Lambda oscillating a bit

[Vlad]

I had bee tuning cars for quite a time now, and have always wondered why, on some engine, is the Lambda (AFR) oscillating. For example, last week i have fitted on my SR20 an M84, and even if i set the same values on the map around the idle point, 0% fuel compensations, and the output DC is constant, lambda still oscillates between 0.96 and 1. Injectors are nismo 555, and idle PW is about 1.6ms, so no problems with short opening time. The car runs great, fast, and smooth, this is just a theoretical question i've been thinking of for the last few years as it happens on quite many cars.
Regards, Vlad

[David Ferguson]

How fast are you logging the Lambda data? Are you sure your exhaust system is completely sealed? At low speeds, it's possible to actually measure the the exhaust pulses from each individual cylinder firing. If you have air leaks, it could draw extra oxygen into the exhaust and the oxygen sensor would measure that. I see more variation of the O2 sensor is located near the exhaust exit.

[Vlad]

My sensor is located at the bottom of the downpipe and i'm pretty sure it has no leaks...Funny fact is that, even during cruising, when i hold a steady, let's say, 15% tps, even with, or without lambda control, it still jumps a bit up and down, no more than 0.2 lambda, but still, it isn't as steady as it was with the stock injectors...maybe the injectors spray angle is a bit different, i did not have time to look at that when i mounted them on...sandy i have little time for my own car...due to client's cars. Beside the exhaust leaks, what could make the lambda jump up and down like that ?

P.S.:beside 4 cyl selection=full sequential and 24 cyl=semi sequential, dose the m84 have other injection modes ?

[David Ferguson]

so it could be individual cylinder effects -- if the injectors are not all flowing the same, or the manifold temp and airflow is different between cylinders. It might be an interesting experiment to try adding or removing 2-5% fuel to one cylinder at a time to see if you can even out single Lambda reading at idle.

If you can idle at 600 RPM, then that would be 10 revs/sec, or 20 combustion events per second for your 4 cylinder -- log the Lambda at 50 or 100 hz and compare the logs with the individual cylinder trims for one cylinder at a time to see if that makes a difference.