5.0 Ho Swapped Comet Lean Backfire Issue

jas280z

Active User
Aug 18, 2017
13
0
1
32
Hello all.


My dad recently picked up a 1964 Mercury Comet, with 5.0 HO and AOD out of an 89-91 Mustang GT swapped in place of the original 289 and C4.


We are having some issues with lean backfiring on slight accelerator application on occasion which I am trying to work through. When the car is able to backfire, it tends to idle roughly. Other times, it idles very smoothly and won’t backfire no matter what is done to the accelerator.


Right off the bat, I have replaced plugs, plug wires, distributor cap and rotor. The issue was happening before and after the swap. The coil has not been changed and still appears to be OEM 5.0 HO.


I checked for vacuum leaks by spraying some carb cleaner around the intake and vacuum lines. I found no signs of any leaks.


I checked fuel pressure at the rail, got ~ 33psi at idle and ~40psi at idle with the FPR vacuum disconnected and plugged.


The car does not have a smog pump, charcoal canister, or a functioning EGR. Those did not make it over in the swap. The EGR valve is in place with no vacuum plumbed, it is just acting as a poor man’s block off plate at the moment.


I have been posting at corral.net, the thread is here. http://forums.corral.net/forums/gen...0-ho-o2-sensor-wiring-check.html#post18059513


My research has shown this site to be a wealth of knowledge, so I figured I see if anyone here has any help to offer. Quickish explanation of what I have done up to now below.




I ran codes on the car, KOEO and KOER. Initially, I got the following...


KOEO:


85 O: Charcoal Canister. This makes sense, since that system did not make it over in the swap. I am planning on living with this code, since it shouldn't affect anything else from what I have seen.


22 CM: Barometric pressure sensor. I replaced the BAP sensor, in doing so I found that the BAP sensor that came in the car was hooked to manifold pressure. I did not hook the new sensor to the manifold. Code 22 has been gone since.


33 CM: EGR. This also makes sense, since the vacuum is not attached to the EGR system. I don't think the whole thing made it over in the swap. I plan to remove the EGR valve and block it off, kit from LMR is on order. I wouldn't mind having EGR if it had been completely swapped over, but at this point it is easier to just completely delete it.


41 CM: I believe this means right hand O2 constantly lean, not switching. More to come on this



KOER:


33: EGR, see above


42: I believe this means right hand O2 constantly rich. Seemingly conflicting with the continous memory code 41.


Seeing what appeared to be a right hand O2 issue, I ordered a replacement O2 sensor (bosch). When I went to install the sensor, I had an issue. The car had NTK sensors, left and right. When drilling the hole in the exhaust pipe on the right hand side, they drilled it JUST big enough for the NTK sensor to fit. The bosch's probe into the exhaust stream is just a tiny bit bigger in diameter, and wouldn't fit. The threaded section is identical, just the physical probes are slighly different in size.


So, I moved the seemingly working left hand sensor to the right hand side, and installed the new bosch sensor on the left hand side, where it fit with no issue. Eventually I will get matching brand O2 sensors in place, but that is not my priority at the moment.



With O2 sensors swapped around, I cleared the codes and tried them out. I still get backfiring out the intake, and codes 41 CM and 42 R. I am still getting 85 CM and 33 CM/R, assume these to be constant going forward.


I verified O2 wiring has continuity from sensor to ECU and ground. I have not yet verified 12V for the heater.



I then used a Fluke 77 DMM and paperclip to jump into pins 43 (left hand) and 29 (right hand) to monitor O2 output. The left hand side appeared to function normally when warmed up, jumping from rich to lean and back every couple seconds. The right hand side was running constantly rich, I thought the signal may have been bad. During this testing, the car was running smoothly and there was no backfiring. I still got codes 41 CM and 42 R



The next day, I tried unplugging the right hand O2 from the harness, and jumping pins 43 and 29 at the ECU. This should have made the compute read the signal from the left hand O2 on both the left and right hand sides. Battery was disconnected for 20 minutes to clear codes. Interestingly, I STILL got codes 41 CM and 42R. No left hand O2 codes, only right hand, even though both sides were getting the same signal.


I hooked everything back up to stock, and have been continuing to monitor pin 29 with a DMM on every drive. I have found that once the O2 starts registering, it tends to read VERY lean for 2-4 minutes, and then in 1-2 seconds jumps up to rich. When the engine is running very lean, it idles poorly and is prone to backfiring. Once the mixture fattens up, it idles MUCH more smoothly and will not backfire.





I took the car for a ride tonight, with my DMM attached to right hand O2 (pin 29) and a fuel pressure gauge sitting next to it. Right off the bat, I will say that the fuel pressure is reading high. I think that is because I had to increase the hose run to ~6 feet to get the gauge safely into the passenger compartment. When I had the gauge hooked up at the fuel rail with the stock length hose, it read 33ish PSI at idle and 40ish at idle with FPR vacuum disconnected. Either way, I was mostly watching for changes in fuel pressure.


View: https://youtu.be/cjERSNT235I



With both gauges hooked up, I went out for a 12 minute ride, videoing the gauges the entire time. The video starts right after initial startup.


It took about a minute and a half for the O2 sensor to start registering. A back fire can be heard a 1:08, and a few more can be heard later on. For the first 7ish minutes, it can be seen that the O2 sensor is registering a very lean condition. During this time, the car idled roughly.


If you are going watch any of the video, I suggest from minute 7 to minute 8. at about 7:15, you will hear some backfiring under very light acceleration. The fuel pressure seems to stay pretty constant, within 1-2 psi during the back fire, and the O2 sensor shows very lean. The fuel pressure does jump a little bit during the back fire, but it appears to actually slightly increase. At 7:30, the O2 sensor shows the mixture pretty quickly goes rich, and idle quality improved. Fuel pressure remained constant during this transition. After the mixture fattened up, there was no more back firing.



Any help or suggestions would be appreciated. I am currently thinking that the fuel trims may be very lean when the ECU is operating in open loop, and it goes rich when it switches to closed loop. I am also leaning towards the ECU potentially being faulty, since it registered 41 CM and 42R but no left hand O2 codes when the left hand O2 signal was fed to both ECU pins. There also could be a leaky EGR affecting things. Once LMR gets the EGR eliminator shipped, I can rule that out.
 
  • Sponsors(?)


imp

Mustang Master
Jul 13, 2017
531
105
83
69
@jas280z
How do you know the complaint is "lean backfiring" at light throttle? From some one of the codes? EEC-IV is very "bendable" when used in conversions such as the Comet. How often is "on occasion"?

I would check initial ign. timing (10`), and watch advance with a timing light while slightly opening the throttle above idle rpms. If advance is not functioning, or inadequate, there may be the problem. Other possibility may be a leaking injector or two, contributing excess fuel which "explodes" in the cats. BTW, do you have cats?? imp
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
@jas280z
How do you know the complaint is "lean backfiring" at light throttle? From some one of the codes? EEC-IV is very "bendable" when used in conversions such as the Comet. How often is "on occasion"?

I would check initial ign. timing (10`), and watch advance with a timing light while slightly opening the throttle above idle rpms. If advance is not functioning, or inadequate, there may be the problem. Other possibility may be a leaking injector or two, contributing excess fuel which "explodes" in the cats. BTW, do you have cats?? imp

Imp, I know it is a lean backfire, since it is clear in person that the "pop" is coming from the front of the car, vs a rich after fire in the exhaust. I have also verified that the right hand O2 is showing a very lean condition when it "pops."

Agreed it can be difficult to determine exactly what you are working with on a swap. I have determined it is an 89-91 5.0 HO. It has mass air, and the fuel relay is in the passenger compartment (and uses the pre-92 relay). It also has 351/HO firing order.

Generally, I am able to get the car to back fire during light acceleration from startup to about 2-4 minutes after startup. During this time the idle is rough as well, like it has a bit of a moderate cam. Something happens 2-4 minutes after startup, at which point the mixture gets fat (per my DMM on pin 29), idle smooths out, and the car will no longer backfire.

It is not JUST a temperature thing. I have verified this by turning the car off for about 2 seconds, shortly after the mixture fattened up. The car went lean and backfirable again for about 2 minutes, then fattened back up.

From my experience, once the mixture goes rich, it stays there for the rest of the time the car is running.

My uneducated guess at the moment is the mixture gets richer upon the ECU switching from open to closed loop. In that case, the car would be running very lean during open loop for some reason.

The car does not have cats.

Checking timing is a good idea. I mean to a few days ago, but it was too bright to get a reading. I will try to get that today. May not be able to check advance without a helper, but I can at least get initial set. I also need to get my vacuum gauge on it. I should be looking for 14+ inhg, right?
 

Dan02gt

10 Year Member
Mar 2, 2003
709
297
113
Greenville, NC
Mine had a lean backfire at part throttle especially when cold and would sometimes idle rough. The codes revealed a O2 sensor code (don't remember which one). Two new O2 sensors cleared it right up.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
Dan02gt, I have done what I can a the moment to replace/swap the O2 sensors (detailed in my first post). That did not help the situation. I have been monitoring the output of the sensors, the appear plausible.

I need to still get the second Bosch sensor installed, but that will involve figuring out how to open up the hole behind the bung to allow to allow me to install the new sensor.
 

General karthief

wonder how much it would cost to ship you a pair
Mod Dude
Aug 25, 2016
18,841
6,218
193
polk county florida
A couple questions
Stock egr plate?
If so, are the water nipples open or plugged off.
What I'm thinking here is you may be getting air leaking in from your egr valve area, either the valve itself or the plate it bolts to, the plate is known to deteriorate inside having had water passing through them and causing electrolysis eating away at the aluminum and/or your egr valve is leaking and as the temps rise leak stops,
Also how far away from the manifolds/headers are your O2's?
Just thinking out loud, my thoughts are like cartoons in my head and are not based on fact or personal experience.
 

jrichker

StangNet's favorite TOOL
SN Certified Technician
Mar 10, 2000
27,374
2,707
234
75
Dublin GA
lowendmac.com
Step 1.) Do a cylinder balance test. Note any codes when you dump the codes to run the test.

Cylinder balance test: use this to find dead or weak cylinders:

Revised 22-Mar-2017 Added disclaimer for the YouTube video that shows how to do a cylinder balance test.

The computer has a cylinder balance test that helps locate cylinders with low power output. You’ll need to dump the codes out of the computer and make sure that you have the A/C off, clutch depressed to the floor and the transmission in neutral. Fail to do this and you can’t do the engine running dump codes test that allows you to do the cylinder balance test.

Here's the way to dump the computer codes with only a jumper wire or paper clip and the check engine light, or test light or voltmeter. I’ve used it for years, and it works great. You watch the flashing test lamp or Check Engine Light and count the flashes.

Be sure to turn off the A/C, have the clutch depressed to the floor, and put the transmission in neutral when dumping the codes. Fail to do this and you will generate a code 67 and not be able to dump the Engine Running codes.


Here's how to dump the computer codes with only a jumper wire or paper clip and the check engine light, or test light or voltmeter. I’ve used it for years, and it works great. You watch the flashing test lamp or Check Engine Light and count the flashes.

583775.jpg


583776.jpg


If your car is an 86-88 stang, you'll have to use the test lamp or voltmeter method. There is no functional check engine light on the 86-88's except possibly the Cali Mass Air cars.

583777.gif


The STI has a gray connector shell and a white/red wire. It comes from the same bundle of wires as the self test connector.

89 through 95 cars have a working Check Engine light. Watch it instead of using a test lamp.

583778.gif


The STI has a gray connector shell and a white/red wire. It comes from the same bundle of wires as the self test connector.


WARNING!!! There is a single dark brown connector with a black/orange wire. It is the 12 volt power to the under the hood light. Do not jumper it to the computer test connector. If you do, you will damage the computer.

What to expect:
You should get a code 11 (two single flashes in succession). This says that the computer's internal workings are OK, and that the wiring to put the computer into diagnostic mode is good. No code 11 and you have some wiring problems.
This is crucial: the same wire that provides the ground to dump the codes provides signal ground for the TPS, EGR, ACT and Map/Baro sensors. If it fails, you will have poor performance, economy and drivability problems

Some codes have different answers if the engine is running from the answers that it has when the engine isn't running. It helps a lot to know if you had the engine running when you ran the test.

Dumping the Engine Running codes: The procedure is the same, you dump the codes and then you start the engine with the test jumper in place. Be sure the A/C is off, clutch depressed to the floor and the transmission is in neutral. You'll get an 11, then a 4 and the engine will speed up to do the EGR test. After the engine speed decreases back to idle, it will dump the engine running codes.

Trouble codes are either 2 digit or 3 digit, there are no cars that use both 2 digit codes and 3 digit codes.

Your 86-88 5.0 won't have a working Check Engine Light, so you'll need a test light.
See AutoZone Part Number: 25886 , $10
583779.jpg




Alternate methods:
For those who are intimidated by all the wires & connections, see Actron® for what a typical hand scanner looks like. Normal retail price is about $30 or so at AutoZone or Wal-Mart.

Or for a nicer scanner see www.midwayautosupply.com/Equus-Digital-Ford-Code-Reader/dp/B000EW0KHW Equus - Digital Ford Code Reader 3145.
It has a 3 digit LCD display so that you don’t have to count flashes or beeps.. Cost is $22-$36.
Order it at Walmart for a better price and free shipping
583780.jpg



Cylinder balance test

If you have idle or IAC/IAB problems and the engine will not idle on its own without mechanically adjusting the base idle speed above 625-750 RPM, this test will fail with random cylinders pointed out every time it runs. The IAC/IAB must be capable of controlling the engine speed to run in the 1300-1500 RPM range. Playing with the base idle speed by adjusting it upwards will not work, the computer has to be able to control the engine speed using the IAC/IAB.

Warm the car's engine up to normal operating temperature. With the test jumper in test position, start the engine and let it stabilize. It should flash a 10 and then a 4 and maybe an 11. If no 11, then there are other codes that will be dumped.
One of the first tests it does is to open the EGR all the way, this will cause the engine to stumble and almost die. If the engine dies here then you have EGR problems.
To start the cylinder balance test, briefly floor the accelerator past 2500 RPM and let off the accelerator. The engine will stabilize at about 1300-1450 RPM and the cut off the fuel injectors one at a time. The engine speed will drop briefly and the computer will turn the fuel injector for the cylinder under test back on. Then it starts the process for the next cylinder. When it has sequenced through all 8 injectors, it will flash 9 for everything OK, or the number of the failing cylinder such as 2 for cylinder #2. Quickly pressing the throttle again up to 2500 RPM’s will cause the test to re-run with smaller qualifying figures.
Do it a third time, and if the same cylinder shows up, the cylinder is weak and isn’t putting out power like it should. See the Chilton’s Shop manual for the complete test procedure

See View: https://www.youtube.com/watch?v=HDXrkKS4jTE
for a visual tour through the process. There is no voice narration so you have to listen carefully for the engine sounds. I posted the link for the benefit of Stangnet members who had questions about how to do a cylinder balance test. I do not own that video and I am not the creator.

Do a compression test on all the cylinders.
Take special note of any cylinder that shows up as weak in the cylinder balance test. Low compression on one of these cylinders rules out the injectors as being the most likely cause of the problem. Look at cylinders that fail the cylinder balance test but have good compression. These cylinders either have a bad injector, bad spark plug or spark plug wire. Move the wire and then the spark plug to another cylinder and run the cylinder balance test again. If it follows the moved wire or spark plug, you have found the problem. If the same cylinder fails the test again, the injector is bad. If different cylinders fail the cylinder balance test, you have ignition problems or wiring problems in the 10 pin black & white electrical connectors located by the EGR.

How to do a compression test:
Only use a compression tester with a screw in adapter for the spark plug hole. The other type leaks too much to get an accurate reading. Your local auto parts store may have a compression tester to rent/loan. If you do mechanic work on your own car on a regular basis, it would be a good tool to add to your collection.

With the engine warmed up, remove all spark plugs and prop the throttle wide open with a plastic screwdriver handle between the throttle butterfly and the throttle housing. Crank the engine until it the gage reading stops increasing. On a cold engine, it will be hard to tell what's good & what's not. Some of the recent posts have numbers ranging from 140-170 PSI. If the compression is low, squirt some oil in the cylinder and do it again – if it comes up, the rings are worn. There should be no more than 10% difference between cylinders. Use a blow down leak test (puts compressed air inside cylinders) on cylinders that have more than 10% difference.

I generally use a big screwdriver handle stuck in the TB between the butterfly and the TB to prop the throttle open. The plastic is soft enough that it won't damage anything and won't get sucked down the intake either.

A battery charger (not the trickle type) is a good thing to have if you haven't driven the car lately or if you have any doubts about the battery's health. Connect it up while you are cranking the engine and it will help keep the starter cranking at a consistent speed from the first cylinder tested to the last cylinder.



Step 2.) Check for vacuum leaks.

Finding vacuum leaks

Revised 2 June 2017 to add picture of engine crankcase vent to throttle body

There is no easy way to find vacuum leaks. It is a time consuming job that requires close inspection of each and every hose and connection.

Small vacuum leaks may not show much change using a vacuum gauge. The range of "good readings" varies so much from engine to engine that it may be difficult to detect small leaks. The engine in my first Mustang pulled about 16.5" of vacuum at 650-725 RPM, which I consider rather low. It was a mass market remanufactured rebuild, so no telling what kind of camshaft it had. Average readings seem to run 16"-18" inches at idle and 18"-21" at 1000 RPM. The only sure comparison is a reading taken when your car was performing at its best through all the RPM ranges and what it is doing now. Use one of the spare ports on the vacuum tree that is mounted on the firewall near the windshield wiper motor.

Use a squirt can of motor oil to squirt around the mating surfaces of the manifold & TB. The oil will be sucked into the leaking area and the engine will change speed. Avoid using flammable substitutes for the oil such as starting fluid, propane or throttle body cleaner. Fire is an excellent hair removal agent, and no eyebrows is not cool...

After you have done the simple visual checks and the check for vacuum leak on the underside of the intake manifold, consider doing a smoke test.
Some of the guys here have built smoke machines used to find automotive vacuum leaks. They seem to work quite well and are made mostly with parts you would have laying around in your garage. Check out smoke machine vacuum leak - YouTube and see if there is one that you could build.

The vacuum line plumbing is old and brittle on many of these cars, so replacing the lines with new hose is a good plan. The common 1/8” and ¼” vacuum hose works well and isn’t expensive.

The PCV grommet and the power brake booster check valve grommet are two places that often get overlooked when checking for vacuum leaks. The rubber grommets get hard and lose their ability to seal properly. The PVC grommet is difficult to see if it is correctly seated and fitting snugly.

Fuel injector O rings can get old and hard. When they do, they are prone to leaking once the engine warms up. This can be difficult to troubleshoot, since it is almost impossible to get to the injectors to squirt oil into the fuel injector mounting bosses. If the plastic caps on the fuel injectors (pintle caps) are missing, the O rings will slide off the injectors and fall into the intake manifold.

Fuel injector seal kits with 2 O rings and a pintle cap (Borg-Warner P/N 274081) are available at Pep Boys auto parts. Cost is about $3-$4 per kit. The following are listed at the Borg-Warner site ( http://www.borg-warner.com ) as being resellers of Borg-Warner parts:
http://www.partsplus.com/ or http://www.autovalue.com/ or http://www.pepboys.com/ or http://www.federatedautoparts.com/

Most of the links above have store locators for find a store in your area.

Use motor oil on the O rings when you re-assemble them & everything will slide into place. The gasoline will wash away any excess oil that gets in the wrong places and it will burn up in the combustion chamber. Heat the pintle caps in boiling water to soften them to make them easier to install.



Diagram courtesy of Tmoss & Stang&2birds
mustangFoxFordVacuumDiagram.jpg



Vacuum leak due to slipped lower intake manifold gasket...

Ask Nicoleb3x3 about the intake gasket that slipped out of place and caused idle and vacuum leak problems that could not be seen or found by external examination. I don't care what you spray with, you won't find the leak when it is sucking air from the lifter valley. It simply isn't possible to spray anything in there with the lower manifold bolted in place.

photodisplay.php?iid=111113.jpg




Determining if you have a leak due to a slipped intake gasket as shown above. This test is only good if you can get the engine to run somewhere in the 1000-1700 RPM range
If your valve cover oil filler & PVC systems are still in the original configuration, try this:
Cap or plug the hose from the intake manifold to the PVC valve with a bolt.
Cap or plug the PVC valve with a piece of hose with a plug or bolt in it.
At that point the only vent for the crankcase is the tube from the oil filler neck to the throttle body.

Disconnect the tube that runs from the oil filler neck to the throttle body. Make sure the oil filler cap is on securely. Start the engine and put your thumb over the end of the tube that comes from the oil filler cap. If you feel suction, there is a leak. Another thing to do is to extend the tubing from the filler neck so that there is enough to stick the end in a jar or cup filled with motor oil. If it sucks up the oil, you definitely have a leak at the underside of intake manifold.

This isn't necessarily the definitive test, but it is the best thing I could come up with on short notice. If there is a lot of blowby, this obviously won't be of much help.

See the picture below to see the breather tube where in connects to the throttle body. It is close to the TPS and runs over the top of the IAC.

The following are diagrams courtesy of Tmoss & Stang&2birds

TPS_IAB_Pic.jpg


See the following website for some help from Tmoss (diagram designer) & Stang&2Birds (website host) for help on 88-95 wiring http://www.veryuseful.com/mustang/tech/engine/ Everyone should bookmark this site.

Ignition switch wiring
http://www.veryuseful.com/mustang/tech/engine/images/IgnitionSwitchWiring.gif

Fuel, alternator, A/C and ignition wiring
http://www.veryuseful.com/mustang/tech/engine/images/fuel-alt-links-ign-ac.gif

Complete computer, actuator & sensor wiring diagram for 88-91 Mass Air Mustangs
http://www.veryuseful.com/mustang/tech/engine/images/88-91_5.0_EEC_Wiring_Diagram.gif

Vacuum diagram 89-93 Mustangs
http://www.veryuseful.com/mustang/tech/engine/images/mustangFoxFordVacuumDiagram.jpg

HVAC vacuum diagram
http://www.veryuseful.com/mustang/tech/engine/images/Mustang_AC_heat_vacuum_controls.gif

TFI module differences & pinout
http://www.veryuseful.com/mustang/tech/engine/images/TFI_5.0_comparison.gif

Fuse box layout
http://www.veryuseful.com/mustang/tech/engine/images/MustangFuseBox.gif
 
  • Like
Reactions: imp

jas280z

Active User
Aug 18, 2017
13
0
1
32
A couple questions
Stock egr plate?
If so, are the water nipples open or plugged off.
What I'm thinking here is you may be getting air leaking in from your egr valve area, either the valve itself or the plate it bolts to, the plate is known to deteriorate inside having had water passing through them and causing electrolysis eating away at the aluminum and/or your egr valve is leaking and as the temps rise leak stops,
Also how far away from the manifolds/headers are your O2's?
Just thinking out loud, my thoughts are like cartoons in my head and are not based on fact or personal experience.

Stock EGR plate, coolant lines still connected. Good thought, I will look into it.

I believe the O2s are roughly stock distance away, there are no extenders on the O2 sensor harness. Hard to tell on a swap though. Exhaust manifold appear stock, no long tubes headers or anything.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
Jrichker, thanks for the suggestions.

Cylinder balance test showed a code 90 when I did it, indicating all good.

I have looked for vacuum leaks, but haven't done the test to look for a leak into the lifter valley. I had that issue on my 351C, that was a real pain. I will need to find a helper to do that test. So far my vacuum leak testing has shown no issues.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
I just pulled the car into the driveway, and let it idle up until the mixture got rich.

It took about 3 and a half minutes for the O3 reading to go from 0.02-0.07 volts to 0.8+ volts.

The switch took about 1 second when it happened.

The water temp was ~170F.

When the car was first started, cold, the vacuum was ~18.5 inhg. As the car warmed up, the idle got rough and the vacuum starting dipping down towards 16.5 inhg. When the mixture got richer, vacuum shot right up to almost 19 inhg.

Fuel pressure appeared to be a constant 38 psi, on my gauge.

I am going to repeat this test several times focusing, watching fuel pressure, vacuum, and rpm depending on the test.

I hope to run a repeatable test and document exactly what it happening up until the mixture gets rich.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
I took some videos of vacuum, right hand O2 output, fuel pressure and RPM during the engine start up lean period. From what I have seen, the engine always idles roughly and the O2 reports a lean condition for 2-4 minutes after starting the engine. It doesn't matter if the engine is hot or cold, it always has this issue for a few minutes after start up. If you drive during this time, the engine will backfire under light throttle.

This is not JUST a heat related issue. If you get the motor warmed up, and to the point where the right hand O2 is reporting rich and turn the motor off and immediately back on, it will still take around 2 minutes for the idle to go smooth and the mixture to not be very lean. If this was being caused by something leaking vacuum when cold, it should run well after a hot start.

I am thinking this may be something to do with the ECU running in open loop. What I found is that idle starts at a good point RIGHT after start up (18.5ish inhg), but falls off slowly to 15 inhg during the warm up period. As the vacuum decreases, idle quality drops and the engine shakes. Idle speed also decreases ~150 rpm during this period.

After 2-4 minutes, something changes. Over the course of ~2 to 5 seconds vacuum increases from 15 inhg to 18.5 inhg, idle speed increases ~150 rpm, O2 reading goes from ~0.02-0.05 V (very lean) to 0.8 to 0.9 V (rich). Fuel pressure stays between 36 and 38 PSI on my gauge during this entire process, I suspect any small changes in fuel pressure are due to vacuum to the FPR fluctuating. I am thinking this relatively sudden change may be the ECU switching to closed loop.

So... is there something that can cause the engine to go lean during open loop? One thought I had is IAC may be malfunctioning in open loop and allowing too much unmetered air past the throttle body. I would imagine that could cause a drop in vacuum and a lean condition. When the ECU switches to closed loop it switches the IAC setting and shuts the vacuum leak. Does this seem plausible to anyone?


Below are videos showing various gauges during the lean warm-up period up to the switch to rich. All videos were shot with the engine around operating temperature. I did not touch the accelerator at all during any of these tests.

View: https://youtu.be/XCZBBgdyXew

The increase is vacuum here corresponds with the O2 sensor switching from lean to rich


View: https://youtu.be/9Cx3P9zpJ2s

This video shows O2 output and fuel pressure.


View: https://youtu.be/D9gTKgXYQ9g

This video shows O2 output and RPM


I am going to try to consolidate all of the gauges into one area, so I can take a video showing everything at once. I think I will also try to get a video monitoring left hand O2 output, or both if possible.

Thanks for any help.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
So, both O2s do appear to be working correctly. At this point, I trust the voltage output I am seeing from them.


After making any significant change, I pull the battery for 20+ minutes to clear any codes. I understand this can clear the ECUs learning as well, as far as learned fuel trims.


I took another video. This is a startup/warmup after sitting for several hours. The engine was not cold, but it was not at operating temperature when starting.


I hooked leads to both O2 sensor pins, and I switch which channel I am looking at several times throughout the video.


Also, timing has been verified at 12 degrees btdc at operating temp and idle.



This is what appears to be happening:


Both O2s take about a minute to warm up. Showing a lean voltage output.


The car idles roughly in open loop.


The car is rev'd several times. It goes rich when revving, but returns to lean once the engine settles. Both O2 sensors show this.


At 2:56, while still in open loop, the left hand O2 shows that the left hand bank is running around stoich. It shows proper switching between rich and lean. The right hand O2 shows that bank is still running lean.


At 4:40, the car likely goes into closed loop. Right hand O2 voltage shows the right hand bank go from lean to rich. There is no switching, once it goes rich it stays rich (at idle). Idle quality improves, and vacuum increases. The left hand bank appears to still be running at stoich.


It took 4:20 for the car to switch to closed loop, assuming the switch to rich on RH is happening when the ECU switches to closed.



View: https://youtu.be/vyJ7i8dfXCA




So, it looks like the left hand bank is running correctly. Whatever is happening APPEARS to only be happening to the right hand bank. If it somehow is a vacuum leak, it would either be an intake runner on the left hand side, or a fuel injector. I am not thinking vacuum leak though, since there is enough fuel delivery to make the right hand rich during closed loop. Since there is enough fuel to go rich, I would expect the computer would be able to target a stoich mixture rather than just going rich.


I am thinking that the computer may be giving incorrect fuel trims to the right hand bank. Does the 5.0 HO computer give unique fuel signals to each bank, or does it give the same signal to each bank? I’m still thinking the issue may lie in the ECU. Looks like I either need a way to read the fuel trims, or find another ECU to test with.
 

jrichker

StangNet's favorite TOOL
SN Certified Technician
Mar 10, 2000
27,374
2,707
234
75
Dublin GA
lowendmac.com
Look for problems in the O2 sensor wiring if you have not already done so.

Check the O2 sensor heater ground, make sure that it is good for both O2 sensors


Code 41 or 91. Or 43 Three digit code 172 or 176 - O2 sensor indicates system lean. Look for a vacuum leak or failing O2 sensor.

Revised 20-July-2017 to add note that the 94-95 uses a 4 wire O2 sensor.

Code 41 is the passenger side sensor, as viewed from the driver's seat.
Code 91 is the driver side sensor, as viewed from the driver's seat.

Code 172 is the passenger side sensor as viewed from the driver's seat.
Code 176 is the driver side sensor, as viewed from the driver's seat.

Code 43 is not side specific according to the Probst Ford Fuel injection book.

The computer sees a lean mixture signal coming from the O2 sensors and tries to compensate by adding more fuel. Many times the end result is an engine that runs pig rich and stinks of unburned fuel.

The following is a Quote from Charles O. Probst, Ford fuel Injection & Electronic Engine control:
"When the mixture is lean, the exhaust gas has oxygen, about the same amount as the ambient air. So the sensor will generate less than 400 Millivolts. Remember lean = less voltage.

When the mixture is rich, there's less oxygen in the exhaust than in the ambient air , so voltage is generated between the two sides of the tip. The voltage is greater than 600 millivolts. Remember rich = more voltage.

Here's a tip: the newer the sensor, the more the voltage changes, swinging from as low as 0.1 volt to as much as 0.9 volt. As an oxygen sensor ages, the voltage changes get smaller and slower - the voltage change lags behind the change in exhaust gas oxygen.

Because the oxygen sensor generates its own voltage, never apply voltage and never measure resistance of the sensor circuit. To measure voltage signals, use an analog voltmeter with a high input impedance, at least 10 megohms. Remember, a digital voltmeter will average a changing voltage." End Quote

Testing the O2 sensors 87-93 5.0 Mustangs
Measuring the O2 sensor voltage at the computer will give you a good idea of how well they are working. You'll have to pull the passenger side kick panel off to gain access to the computer connector. Remove the plastic wiring cover to get to the back side of the wiring. Use a safety pin or paper clip to probe the connections from the rear.

Disconnect the O2 sensor from the harness and use the body side O2 sensor harness as the starting point for testing. Do not measure the resistance of the O2 sensor , you may damage it. Resistance measurements for the O2 sensor harness are made with one meter lead on the O2 sensor harness and the other meter lead on the computer wire or pin for the O2 sensor.

Computer wiring harness connector, computer side.
88243


Backside view of the computer wiring connector:
a9x-series-computer-connector-wire-side-view-gif.71316.gif


87-90 5.0 Mustangs:
Computer pin 43 Dark blue/Lt green – LH O2 sensor
Computer pin 29 Dark Green/Pink – RH O2 sensor
The computer pins are 29 (L\RH O2 with a dark green/pink wire) and 43 (LH O2 with a dark blue/pink wire). Use the ground next to the computer to ground the voltmeter. The O2 sensor voltage should switch between .2-.9 volt at idle.

91-93 5.0 Mustangs:
Computer pin 43 Red/Black – LH O2 sensor
Computer pin 29 Gray/Lt blue – RH O2 sensor
The computer pins are 29 (LH O2 with a Gray/Lt blue wire) and 43 (RH O2 with a Red/Black wire). Use the ground next to the computer to ground the voltmeter. The O2 sensor voltage should switch between .2-.9 volt at idle.


Testing the O2 sensors 94-95 5.0 Mustangs; note that the 94-95 uses a 4 wire O2 sensor.
Measuring the O2 sensor voltage at the computer will give you a good idea of how well they are working. You'll have to pull the passenger side kick panel off to gain access to the computer connector. Remove the plastic wiring cover to get to the back side of the wiring. Use a safety pin or paper clip to probe the connections from the rear. The computer pins are 29 (LH O2 with a red/black wire) and 27 (RH O2 with a gray/lt blue wire). Use pin 32 (gray/red wire) to ground the voltmeter. The O2 sensor voltage should switch between .2-.9 volt at idle.


Note that all resistance tests must be done with power off. Measuring resistance with a circuit powered on will give false readings and possibly damage the meter. Do not attempt to measure the resistance of the O2 sensors, it may damage them.

Testing the O2 sensor wiring harness
Most of the common multimeters have a resistance scale. Be sure the O2 sensors are disconnected and measure the resistance from the O2 sensor body harness to the pins on the computer. Using the Low Ohms range (usually 200 Ohms) you should see less than 1.5 Ohms.

87-90 5.0 Mustangs:
Computer pin 43 Dark blue/Lt green – LH O2 sensor
Computer pin 29 Dark Green/Pink – RH O2 sensor
Disconnect the connector from the O2 sensor and measure the resistance:
From the Dark blue/Lt green wire in the LH O2 sensor harness and the Dark blue/Lt green wire on the computer pin 43
From the Dark Green/Pink wire on the RH Os sensor harness and the Dark Green/Pink wire on the computer pin 29

91-93 5.0 Mustangs:
Computer pin 43 Red/Black – LH O2 sensor
Computer pin 29 Gray/Lt blue – RH O2 sensor
Disconnect the connector from the O2 sensor and measure the resistance:
From the Red/Black wire in the LH O2 sensor harness and the Red/Black wire on the computer pin 43
From the Dark Green/Pink Gray/Lt blue wire on the RH Os sensor harness and the Gray/Lt blue wire on the computer pin 29

94-95 5.0 Mustangs:
Computer pin 29 Red/Black – LH O2 sensor
Computer pin 27 Gray/Lt blue – RH O2 sensor
From the Red/Black wire in the LH O2 sensor harness and the Red/Black wire on the computer pin 29
From the Dark Green/Pink Gray/Lt blue wire on the RH Os sensor harness and the Gray/Lt blue wire on the computer pin 27

There is a connector between the body harness and the O2 sensor harness. Make sure the connectors are mated together, the contacts and wiring are not damaged and the contacts are clean and not coated with oil.

The O2 sensor ground (orange wire with a ring terminal on it) is in the wiring harness for the fuel injection wiring. I grounded mine to one of the intake manifold bolts

Check the fuel pressure – the fuel pressure is 37-41 PSI with the vacuum disconnected and the engine idling. Fuel pressure out of range can cause the 41 & 91 codes together. It will not cause a single code, only both codes together.

Make sure you have the proper 3 wire O2 sensors. Only the 4 cylinder cars used a 4 wire sensor, which is not compatible with the V8 wiring harness. The exception is that the 94-95 uses a 4 wire O2 sensor.

Replace the O2 sensors in pairs if replacement is indicated. If one is weak or bad, the other one probably isn't far behind.

Code 41 can also be due to carbon plugging the driver’s side Thermactor air crossover tube on the back of the engine. The tube fills up with carbon and does not pass air to the driver’s side head ports. This puts an excess amount of air in the passenger side exhaust and can set the code 41. Remove the tube and clean it out so that both sides get good airflow: this may be more difficult than it sounds. You need something like a mini rotor-rooter to do the job because of the curves in the tube. Something like the outer spiral jacket of a flexible push-pull cable may be the thing that does the trick.

If you get only code 41 and have changed the sensor, look for vacuum leaks. This is especially true if you are having idle problems. The small plastic tubing is very brittle after many years of the heating it receives. Replace the tubing and check the PVC and the hoses connected to it.
 

jas280z

Active User
Aug 18, 2017
13
0
1
32
I have checked the O2 wiring, everything looks good and has continuity from one end to the other. Both O2 sensors appear to be sending the correct voltage to their respective pins at the ECU.

There are only three vacuum lines left on the car, FPR, PCV, and my vacuum gauge. None are leaking from what I can tell. If any were leaking, I would expect them to affect both banks of the engine. This issue appears to only affect the right hand bank.
 

Mustang5L5

i'm familiar with penetration
Mod Dude
Feb 18, 2001
35,081
11,210
224
Massachusetts
This is a mass air equipped car correct?

170 degrees when fully warmed up for a EFi 302 is too cold. You really want it in the 185-195 range.

I know you are chasing O2 sensors right now, but another cause for o2 sensors reading abnormally and tripping codes is a vacuum leak somewhere that is after the meter. Common cause is a slipped gasket on the lower intake manifold to block that won't register as an internal vac leak. Typically 9 times out of 10, when I see a lean backfire with no obvious signs as to why, it's usually been a slipped lower intake manifold gasket.
 
  • Like
Reactions: General karthief

jrichker

StangNet's favorite TOOL
SN Certified Technician
Mar 10, 2000
27,374
2,707
234
75
Dublin GA
lowendmac.com
This is a mass air equipped car correct?

170 degrees when fully warmed up for a EFi 302 is too cold. You really want it in the 185-195 range.

I know you are chasing O2 sensors right now, but another cause for o2 sensors reading abnormally and tripping codes is a vacuum leak somewhere that is after the meter. Common cause is a slipped gasket on the lower intake manifold to block that won't register as an internal vac leak. Typically 9 times out of 10, when I see a lean backfire with no obvious signs as to why, it's usually been a slipped lower intake manifold gasket.
Just in case you missed this, there is a test for internal vacuum leak due to a slipped lower intake manifold gasket...

Determining if you have a leak due to a slipped intake gasket as shown above. This test is only good if you can get the engine to run somewhere in the 1000-1700 RPM range
If your valve cover oil filler & PVC systems are still in the original configuration, try this:
Cap or plug the hose from the intake manifold to the PVC valve with a bolt.
Cap or plug the PVC valve with a piece of hose with a plug or bolt in it.
At that point the only vent for the crankcase is the tube from the oil filler neck to the throttle body.

Disconnect the tube that runs from the oil filler neck to the throttle body. Make sure the oil filler cap is on securely. Start the engine and put your thumb over the end of the tube that comes from the oil filler cap. If you feel suction, there is a leak. Another thing to do is to extend the tubing from the filler neck so that there is enough to stick the end in a jar or cup filled with motor oil. If it sucks up the oil, you definitely have a leak at the underside of intake manifold.

This isn't necessarily the definitive test, but it is the best thing I could come up with on short notice. If there is a lot of blowby, this obviously won't be of much help.

See the picture to see the breather tube where in connects to the throttle body. It is close to the TPS and runs over the top of the IAC.
 
  • Like
Reactions: Mustang5L5

jas280z

Active User
Aug 18, 2017
13
0
1
32
The car is mass air equipped.

At this point, I believe the O2 sensors are working correctly.


If a vacuum leak were the cause, why would it be rich in open loop? Not saying it can't be, it just wouldn't be what I would suspect.


At this point, I am thinking the reman ECU that is in the car is not correct. From what I have seen, if a manual A9L computer were used in the car with the auto O2 harness, it can cause some of the issues I have seen. Pins 30 and 29 are connected through a resistor in the ECU. If they put 12VDC into a manual computer expecting 5VDC at pin 30, it could have blown out some of the components on pin 29.

http://forums.corral.net/forums/eec-tech/540841-what-computer-use.html

"E9ZF-12A650-AA (sometimes A2A), catch code A9L, derived from 89-92 Manual transmission Mustang. If you use this processor, be absolutely certain that you also use an O2 sensor wiring harness from manual trans vehicle or you will likely damage the processor. This has to do with the neutral sensing circuit at pin 30 of the processor. Manual processors utilize a 5 volt signal return, while Auto processors utilize a 12 Volt signal. If you feed 12 volts to the manual processor that is designed to work in conjunction with a 5 volt signal, you will usually lose the KAM and the computer may nat communicate with a scanner. The computer will still control the engine, but, you will generally lose the adaptive fuel strategy. This signal is received through a looped wire within the O2 sensor harness. If you cannot obtain the proper harness, get a schematic and move the proper wire. Depending on the year of the harness, it is generally a Purple wire with a Yellow stripe, or a Light Blue Wire with a Yellow stripe."

This would explain a lot of my symptoms. I have had a hard time getting the code reader to get the car into test mode at times as well.

The ECU in the car is NOT original, it is pretty clearly a reman. I cannot tell if it was an A9L or A9P computer.

llEDouj.jpg


ttyG1mf.jpg
 

jrichker

StangNet's favorite TOOL
SN Certified Technician
Mar 10, 2000
27,374
2,707
234
75
Dublin GA
lowendmac.com
The car is mass air equipped.

At this point, I believe the O2 sensors are working correctly.


If a vacuum leak were the cause, why would it be rich in open loop? Not saying it can't be, it just wouldn't be what I would suspect.


At this point, I am thinking the reman ECU that is in the car is not correct. From what I have seen, if a manual A9L computer were used in the car with the auto O2 harness, it can cause some of the issues I have seen. Pins 30 and 29 are connected through a resistor in the ECU. If they put 12VDC into a manual computer expecting 5VDC at pin 30, it could have blown out some of the components on pin 29.

http://forums.corral.net/forums/eec-tech/540841-what-computer-use.html

"E9ZF-12A650-AA (sometimes A2A), catch code A9L, derived from 89-92 Manual transmission Mustang. If you use this processor, be absolutely certain that you also use an O2 sensor wiring harness from manual trans vehicle or you will likely damage the processor. This has to do with the neutral sensing circuit at pin 30 of the processor. Manual processors utilize a 5 volt signal return, while Auto processors utilize a 12 Volt signal. If you feed 12 volts to the manual processor that is designed to work in conjunction with a 5 volt signal, you will usually lose the KAM and the computer may nat communicate with a scanner. The computer will still control the engine, but, you will generally lose the adaptive fuel strategy. This signal is received through a looped wire within the O2 sensor harness. If you cannot obtain the proper harness, get a schematic and move the proper wire. Depending on the year of the harness, it is generally a Purple wire with a Yellow stripe, or a Light Blue Wire with a Yellow stripe."

This would explain a lot of my symptoms. I have had a hard time getting the code reader to get the car into test mode at times as well.

The ECU in the car is NOT original, it is pretty clearly a reman. I cannot tell if it was an A9L or A9P computer.

llEDouj.jpg


ttyG1mf.jpg
The O2 sensor mismatch will cause a failure of pin 46 with is the signal ground for all the engine mounted sensors. The tip is that you cannot get the computer to go into diagnostic mode, and it will not dump codes.

If I recall correctly you have successfully dumped the computer codes with the computer that you are currently using.



If that is not correct or you desire to look further into the computer, here is some help.


Computer will not go into diagnostic mode on 86-90 model 5.0 Mustangs

Disconnect the battery positive terminal before making any resistance checks.
The voltage drop in the ground cable will cause incorrect resistance readings.


How it is supposed to work:
The black/white wire (pin 46) is signal ground for the computer. It provides a dedicated ground for the EGR, Baro, ACT, ECT, & TPS sensors as well as the ground to put the computer into self test mode. If this ground is bad, none of the sensors mentioned will work properly. That will severely affect the car's performance. You will have hard starting, low power and drivability problems. Since it is a dedicated ground, it passes through the computer on its way to the computer main power ground that terminates at the battery pigtail ground. It should read less than 1.5 ohms when measured from any of the signal ground connections to sensors on the engine harness. Use the battery pigtail ground as the other reference point for the ohmmeter probe.

What sometimes happens is that the test connector black/white wire gets jumpered to power which either burns up the wiring or burns the trace off the pc board inside the computer. That trace connects pins 46 to pins 40 & 60.

The STI (Self Test Input ) is jumpered to ground to put the computer into test mode. Jumpering it to power can produce unknown results, including damage to the computer. The ohm test simply verifies that there are no breaks in the wiring between the test connector and the computer input.

How to test the wiring :
With the power off, measure the resistance between the computer test ground (black/white wire) on the self test connector and battery ground. You should see less than 1.5 ohms.

attachment.php?attachmentid=58312&stc=1&d=1242744354.gif


If that check fails, remove the passenger side kick panel and disconnect the computer connector. There is a 10 MM bolt that holds it in place. Measure the resistance between the black/white wire and pin 46 on the computer wiring connector: it should be less than 1.5 ohms. More that 1.5 ohms is a wiring problem. If it reads 1.5 ohms or less, then the computer is suspect. On the computer, measure the resistance between pin 46 and pins 40 & 60: it should be less than 1.5 ohms. More that that and the computer’s internal ground has failed, and the computer needs to be repaired or replaced.

See http://www.stangnet.com/mustang-forums/749974-computer-issue.html#post7490537 for Joel5.0’s fix for the computer internal signal ground.

If the first ground check was good, there are other wires to check. Measure the resistance between the STI computer self test connector (red/white wire) and pin 48 on the computer main connector: it should be less than 1.5 ohms. More that 1.5 ohms is a wiring problem

The following is a view from the computer side of the computer wiring connector: it is for an A9L, A9P computer.
eec-iv-computer-connector-for-5-0-mustang-gif.88243.gif


a9x-series-computer-connector-wire-side-view-gif.71316.gif


Diagram courtesy of Tmoss & Stang&2birds

Check out the diagram and notice all the places the black/white wire goes. Almost every sensor on the engine except the MAF is connected to it.

88-91_5.0_EEC_Wiring_Diagram.gif


See the following website for some help from Tmoss (diagram designer) & Stang&2Birds
(website host) for help on 88-95 wiring http://www.veryuseful.com/mustang/tech/engine

See the graphic for the 10 pin connector circuit layout.
?temp_hash=3ef2497fff29a7a9daee955cf93e5805.jpg