gone. Lift off the throttle, the speed of the car drops off along with engine RPMs. Come to a full stop in drive, the car idles down to a smooth 650. So maybe I was overcomplicating things by theorizing a rich mixture due to a misfire and consequent ECM opening of the IACs, when in fact it appears the IACs were simply getting stuck open.

As for the CAT, I also tend to think it is not clogged. I actually had a CAT get clogged on a 97 Grand Cherokee some years back, and the 82 is not acting like that at all. The Jeep would stumble and lose power under many more circumstances, and feathering the throttle did little to bring the Jeep back. With the 1982, it will hesitate or stumble, but not a consistently as the Jeep with the clogged CAT. I realize this is a loose comparison, and experience with one clogged CAT may not extend to others, but it does seem like the 82 has moments where it functions normally, in a way the Jeep with the clogged CAT never did. Once that CAT clogged, the Jeep became dangerous to drive, and I had to replace the exhaust system CAT back. Which completely fixed the problem.

I still think from everything I've read and done that this is most likely an ignition problem. It may not be, but that appears to be the strongest candidate, given no CODES, no apparent vacuum leaks, correct timing, good fuel pressure, good fuel cones on both injectors, and the fact that the car is strong other than uphill under load. I suppose it could still be a fuel or vacuum problem, but I'm thinking getting the ignition system fully vetted (no pun intended) and known to be working properly would be the most prudent next step. If the problem persists after that, it will have to be either air or fuel.

Still interested in any thoughts and opinions, and appreciate everyone who has been helping with this.

Red1990VT

Oh, just to be clear, any abrupt opening of the throttle plates will still stumble and hesitate the car, so it isn't just uphill. But the problem is most pronounced above second gear, with the engine reving above 1000 RPM - the FSM says that a car that idles smoothly can still have ignition problems at higher speeds, as mine seems to have - hence my leaning toward the ignition system.

Cleaning up the ignition system will take a bit of time given all else I have to deal with (including a leaking fuel injector in a 1990 and a leaking power steering system in a 1978, sigh) so stay tuned...

lionelhutz

Wow. Nope, I did not. You just failed to comprehend what I wrote.

If you only select AFR on the gauge then the gauge is ASSUMING 14.7:1 for the ideal AFR. If you select AFR and entered 14.7 then you have ASSUMED 14.7:1 is the ideal AFR.



NO, THE RECORDED O2 DATA IS NOT TRENDED OR AVERAGED. You have repeatedly demonstrated a complete lack of understanding of how the O2 sensor voltage is used internally in the ECM. The ECM simply uses the instantaneous O2 voltage to determine if the engine is rich or lean and adjusts the fuel accordingly.

You have to watch the INT and BLM if you want to determine if the engine is running rich or lean while in closed loop.

I do question the reliability of the data, because for all the recording above 2k rpm the closed loop flag is on yet the INT is locked to 128. At > 60% TPS it should also be in power enrichment mode, not closed loop. But then, for most of the recording the learn flag is off so maybe that flag really represents closed loop.

I don't believe the ECM is very sophisticated in the way it transmits data. I don't believe all the data for each row in the log represents the exact same point in time. The code will just grab each value as it transmits it, meaning some are grabbed earlier in time than others. This could explain the low %TPS and high MAP lines if you were snapping the throttle closed.

Red1990VT

I'm a little hesitant to step into the middle of this, but...

I thought the ECM did average the O2 data over time in order to fill in the Block Learn table entries. That said, whatever is in the spreadsheet is a 160 baud readout of the data stream as it's happening, and would not necessarily reflect any averaged O2 data...except possibly the entries in the Block Learn table? If that is not correct, I appreciate any needed updates...

I've long wished I could find some technical documentation on that ECM...I'm an electrical and systems engineer who used to design and lead teams designing defense electronics - I still have a pretty good understanding of digital systems technologies and architectures. Especially from the late 70s and early 80s when I was actually doing hands on digital design work prior to being promoted out of that. I've long searched for any kind of spec or interface control document for the GM fuel management systems from that era with no luck. If anyone has that kind of documentation, I'd appreciate receiving copies of it.

Also, as an aside for aspiring tuners...the PROM for the 82 and 84 Crossfire systems, AM27S181, is a hard to find part that requires a very expensive and hard to find programmer (I bought several blank PROMS from China before finding out about the special programmer)...so even with updated wideband information, getting it into a suitable PROM might be nearly impossible to actually do (unless you're with Hypertech or willing to spend $2200 for an old programmer, if you can find one). So, I don't really entertain thoughts of ever tuning my crossfire system, I just want it to work as it did when it left Bolling Green.

All that said, it seems the processor and sensors are functioning normally. There are no codes, good idle, ok performance except under load. Making it seem that there is something else broken in the system, like an overheating ignition module, or clogged CAT (which I think we've ruled out). A wideband AFR would be interesting to use, but I would think mostly provide additional detail on what we already seem to know...the car is appears to be running very lean (no power enrichment?) under load.

If the Wideband AFR would tell more than that, it might be worth trying, other than the cost, which I haven't investigated but sounds relatively costly from what I've heard and read.

Red1990VT

One other thought involving gas, ethanol, and tuning. I wondered if modern gasoline with ethanol might not work well with the tuning tables from the 1980s, when gas was just gas.

Turns out, my 1984 Z51 (with essentially the same engine and crossfire system as the 1982) answered that question - it always ran fine on modern regular unleaded fuel.

Kacyc3

Disconnect the tps and take it for a drive, I had a similar issue with a truck and this helped me figure out the tps was bad even though the ecm never threw the code.

Red1990VT

Good idea thanks, I'll give that a try and post the results. I've tested the TPS with an ohm meter, sweeping it and not seeing any dropouts or flat spots, and watched the WINALDL readout of the TPS, but I suppose it could still be problematic despite appearing to a pass those tests. Thanks.

terry82

something I would try,uplug ecm .check connections.there is something simple being over looked.it sounds like you have tried all the right things do look for .have you thought about trying another ecm?

Red1990VT

Terry,

Thanks, actually the ECM I'm using is a Cardone rebuilt replacement for the original. That one was corroded from being mounted next to the battery for so many years.

The prior owner had the prior ECM being used with a Hypertech PROM, and the car exhibited all manner of strange behaviors, especially once the ECM got hot. I replaced that ECM with a new/rebuilt one, being very careful to clean all the connectors well. I also added a new spark control module and a factory PROM (from Tony's Corvettes in Gaithersburg MD, thanks Tony) at the same time.

Since then, the electronics have all seemed to work well, bearing in mind that the electronics in the distributor could still be original and causing problems.

I wouldn't rule out the ECM by any means, especially given it is a rebuilt unit, and I'll try pulling and reseating the connectors...but especially given the behavior I saw with the original failing ECM, this one does seem to be operating ok.

Fun trying to solve a problem with so many interdependent variables all at play at the same time...

Drawmain

I've followed this thread for a while now. I still think you need to check your injectors. Here's why:

I rebuilt my engine last year. New heads, new cam, new everything pretty much except the block. The engine ran great until January, then I started having the same behavior you are now experiencing. I replaced the entire ignition, wires, plugs, distributor, re installed the intake to ensure no leaks. The only vacuum ports are for the brakes and the climate control system, no egr, and my lights are electrically actuated, cruise control was removed when the cable quit. Remember, this is a freshly rebuilt engine with new injectors that's stumbling all over the place. Next I replaced the ECM with the EBL unit. Same behavior. Fuel pressure fine at 14 psi, injectors appeared to be fine with great cones at idle, balance set and verified, re-verified, and re-re-verified. Datalogs showed VE at 100 at many places in the map. That meant injectors, so I pulled my old injectors out the box and reinstalled the old injectors and all my issues went away. No stumbling, no bogging, no issues. Try it. What have you got to lose? The injectors I removed were only 8 months old. If I'm wrong at least you've eliminated the injectors as the problem.

cardo0

NO, THE RECORDED O2 DATA IS NOT TRENDED OR AVERAGED. You have repeatedly demonstrated a complete lack of understanding of how the O2 sensor voltage is used internally in the ECM. The ECM simply uses the instantaneous O2 voltage to determine if the engine is rich or lean and adjusts the fuel accordingly.

I do question the reliability of the data, because for all the recording above 2k rpm the closed loop flag is on yet the INT is locked to 128. At > 60% TPS it should also be in power enrichment mode, not closed loop. But then, for most of the recording the learn flag is off so maybe that flag really represents closed loop.

No! That fact is u havent explained it correctly either - because u dont fully understand it either but try to credit yourself as such.
Ok after some research, one last time. The short term fuel trims = integrator cause the O2 sensor to cycle between .1V and 0.9v in closed loop - not by the O2 sensor itself. The output is a direct reading but always changing due the integrating fuel trims. FYI the long term fuel trims = BLM follow and lag the integrator to neutral it out at 128 value (it just a number meaning neutral integrator).

Once in open loop due to throttle position or engine speed or manifold pressure or whatever the integrator value goes to 128 = neutral something u clearly stated u dont understand. And the BLM follows to near 128 (in Red sheet its 129). But the O2 sensor output is what it is at that time. It should be somewhat steady (and it is in Red's above 2000rpm and at high MAP) but do to engine conditions/performance at high power it still changes a little (and it does in Red's chart).


Now back in closed loop the integrator followed by the BLM is creating a varying output which is whats happening oxygen wise in the exhuast. So yes in closed loop it reads the present value but the changing due to the short term fuel trim and controls fueling by the output voltage trend. Yes a output trend >450mV will be rich and cause the short term integrator to lean the a/f.

This from post #25 is just not true: One other thing to note, when in closed loop you can't just examine a single line and say the O2 is reading <0.45V or >0.45V so the engine is lean or rich at that point like cardo0 posted. It just doesn't work that way. When in closed loop, the O2 sensor voltage will constantly oscillate above and below 0.45V as the control loop centers around stoichiometric. The only time the O2 can be of some indicative value that way is when you are in open loop power enrichment mode where the O2 should show read rich.


Too bad the spreed sheet doesnt have pulse width or timing values.

Good luck Red and good night.

lionelhutz

No, it does not. The ECM has limits like 0.44V and 0.46V. These voltages can vary by vehicle and the is also typically a table that varies it as the MAP or TPS changes. Often, the limits are not centered around stoichiometric so the engine can run slightly rich or slightly lean. The limits are always set right around the lamda =1 point of the O2 sensor. This is common with GM fuel injection through many years.

The control loop increases the fuel until the O2 voltage goes above the higher voltage limit and then it lowers the fuel until the O2 voltage goes below the lower voltage limit.

The controller never tries to swing the O2 voltage between 0.1V and 0.9V. The O2 goes past those voltage limits because it's so sensitive to changes in the AFR. So, the voltage might reach 0.1V by the time the control loop reverses. BUT, it does not need to reach 0.1V. It might only drop to 0.3V or 0.2V or something like that. Same applies to the higher voltages.

I've already posted this so you're not explaining anything new here but rather just posting wrong information.


Huh? I just posted that. High throttle and integrator locked at 128 should mean open loop and power enrichment mode.



The BLM doesn't create the varying O2 voltage in closed loop. The BLM is a correction factor. To be technical, the integrator doesn't create the swinging O2 voltage either, the control loop does and the integrator is the just integrator factor from the control loop.

Overall, good job copying an explanation from somewhere and still getting it wrong....

Red1990VT

Hi Everyone. Great to see this thread still going strong.

Last evening I disconnected the TPS and took the car for a drive. Unfortunately it did not change matters as far as bucking and stumbling. The check engine light came on with the TPS disconnected, and went back off when I reconnected it to drive home. (Drove the car about ten minutes over varying terrain at operating temperature. With the TPS disconnected, the car stalled often, as I expected it would - but it did not stop the hesitation and stumble under load, in fact it seemed worse, which made sense since the ECM could only "think" the throttle plates were closed the entire time). But it was worth a try.

INJECTORS - yes, I've been thinking this may still be the problem, and even toyed with the idea of trying to mount a small video cam to record injector behavior on a test drive. But, actually doing that is such a hassle it may be worth simply sending the injectors off for reconditioning and flow testing. Since the car behaves the same way my first Silver Anniversary did when that car needed a new accelerator pump in the 4BBL carb, this 82 might simply not be getting enough fuel injected when it needs it most. The car will be down for the time it takes to send the injectors off and get them tested or replaced, but that may be worth it versus troubleshooting this forever. And finding out the car needs fuel injectors in the end anyway. Been my suspicion for a while, and I think it might finally be time to pull the injectors and answer that question once and for all. Thanks for the thought.

As for the discussion of INT, BLM, and O2, it is really interesting. I wish the WINALDL table had pulse width too.

Does anyone have a readout or document of what's in the PROM? It is only a 1K X 8 PROM, so it cannot be a lot of data. I found a bin on line, but that is probably just a hexidecimal dump - I'd really like to know what the VE table and whatever else is in there has in it from the factory. It would be tedious but I've even thought of setting up something to manually toggle address lines, and read leds off the output and then try to decipher what it all means. But at this point that's a bigger pipe dream than putting a video camera near the fuel injectors. If anyone has a map of what's in the PROM, please share it - it might not change a thing on this thread but it would be interesting to know.

Thanks to all who are continuing to help and participate in this thread - its' really nice to see and I'm learning a lot.

Red1990VT

Hi Everyone,

Just got off the phone with John at Fuel Injector Connection - he really knows Crossfire systems. After talking a while John said he doesn't think the problem is my injectors, although I could still get them serviced if I like, but he does not think that will solve the problem.

What he did recommend however, and I have not previously thought of it, is looking at what the knock sensor is doing under load. If it is for any reason causing the ECM to pull timing out, the car will die, just as it has been doing. This makes sense, and I've really been kind of taking the knock sensor for granted. After all, base timing is set at 6 per FSM, reconnecting the lead shows advanced timing that varies with throttle, and I never hear knocking audibly while driving. But all that means little in terms of the knock sensor, which could be picking something up under load and retarding timing to the point the car stumbles. Interesting lead to follow.

Also, some of the things that could trigger the knock sensor would be a misfire (!) or a rattling catalytic converter, among other things. I've been assuming the slight roughness at 2000 plus is an ignition problem, when in fact it could be a timing problem. Or an ignition problem causing a timing problem. Got to think about this...

John also recommended a performance expert local to me and I'm going to call that person and if he is up for it, bring the car over to him.

Stay tuned...thoughts and ideas always appreciated.