My '91 TBI Firebird couldn't pass emission testing, as there is way too much oxygen in exhaust (lambda around 1.3/oxygen 5% at 2000 rpm). I have checked all vacuum hoses for leaks, nothing, changed oxygen sensor, nothing, new plugs, distributor cap, rotor etc (to be sure there's no misfiring), adjusted timing, checked exhaust for leaks, nothing. Now I have no ideas what to check - intake manifold gasket? Not sure how to check it for leaks - as I understand, it must be pretty hard leak to get the oxygen level in exhaust so high? I suspect that there must be an intake leak and the engine runs lean, as the temperature of the exhaust seems to be too high. The plugs seem OK, oxygen sensor reads approx. 350-375 mV at idle. Any ideas are appreciated...

The oxygen sensor should be reading around 450mV at the stochiometric 14.7 A/F ratio. A lower reading would mean that either the sensor is bad (they don't last forever) or the mixture is lean. (which the emissions test confirmed)
Lean mixture can be caused by many factors but unless you hook up a laptop with datalogging software to the ECM, you won't likely find out which one.

Although, you could test all accessible sensors that provide information to the ECM. That means MAP, TPS, CTS, IAT and O2 (you already measured that one).

I don't see in your post whether you tried replacing the O2 sensor.

If you need specific info on how to test each sensor, do a search and if you can't find what you need, let me know, I can post some.

Lou

The O2 sensor is new Denso one and it seems to work properly. My dignostic guy hooked computer and said that all sensors are working OK.

You can measure the fuel pressure, it should be between 35 and 38 psi when idling and over 40 psi at WOT (or with the engine off but fuel pump primed).

Do you still have all the emissions equipment on your engine? If so, I would check the AIR diverter valves. If they stick open, they can cause extra oxygen in the exhaust.
Did you have high NOx readings during the test? Or just high O2?

Also, you would be surprised how much havoc aging injectors can wreak onto the emissions control. An injector flow test would show injectors that supply less or more fuel than expected. I performed the injector test at home in my own garage, it just takes time to do. Oh and since you have a TBI system, it's very easy for you (it's not easy on TPI).

Lou

Lou, could you explain me injector flow testing method?

I suspected the fuel pressure myself too, but shouldn't that be more an issue on higher rpm-s...? The engine seems to work OK, acceleration is pretty good and it don't hesitate on acceleration.

Edit: I think that AIR diverter valves are not an issue, as the oxygen from there could not cause exhaust overheating, or could it?

Depending on the condition of the FPR, the fuel pressure might be lower only with no load (high vacuum) but might be good under load (acceleration) so it's hard to tell without putting an actual gauge on the fuel line.

I am not sure why you mentioned exhaust overheating, do you have this problem? That would point to a severe lean condition (which could cause decent damage to the motor). The AIR diverter valve sends air to different parts of the exhaust, as the ECM commands it. If the ECM expects air to be (or not be) injected in the exhaust, it will compensate for it but if the AIR valves are bad, the mixture will be way off. That's why I was asking about your NOx levels from the emissions test, that would tell us whether you're running lean or just have a leak in the exhaust system.

The injectors have two basic tests: static flow and dynamic flow.
With static flow, you measure how much fluid comes through at constant pressure during a specific period of time. That would show a clogged injector, for example.
A dynamic flow is one that is not static, it is constantly interrupted. The injector is periodically cycled on & off to reveal dynamic flaws, such as a weak spring.

The static flow can be done with a simple stop watch and a cup. You turn on the fuel pump (NOT the motor) and let one injector fill the cup for let's say a minute or two. Afterwards, you mark the fluid level in the cup with let's say a Sharpie and dump the fuel back into the gas tank. Then you test the second injector for the same period of time and mark the fuel level with a Sharpie. You do so for all injectors (if you have more than two) and compare the levels. I've seen injectors with 20% difference in static flow.

Dynamic flow is performed in the same manner, except you use a multivibrator circuit to actuate the injector a few times per second. Use of relays is NOT recommended to drive the injector as they introduce a very imprecise delay.

Hope this helps.
Lou

I replaced the intake gaskets - someone has put the old ones back with some kind of glue or silicone and as I saw few rusty spots on the surfaces of the cylinder heads, then they were not too hermetically sealed. But it didn't resolve the problem.

Now I built a ALDL cable and used WinALDL to see, what is going on. And I discovered, that when the engine is cold and ECM works in open loop, it idles very rich - O2 sensor reading was constantly over 800 mV. Warmed up, the ECM went in closed loop and then very strange behaviour started - the O2 reading switched constantly from extremely rich to very lean - from ~900 mV to ~200 mV and the RPM reading switched between 450 and 700 RPM too. I suspected defective ECM and ordered new one from Rockauto, but the replacement they sent, seemed to be bad too (after replacement, there was no error codes, after starting engine, it idled very high, then the RPM slowed down, almost stalled, misfired and then idled high again. The SES light flashed chaotically when that happened. I stopped the engine and tried to read error codes - nothing, only 12 repeating and then suddenly chaotic flashes again). So I try to exchange the suspicious ECM and get the readings then.

Has anybody any idea what could give that kind of symptoms? Bad O2 sensor (its brand new Denso - I haven't heard too much good stories about these)? Something with fuel system? I haven't got a fuel pressure manometer yet to measure my fuel pressure

Got new ECM yesterday and the RPM reading is stable. But O2 changes still - after startup it is about 450-470 mV, then slowly raises to about 800 mV, then computer shows rich flag and the O2 reading goes rapidly to 200 mV-s. After that the reading raises again. BLM is constantly 117 and INT changes between 122-134. I try to log the data and post it later, as I go home.

Edit: As I promised, I attached some WinALDL logs.

Are any other sensors reading off a bit?

Your open/closed loop statement makes sense, but think about it like this.

Sorry to those that I am boring/reteaching.

Open loop, the computer runs the car off of it's "map" until it decides it's time to switch to closed loop via coolant temp. sensor. and one other sensor that I can't think of (I wanna say IAT but I don't see how it could be that one. Maybe the o2 actually, it switches when the readings fit the "normal" parameters)

Anyway, the comp. switches to closed loop and the car runs off of the computer making decisions based on what the sensors say.

Now what could be happening is another sensor/actuator up the line could be causing the "Erratic" o2 readings. Perhaps something is not allowing the normal flow of fuel, something along those lines.

I'm just trying to shoot it from another direction.

I recommend to post this in the tuning forum, lots of smart people there with ECM knowledge.

Lou