Some people have claimed that they can monitor/measure these, but nothing specific as yet.

So it's a question for all: how can I measure, even if approximately, air fuel ratio and the amount of CO, CO2, HC, NOx, and O2 (or at least some of them) in the exhaust gases, without spending a small fortune on test equipment? I guess for O2 I can use the built in sensor, but I'm not aware any automotive sensors for the rest. And I want something a little more accurate then eyeballing the spark plugs.

I want to run my car a little leaner, and because of the electronics ('95) that's not as easy as adjusting the carb,

Here are some reading on the why:
Improving IC Engine Efficiency
Lean burn - Wikipedia, the free encyclopedia

 

For simple air/fuel ratio measurement, I would use a wide-band O2 sensor and gauge. Innovate makes a good, reasonably priced kit for this (I think that the LM-1 is what I've used). As for the rest of those things, a 5-gas analyzer would do the trick, but as far as pricing of those units, I have no idea.

Innovate Motorsports Wideband Air/Fuel Ratio Tuning

 

Yes, a wideband O2 is what you need for the A/F ratio. However.. there is no way to change the amount of fuel the computer delivers without a tuner of some sort, and even then this won't be a trivial accomplishment. This is because the whole EFI system is based around the stoichometric ratio of 14.7:1, the computer just calculates deviation from this point in closed loop operation, and the narrow band O2 sensors employed in all EFI systems these days just switch between high and low output at this A/F ratio. To get the computer to operate at any other A/F ratio, you will have to force it to open loop mode, and then trim all the fuel tables by hand to achieve the goal. Racers use this method when tuning for maximum performance, they target about 13:1, and achieve pretty good results at times. But they aren't too concerned about idle manners or emissions, just WOT performance. Targeting a ratio leaner than 14.7:1 will also introduce the potential for detonation, so having a motor with a knock sensor and a computer that has the ability to retard ignition when knock is sensed will be helpfull.

 

Would a scangauage be of any help with this? I don't know, so I am more asking than anything else?

I have heard of websites and wikis that have some ecm hacking done, perhaps stuff like that may be of help somewhat?
Welcome to diy_efi/efi332

 

Actually, there is. Modifying the output of the O2 sensor, for example, is one.

If I'll have plenty of time, I might look into reverse engineering the computer, but I think that would much harder than modifying/faking the output of a few sensors.

 

That's easy to say but a lot harder to do. The stock sensor output toggles between a low voltage and high voltage as the A/F ratio passes through 14.7:1, and the computer is programmed to interpret this switching as representing unity.. or the A/F ratio it is programmed to work towards. What you need is a sensor that toggles at 15:1 or so, I don't know if these exist in a simple narrowband sensor form, and if not you would probably have to start with a wideband sensor that has some custom electronics to simulate a narrow band sensor that switches at your desired ratio.

Innovate Motorsports has such a device, I actually have the sensors kit below but have not installed it yet. This unit has programmable analog outputs that simulate a narrowband sensor, and it appears you can adjust the A/F ratio it switches at. Have a look...
LC-1 Lambda Cable with 02 Sensor : Wideband Controller Cable for Dyno, ECU, Data Acquisition, or Gauge Applications

 

O2 sensor is not that easy. The output from the sensor looks like a Sinewave as it switches on and off. Its not a simple put a resistor in line deal.

 

I didn't scope any O2 sensor (especially since my scope was stolen in a break-in) but I highly doubt it would be a sine wave, Maybe a square wave.

 

Lambda sensors / exhaust oxygen sensors

Its a sine wave

 

The sensor itself is responding to the computer varying the fuel mixture for best emissions performance. The reason it swings rich to lean is because one pollutant reacts in the catalyst better at a rich mixture, and another pollutant reacts at a lean mixture, so the mixture bounces between the two to maximize converter efficiency. Without that, the reading would be fairly steady. It's not exactly a sine wave either, according to the graph on the page linked to above.

It shouldn't be to hard to build the electronics to shift the curve of the sensor output to trick the computer into a leaner mixture, then tune for the proper mixture at high loads while in open loop.