EFI, Why AFR different Left to Right?
 

I am in the shakedown stage of my EFI conversion and have much higher AFRs coming from my right bank opposed to my left bank...

Do I have something wrong? Bad injectors, timing, fuel pressure, spark?

Engine starts instantly, idles perfectly, revs smooth and builds boost. I have verified timing with timing light, fuel pressure is steady, no vacuum leaks...

Parameters: 3.4, ported heads, Carrera intake, 964 cams, 75mm TB, twin 044 fuel pumps, -10 fuel lines/rails, 30 PSI fuel pressure, twin plug, dual EDIS, MSII, 72lb injectors, k14/24 zero clearance turbos...

Is it possible there may be a leak in the tail pipe drawing in extra O2? I just finished a similar build using MSII with extra code. twin Turbonetics t3 t4s. http://forums.pelicanparts.com/uploa...1284347014.jpg
http://forums.pelicanparts.com/uploa...1284347119.jpg

Are you certain your wastegates are synchronized? Maybe you got one turbo working over the other. Are you taking the wastegates boost reference from the same location? I would guess your fuel rails are identical and run out of a single fuel pressure regulator with similar feed to the rails? Otherwise this could be another problem. Have you swapped the sensors? Maybe they are not calibrated equally. Could be so many things....Hope you get to the end of this soon.

Turbo Nut

Nice build. What coil over plugs are you using?

Thanks

The latest VW 1.8T coils and 964 covers with the 6mm bolt holes drilled to 8mm coils fit like glove top and bottom. The coils are like $12.00 each with the good buddy deal from the dealer.

Are your WBO2 calibrated? Are they grounded at the same place? How different are the readings? Are you logging the two WBO2 readings in Megatune or TS?

I had a similar problem with mine. I only have one Lambda sensor so I didn't see it there, but I did see differences in my left and right spark plugs. Turns out I was a little off on my cam timing left to right.

Thanks.

Where are the O2 sensors located (pre or post turbo, near tailpipe), what O2 controllers are you using, and are they both calibrated the same?

Not to hijack the thread but very nice. How are you retaining the lower coils and how are you splitting the +5V logic signal to the coils?

Also, have you verified that each of your injectors are flowing the same? Have you checked for contact resistance at the injectors?

How did you wire your injector banks? Did you do a L bank and a R bank or did you put 1-6-2 and 4-3-5 together to match the crank rotation?

I just wonder, are your fuel rails in series? If so might check pressure at the end of each rail at WOT for balance if possible.

Are you using one fuel pressure reg or one at the end of each rail? If two, might want to test that they are balanced on boost.

Looks like a single TB so probably not a wast gate balance thing unless off enough to increase back pressure / reversion on that side.

Good luck, crazy wild build.

So we dont hijack the thread PM me and I will give you the rundown.

Hi David,

First, what an absolutely great build! I've got some questions but I'll send them outside of this thread. Regarding the differences in O2 between the banks, here are my guesses:

1. O2 sensor or controller? If you are using two then perhaps you could swap to see if it is a sensor calibration issue.

2. Stuck fuel injector? Check your plugs to see if a cylinder looks lean or pull the lean bank and flow test them.

3. Difference in airflow in the intakes? I've heard that flow matching is a good idea if you're going to go for big hp. Problem if you're tuning a cylinder lean.

Again, looks terrific.
Mike

First of all, I want to thank each of you for your input! Chasing these issues can be frustrating, but it helps knowing that such a wealth of expertise exists here at PP. With the amount of modifications done to the car/engine, I went into this expecting to chase issues ranging from fuel/oil leaks to wiring to sensor calibrations to vacuum leaks, etc. I was however amazed that when I turned the key for the first time to check the fuel system for leaks, the engine fired. I download a base tune provided by Michael Pickett, tweaked a few settings and had it idling in about 15 minutes. Now to work out the kinks…

Turbonut: I do have leaks between the cats and the muffler. The muffler is just barely slip fitted to the cats at this time and will be finish welded to align properly with the bumper cover. I hadn’t considered this, and will remove the muffler while tuning (just gets a little loud). BTW, very nice setup…if doing it again, I would have went COP...have always hated all the plug wires and spent a lot of time routing them away from heat and visually out of sight.

IMONBOOST: The turbos have internal wastegates and I believe that they are set the same. Both are getting vacuum from the same source. Even off of boost, I see inconsistencies from left to right. The fuel rails are identical, with equal length feed lines and exit through a single fuel pressure regulator. I did swap the sensors left to right and saw consistent results

Jpnovak: I did perform a free air calibration, but the controllers are not grounded to the same place, hadn’t considered that. When seeing 14 on one side, I will see anywhere from 11-12 on the other. I am using TunerStudio but am only logging one WB but monitoring with 2 gauges. I have switched the input to MS between the two controllers and confirmed that the gauges match what MS is receiving.

David: I was hoping that the cam timing wasn’t the issue, but will add it to the list to verify left to right. Due to the packaging constraints, pulling plugs is a real pain…I was hoping that I would not need to so soon, but am thinking this is the only way to see which plugs are getting too much/little fuel.

Sjf911: O2 sensors are after the turbos and before the cats, both sides are in the same location. I am using two techedge controllers with NTK L1H1 sensors and believe they are calibrated similarly. If I switch the controllers and sensors left to right, I am seeing the same result with the right side being richer than the left.

Jpnovak: I have not verified that the injectors are flowing the same…how would I do this? When I purchased them, they were flow matched, but that was a couple of years ago and they have been on the shelf since them. I do have two sets of injectors and I would not rule out the possibility that I may have mixed them up. I checked the resistance at each injector when I built the harness, but it would be worth checking again. They are wired left bank, right bank.

911st: My fuel rails are in parallel, fuel filter runs to a fuel log splitting to each rail and then to a single FPR


Thoughts on diagnostic steps:

1. Remove muffler to eliminate possible issues with exhaust leaks
2. Ground both WBO2 to the same place (does this need to be the same ground point as the MS ECU)
3. Use a single O2 sensor and controller and swap side to side to get baseline readings and rule out differences between WB controllers/sensors
4. Pull plugs to determine if an over/under fueling condition exists in a specific cylinder (does it matter if I pull top or bottom or do I need to look at both?)
5. Pull injectors and verify each is firing and has a reasonable spray pattern
6. Verify consistent flow across injectors (how can I do this?)
7. Verify cam timing

Since pictures are worth a thousand words...



http://forums.pelicanparts.com/uploa...1284418715.jpg


http://forums.pelicanparts.com/uploa...1284418726.jpg


http://forums.pelicanparts.com/uploa...1284418737.jpg


http://forums.pelicanparts.com/uploa...1284418749.jpg


http://forums.pelicanparts.com/uploa...1284418761.jpg


http://forums.pelicanparts.com/uploa...1284418773.jpg


http://forums.pelicanparts.com/uploa...1284418782.jpg


http://forums.pelicanparts.com/uploa...1284418795.jpg


http://forums.pelicanparts.com/uploa...1284418803.jpg


http://forums.pelicanparts.com/uploa...1284418813.jpg

Beautiful install....

You could install a dual unit with fresh O2 sensors and do a careful free air calibration. I am bit biased for this unit since I have a single version of the this Wego IV instrument in my 1973 911 MFI and love it. Great instrument and great value for the money.

With this great motor of yours I would install a dual AFR permanently and have real time readings plus the bonus of 2 hours of data logging to look at on your laptop at any time.



Our host sells this one.....here is a photo:

http://forums.pelicanparts.com/uploa...1284419913.jpg

Send your injector to RC Engineering and have them cleaned and flow tested. They will then send you back the injector and will tell you the flow reading on each one. The best way would be to send them ALL the injectors you have. Its easier to match 6 out of 10 injectors that way you will get the closest matching ones to install on your beautiful motor

Megasquirt has an injector test mode if you want to test them yourself:

MegaSquirt-II Injector Test Mode

MyTurbo, I went round and round when I had the issue. First Simon at Motec said it sound's like the cams. I didn't believe it since I was sure I had them right so I completely replumbed the fuel system on the engine. There was no change. Then my tuner said it sounds like the cams. I didn't believe and thought it could be some other fueling issue so I tried adding more fuel to one side and taking fuel from the other with ECU programming. There was very little change.

So I pulled the engine and checked the cams. Sure enough they were off, probably more than I'd like to admit.

At least try to pull a plug from each side and see what they look like. All 6 left plugs looked almost identical and all 6 right did too, they just looked like they came from different engines.


BTW, the build looks awesome!!!!

Maybe switch the O2's and or the connections from side to side to see if the richness readings follow the sensors or leads.

What happens if there is some oil blow by on one side? Dose it effect the O2 readings?

Just thinking out loud. How much can playing with cam timing effect VE?

I am guessing it can move peak TQ up or down about 500rpm if moved to both extremes but I do not know this for sure. I would not think it would effect how much peak TQ would be made very much. Just where.

There is a thread here some where, where a PCA club racer re-timed his 3.2 normally aspirated cams and dynoed in search of more corner exit speed. It only effected peak HP about 3% in his case. However, Porsche re-timed the cams between 77&78 and Hp dropped from 200 to 180hp and HP peak moved from 6000 to 5500rpm which is 10% if that was the only change and the reported spec's are accurate. I am suspect as the reported TQ peak rpm point did not change between the years but peak TQ changed from 188 to 165# suggesting a change in VE's.

If it is the cams and the variance is significant then it seems there should be shift in which side is rich some where around the TQ peak area. And in the area where the TQ curve is mostly flat the AFR's might be ballanced.

Before going too far, I figured I should pull the plugs and see what they look like…I would say this is the biggest drawback of having the turbos on the side and the intercoolers pipes running up next to the valve covers. Fortunately, it’s as simple as removing three nuts and disconnecting 4 hoses.

Lower plugs 1, 2 and 3 have black, sooty dry-looking deposits; this is consistent with this bank showing a rich condition. Plug 4 is sooty with 5 and 6 showing black on only the porcelain with the strap and base ring mostly clean. Since they are not consistent across all 6, I am thinking this is a fueling issue, probably injectors. Thoughts?

How quickly do the plugs become sooty with a rich condition? Could this have happened from my initial map being overly rich before I made adjustments? Should I clean the plugs and check them again or move on to having the injectors checked and flow matched?

As an aside, what are your thoughts on the heat range of the plugs? The uppers are Denso PK20PR-P8 and the lowers are NGK PJR7A.


http://forums.pelicanparts.com/uploa...1284443333.jpg


http://forums.pelicanparts.com/uploa...1284443345.jpg


http://forums.pelicanparts.com/uploa...1284443354.jpg


http://forums.pelicanparts.com/uploa...1284443364.jpg

Those aren't nearly as obvious as what I had. Mine were literally black and white.

When I first started my engine after EFI, I put in plain jane NGK plugs knowing I'd probably throw a few sets away. Those look pretty expensive so you could try cleaning them and then a quick drive around the block before pulling them out.

Unfortunately I don't have 2 lambda sensors so I have no numbers for comparison.

I did go from a simple in series fuel system where fuel entered one rail, crossed to the other, and out the other rail. Even though I'd gotten the OK from a reputable tuner that this was OK, I thought it was the problem and converted to a system like your's. I put the rails in parallel with dampers on each rail. There was literally no change after all that work and money. So I wouldn't worry about the fuel delivery system.

If you see some inconsistancy between different cylinders on the same side, I guess it's possible to have just one or two bad injectors causing the problem. Could you switch injectors side to side very easily?

Again, what happens to O2 /AFR readings if the rings are not seated yet or some such and some oil gets into the combustion chamber?

Have you tryed moving the O2 sensors side to side.

Here is a picture guide to reading spark plugs: Spark Plugs


I am not an expert but it dose not look like there is a lean condition


Also could be an ignition issue and the plugs are not getting hot enough:

"Dry and Wet Fouling:
http://forums.pelicanparts.com/uploa...1284493316.jpg

Although there are many different cases, if the insulation resistance between the center electrode and the shell is over 10 ohms, the engine can be started normally. If the insulation resistance drops to 0 ohms, the firing end is fouled by either wet or dry carbon."

"As an aside, what are your thoughts on the heat range of the plugs? The uppers are Denso PK20PR-P8 and the lowers are NGK PJR7A."

I've been running the '7' range NGKs for 4 years with no fouling. They're on the cool side for my setup.

Today’s update…

Since my wideband controllers and sensors were suspect, I borrowed an LM2 and verified by swapping it left to right that the AFR readings are different. However, the readings were not completely consistent with my TechEdge. Looks like it is time to upgrade…I think I will order the Daytona dual channel that dickague recommended.

Since the lower plugs didn’t look great, I decided to pull the top plugs as well. Only problem is that the right rear plug is impossible to get out due to clearance issues with the power steering pump. As a result, a partial engine drop was in order to gain enough access to pull the final plug. Top plugs were consistent with lower plugs.

Next I decided to test the injectors using the Megasquirt injector test mode as suggested by Mike. As suspected, two of the injectors flowed much less than the other four. Since this was the side with the wideband output going to MS, looks like I was adjusting the fuel to compensate for the under flowing injectors and as a result was over fueling the other side. Turns out I have a total of 16 of these injectors (not sure why) so it is entirely possible that I inadvertently mixed them up. As suggested by esotoracing, I have packed the entire lot and will send them off to RC Engineering to have cleaned and flow matched.

In double checking the fuel system, it turns out that the fuel return line from the fuel pressure regulator back to the tank is the same -8an as the line going to the fuel rails and then to the FPR. Even though I laid it out to be smaller, looks like I messed up…while the injectors are being serviced, I will replace this with a -6an return line.

Since the engine is mostly out and I am in a holding pattern with the injectors, I will pull it completely and check the cam timing tomorrow. I will also double check the coil packs and the balance of the ignition system to ensure that all 12 plugs are getting good spark.

Can I disconnect the fuel system and turn the engine over without having it bolted to the car, or is this a bad idea?


http://forums.pelicanparts.com/uploa...1284516800.jpg


http://forums.pelicanparts.com/uploa...1284516810.jpg