911sc lean cylinders 1, 2 3

[L8apex32]

1981 911SC euro, stock engine, high miles, stock racing class use only

The engine is running lean in cylinders 1, 2 and 3. You can see the white deposit on the exhaust pipes and the spark plugs, left side only. My mechanic put it on the dyno to analyze air/fuel ratio and to diagnose. He found that moving the fuel lines - hard lines between the fuel distributor to the injectors - from cylinders (1, 2, 3) to (4, 5, 6) and vice-versa moved the lean condition from the 1, 2, 3 side to the 4, 5, 6 side of the engine. It has true dual exhaust, no cross over, so the analyzer sees one side or the other. That seems like a pretty good diagnosis. But, is there anything else it could be?

I have two further subjects to ask about. Before the most current diagnosis, my mechanic proposed that it may be brake related in that the vacuum line to the brake booster comes from the left side of the engine. Could a leak in the vacuum line to the brake booster make the 1, 2, 3 cylinders go lean and not 4, 5 and 6? (I did feel my brake feel rock hard during a double yellow, which indicates a loss of vacuum to the brake booster.)

Finally, it seems that all 3 cylinders are lean on the left side. Is there anything in the fuel distributor that would affect 3 cylinders at once on the same side?

Similarly, is there anything in the air intake path that could mess with cylinders 1, 2 and 3 but not 4, 5 and 6?

Thanks.

[psalt]

Tell your mechanic to do the Injected Quantity Comparison Test outlined in the Bosch manual, if he hasn't done it already. This will isolate if it is fuel related. One possibility is clogged hard lines, this has fooled many, sometimes they are impossible to clean and need to be replaced.

[stlrj]

Bad injectors are not that uncommon. Replace the three lean ones and go from there.

[T77911S]

if the injectors where not moved, then it is not the injectors. i assume the lines where swapped AT the FD side only.

the 81 has adjustable fuel flow for each port on the FD. either the FD itself is bad/dirty or someone has messed with the adjustments on the FD or they where never right to begin with.

i checked the fuel volume out of my FD by placing the injectors in 6 containers. with the FP on, i raised the sensor plate all the way up to fill the containers. then i each container into a common container, one at a time to eliminate the contaier weight, then weighed the fuel on a postal scale. i also did the test with the just the fuel lines and no injectors to take them out of the equation. i did this several times with different heights of the sensor plate.

i had one injector line (it was a bad line) that was so bad i did not have to weigh it see that it was bad. the line was blocked only allowing about half the fuel as the others.

there was a guy on here, alan sonething, that went through this with his 930 FD.

[L8apex32]

Thanks for the advice. I hope to report what we find.

[L8apex32]

I have acquired the fuel pressure gauge so I can take measurements. The system pressure and control pressure are good.

While taking these readings, the gauge fluctuated a little, from 67 psi to 69psi. It bounced around. The target range is 65-75psi so I'm in the range - is a little fluctuation in the pressure ok?

Now its off to do the Injected Quantity Comparison Test.

[boyt911sc]

Originally Posted by L8apex32

I have acquired the fuel pressure gauge so I can take measurements. The system pressure and control pressure are good.

While taking these readings, the gauge fluctuated a little, from 67 psi to 69psi. It bounced around. The target range is 65-75psi so I'm in the range - is a little fluctuation in the pressure ok?

Now its off to do the Injected Quantity Comparison Test.

L8apex,

Do you mind sharing the values of your fuel pressure measurements? If they are good, then you should have numbers for them. Thanks.

Tony

[L8apex32]

The system pressure is 66 to 67 psi. Specified range is 65-75psi.

The control pressure is 2.0 bar/30psi. This is at 70degF with the wires connected. 70 deg F = 21.1deg C and the specified range for this temperature, cold, is 2.05 - 2.45bar. I'm just below that range. However, I'm also in my shaded garage and the engine may be cooler than the ambient air.

I have also done an injector quantity comparison test. Did all 6 at once. I took the hard lines off the injectors and collected for a period of time. The results per cylinder are:
1: 102g
2: 101g
3: 102g
4: 100g
5: 100g
6: 102g

[psalt]

Are you saying you removed the injectors for this test ?

[Grady Clay]

Brian,

Your measurements:
1: 102g
2: 101g
3: 102g
4: 100g
5: 100g
6: 102g
Mean of 101.17 g, standard deviation of 0.98 g.

Considering cylinders 1-2-3 only:
Mean of 101.67 g, standard deviation of 0.58 g.
Considering cylinders 4-5-6 only:
Mean of 100.67 g, standard deviation of 1.15 g.

Compared to the mean, the 1-2-3 side is +0.49% of the mean.
Compared to the mean, the 4-5-6 side is -0.49% of the mean.
The ‘worst case’ cylinders (1,2,5) are -1.16% of the mean.
“+” is ‘rich’ and “-“ is ‘lean’.

The question is:
Is a L-R 1% difference in fuel flow significant for a visible difference in combustion mixture?


I would like to see more data, greater sample size (1 kg, ~1.3 l) and greater precision (measure to 0.1 g).
A major consideration when measuring gasoline is evaporation during the experiment.




Working with gasoline is VERY DANGEROUS!
Take every possible precaution.

REPEAT

Working with gasoline is VERY DANGEROUS!
Take every possible precaution.
Do this outside, away from structures.
Wear your Nomex. A burn injury is most devastating.
Have the rear of the 911 down-wind.
Disconnect the CDI.
Have two friends with large fire extinguishers positioned up-wind.
Much more is not ‘too much’.

Best,
Grady

[maniqz]

Originally Posted by T77911S

if the injectors where not moved, then it is not the injectors. i assume the lines where swapped AT the FD side only.

the 81 has adjustable fuel flow for each port on the FD. either the FD itself is bad/dirty or someone has messed with the adjustments on the FD or they where never right to begin with.

i checked the fuel volume out of my FD by placing the injectors in 6 containers. with the FP on, i raised the sensor plate all the way up to fill the containers. then i each container into a common container, one at a time to eliminate the contaier weight, then weighed the fuel on a postal scale. i also did the test with the just the fuel lines and no injectors to take them out of the equation. i did this several times with different heights of the sensor plate.

i had one injector line (it was a bad line) that was so bad i did not have to weigh it see that it was bad. the line was blocked only allowing about half the fuel as the others.

there was a guy on here, alan sonething, that went through this with his 930 FD.

I would love to do this test, what's FD and FP? How do I go about it? TIA.

Joe

[L8apex32]

Originally Posted by psalt

Are you saying you removed the injectors for this test ?

I removed the hard fuel lines from the injectors and had them dump into individual catch pans. I then weighed the results.

[Grady Clay]

When you do this test, include the injectors.

Put the injectors each in a simular weight & size small-mouth container. A galss 1 gal bottle works well.
Measure the weight (+/- 0.1 g) of each empty container and record.
If you can use a cap (lid) that you can cut, notch it to clear only the hose.
The container needs to be vented but have all six the same.
Measure the weights of the 'full' containers. Subtract the weight of each empty container. Record the difference.
You want to measure small differences in gasoline weight among the six containers at the various positions of the sensor plate.

The more and better data, the better anylisis can be made.

Best,
Grady

[L8apex32]

Here are the results from an injector fuel quantity test, this time through the injectors (in grams):
1: 131
2: 131
3: 130
4: (scale turned off as I poured in fuel, lost sample)
5: 128
6: 130

[Grady Clay]

Joe,

FD = Fuel Distributor – the device on top of the fuel injection with all the pipes/hosed coming and going.

FP = Fuel Pressure – with CIS measured with a set of unique gauges, hoses and valve.

L8apex32, can you post some images for Joe?
See if you can find some scales that are 0.1 g or better.
Record the position of the sensor plate each test. Use small wood blocks to repeat the sensor plate position. Try and estimate % of full.
Record the amount of time for each test (this can help spot a flaw in the test method).

I can’t tell anything useful from this last set of measurements.

Someone needs to chime in with a relationship between a 1% difference in mass fuel flow as to how it relates to exhaust gas mixture.

Best,
Grady

[boyt911sc]

Originally Posted by L8apex32

Here are the results from an injector fuel quantity test, this time through the injectors (in grams):
1: 131
2: 131
3: 130
4: (scale turned off as I poured in fuel, lost sample)
5: 128
6: 130

L8apex32,

It would be nice to take multiple readings. Example do 3 different measurements to determine the trend of the data. Instead of using weight measurement, I would switch to volumetric measurement. A small plastic burette would be ideal for this kind of test. Then, repeat the test by randomly switching the fuel injectors or simply switch the 1-2-3 injectors with that of 4-5-6. The goal is to determine if the fuel injectors are the culprits or not. Keep us posted.

Tony

[L8apex32]

I have now taken off the fuel distributor and cleaned it. It was dirty on the outside but clean on the inside. The air flow sensor plate moves as described as does the piston in the fuel distributor.

The rubber gasket sealing the fuel distributor to the intake was damaged. A chunk of it was missing, maybe 1/12th of the circumference. I have replaced this gasket.

I re-took the fuel pressure readings, according to the Bosch K-Jetronic Service Manual. It has checking the WUR first. Those readings are:
control pressure cold, no vacuum: 2 bar (spec is 2.4-2.8 bar at 24.4deg C no vacuum; spec is 1.8 - 2.2bar with vacuum)

control pressure warm, no vacuum: 2.7bar (spec is 3.4 to 3.8 bar; spec is 3.4 to 3.8bar with vacuum [no change])

When you get bad readings for the cold test, its says to check:
* fuel delivery to control pressure circuit (60 sec -> 165g = .232 liter/minute: ok)
* fuel return (not checked as this would cause a "too high" pressure, not too low)
* warm-up regulator defective

When you get low readings for the warm test, it says to check:
* no current to warm-up regulator (not sure how to check current)
* voltage at WUR too low (I saw 11-12v on my needle, spec is 11.5v)
* fuel delivery at WUR too low (measured ok above)
* warm-up regulator defective

Of course, all of these measurements depend on good fuel pressure and flow. The flow measured 1100ml after 30 sec at the fuel accumulator. The system pressure is 66psi/4.55bar between the fuel distributor and the WUR (spec 4.5 to 5.2 bar). Those are in spec, but the pressure is at the lower end.

This seems to point to a bad WUR. Any thoughts?




WUR - fuel pressure gauge connected


Fuel distributor, all injector lines removed, fuel pressure gauge connected

[psalt]

Remove the electrical plug at the WUR and test for 12 volts with the key "ON".

Low warm control pressure should cause AFR to be rich everywhere, did your test show this ? If the dyno test in the OP was on a hot running engine, the control pressure could have been higher from the engine heat. What were the AFR numbers ?

[L8apex32]

I reported the voltage at the WUR in post #17. I know there is a lot to read in this thread.

Two dyno pulls showed A/F of 12:1 throughout the range, except, strangely, it was really lean around 3500 rpm.

I'm wondering if someone adjusted the mixture setting to compensate for a bad WUR to get the A/F to look right.

Unfortunately, I don't see a connection from this scenario to "suddenly going lean at the track".

I assume that I could do a dyno run now with good A/F, but that it could go lean at any time.

[psalt]

My advice would be to drive the car until the engine is really hot, then check warm control pressure.

What is "really lean around 3500 rpm" in AFR ? The word "lean" means above 14.7, if you really mean this, no amount of idle screw twiddling will cause that.