Wideband reading erratic, poor performance

[patkeefe]

The short story is I got the turbo SC running enough to put it on a truck to ship it to FL. I drove it today to fuel it up, it was popping in the exhaust, ran generally crappy, sounded like it was detonating, and the AFR was somewhat erratic. I ran it long enough to get the il warmed up to 210 on the gauge, and brought it home.

An example is driving along at a few thousand RPM in third, the AFR gauge is reading 14 is, the jumps to 11, then back up. All of this within a few seconds, with no change in load or throttle position. Not making boost at the time.

AFR O2 sensor is Bosch, AEM gauge, mounted after the turbo. Maybe there is a leak at the V Band? Is it possible the sensor/gauge is reading incorrectly? The O2 sensor is basically controlling the SDS EM-5 fuel/ignition system I put in.

I am open to suggestion. I am thinking of bringing it to my friends shop next week, and have him put his gas analyzer in the tailpipe, so I can compare O2 readings.



BTW, I just got it back from the paint shop...

[scottrx7tt]

Sounds like maybe an ignition problem. If you have a cylinder that isn’t getting fire, it will show rich on the wideband due to the unburned fuel passing through to the exhaust.

[993Speedster]

Originally Posted by scottrx7tt

Sounds like maybe an ignition problem. If you have a cylinder that isn’t getting fire, it will show rich on the wideband due to the unburned fuel passing through to the exhaust.

Misfires read lean, not rich. It's the oxygen (or lack thereof) that is being read, not the fuel. This is typically why you tune on the rich side and give your O2 feedback more authority to lean out than enrich (and also why you can't trust your AFR gauge during launch control/anti-lag). A situation where you're already misfiring from too much fuel can cause the O2 to read lean and fatten up the mixture even more so and really drown out in fuel.

Exhaust leak is possible but usually more prevalent at really low-loads/RPM when you're not flowing a lot of exhaust volume. Cruising with some load and RPM I'd suspect you'd be flowing enough exhaust for a little leak to be negligible. And once again, extra air entering would be showing as a lean-trend on your AFR gauge, not big rich swings.
Can you give some more info about your turbo setup and what you're using to control fueling?

EDIT: saw the SDS EM5 note and don't know much about them but was curious if your AFR gauge was simply a feedback for the driver or if it was feeding closed-loop control to an ECU of some kind. Was it running A-ok at one point and now like this or you still working out the tuning aspect?

[patkeefe]

Still working out tuning. It the O2 sensor is feeding information back to the SDS ECU. I have done leak down tests but they were OK so I have ruled out mechanical issues. Valve adjustment is OK.
Thanks for the input
Pat

[flightlead404]

Does the engine change the way it runs, or is there exhaust popping or anything else when it suddenly changes?

What you describe sounds more like a bad electrical connection than actual changes in AFR

[patkeefe]

No change in the way it runs when the O2 gauge wigs out. I am going to bring it to Anthony and Frank this week to see if we can verify the O2 readings.

[Rawknees'Turbo]

Pat, that rubber, oil discharge hose, with worm clamp on the turbocharger side, looks like walletbreak waiting to happen, bro (pretty close to the turbine and its feed pipe for rubber).

[patkeefe]

Originally Posted by Rawknees'Turbo

looks like walletbreak waiting to happen, bro

HA. That is correct! As soon as I get the car moved to its new home in Florida next week, that is changing to AN.

That is the test into the sump, as opposed to the electric scavenge pump I had...guys on this board said to go right into the sump. It does work better, and I used the space from the electric pump to put in a fenderwell condenser.

[spuggy]

Originally Posted by 993Speedster

Misfires read lean, not rich.

This.

Kind of a giveaway; it reads lean, add fuel, reads leaner...

Twiddling with my hot idle, I had a rich misfire that I couldn't hear or feel - it read leaner than I'd prefer, but fattening up the fuel went the wrong direction...

Originally Posted by 993Speedster

A situation where you're already misfiring from too much fuel can cause the O2 to read lean and fatten up the mixture even more so and really drown out in fuel.

Yeh, I'd be wary of using an O2 sensor and a piggy back to drive fueling. Ideally, you want a means of providing load (eg MAP) and fuel tables that Just Work. I have the option to run closed-loop lambda - but don't, I just log it.

Recently had an LSU 4.9 change calibration to the point where the motec decided the values it gave were out-of-range/invalid; got 2 minutes worth of "valid" lambda data logged on a 30 minute drive.

If they change, you can free-air calibrate them - but as they age, they react slower as well. They can also get contaminated (as opposed to just dirty).

[patkeefe]

If I tune it a little rich (like into the 13's), it is pig rich...I can smell fuel, it gets in the oil, plugs are black, etc. I leaned it back out, put in 30# injectors (I had 44#), and now it smells better, but runs crappy, as I have stated. It seems to run the same either in closed or open loop.

It probably doesn't help that I never got the initial tuning done. I have driven the car three times since I got this engine back in-to and from the painter, and to the gas station.. I will piddle with it next week, before I bring it to the shipper to go South.

[Porsche 935]

Then your oxygen sensor is not reading right or you have another problem. 10's are rich, 13's not going to cause those issues.

[patkeefe]

So, I have been thinking about MAP sensors. How much enrichment do you EFI guys do related to manifold vacuum/pressure? Is that linear (fuel vs MAP) or does it become more of a parabolic (exponential) curve as the boost goes up? I am thinking I should maybe be linear through the negative (vaccum) range, and get more exponential as boost starts to come in (about 2 PSIG).
Thanks!

[spuggy]

Originally Posted by patkeefe

So, I have been thinking about MAP sensors. How much enrichment do you EFI guys do related to manifold vacuum/pressure? Is that linear (fuel vs MAP) or does it become more of a parabolic (exponential) curve as the boost goes up? I am thinking I should maybe be linear through the negative (vaccum) range, and get more exponential as boost starts to come in (about 2 PSIG).
Thanks!

That's one beauty of a MAP table; fuel map ends up almost entirely flat. And anywhere it isn't is usually wrong...

My map pulls fuel at higher boost (K27/HF runs out of airflow/AFRs go too rich) and I diddle some out at idle, but otherwise fuel is fairly flat - at least in the areas the motor can actually reach.

Like this:

[Uwon]

Personal experience with my turbo hot rod- I got better more consistent AFR readings once I moved the O2sensor away from the really hot turbo downstream to the other side of the muffler.
Hope this helps.
Johan

[patkeefe]

I did some very minor tuning today. I basically put in some more advance across the board, and have the retard feature to pull advance back when the MAP sees higher pressure. This seems to have helped a lot.

As I am moving, and sending the car on a truck to sunny Florida, this may be my last chance at anything. I grossly underestimated the amount of time and effort necessary to downsize and move.

Spuggy, that is a great image on that graph. Did the software you have for EFI generate that?

[spuggy]

Originally Posted by patkeefe

I did some very minor tuning today. I basically put in some more advance across the board, and have the retard feature to pull advance back when the MAP sees higher pressure. This seems to have helped a lot.

Sweet

I posted my ignition map in the 911 forum; the motor loves advance. I tried a twin-plug map that was a lot more conservative - and it ran quite sluggish/poorly in comparison.

I'm twin-plugged and running ~26 degrees on boost, up to 35 N/A WOT and up to 37.5 on part-throttle cruise.

The Safeguard doesn't have jack to do, AFRs are whatever I'd like them to be - and the thing is a rev-monster. it just wants to go. I'm quite sure it'd like even more advance.

Originally Posted by patkeefe

I grossly underestimated the amount of time and effort necessary to downsize and move.

Ha. You would think, the number of times I've done exactly that, I'd know better. And yet, everytime, takes me by surprise...

Originally Posted by patkeefe

Spuggy, that is a great image on that graph. Did the software you have for EFI generate that?

Yeh, that's the MoTeC ECU Manager; it's all a live display of bar graphs/dials/sensor outputs (user configurable; build your own display) when connected to the ECU/motor is running.

I just set it to "Fuel - Main" view and moved/resized that table to show most of the cells, and then did a print screen and cropped it.

Here.s the ignition map:

[patkeefe]

I drove the car about 35 miles today to put it on the truck to Florida. It really likes that extra advance...I think I have 32 total advance, but it gets pulled back a few degrees on boost.

I just got my account here unlocked. Seems there were some email issues for the past few days?

[Uwon]

Pat, so at what point would you start worrying about pinging/denotation? Do you have a knock sensor installed? I note that you are running 35 degrees at 100kpa which is on the power, albeit mild.
Johan