how do you know that your lambda measurement is accurate ?
first I though narrow band oxygen sensors were very non-linear
basically you can only tell if it is running rich or lean and the computer uses this to adjust the injection time so that the sensor will oscillate between rich and lean constantly
so I guess to know if your engine is healthy you wouldn't just display the actual value of the sensor nor an averaged value
but you would display, maybe, how often per unit of time the sensors switched between rich and lean (I would guess the more often the better, and if it stayed at one value for too long that would be a fault condition and if it stayed lean for too long that would be a "master fault" requiring engine shutdown or limp-home mode)
plus there are at least 3 main types of sensors
zirconia narrowband, this is the most common
zirconia wideband, I guess this is what new cars and performance cars have , also what they use in aftermarket A/F gauges
and titania sensors which is rare (but I just happen to find some on a nissan RB26DETT engine, these are coupled with EGT sensors I guess because their response is temperature sensitive)
a good O² sensor controller should support all of these I guess, so that all can benefit without buying new sensors
but maybe people playing with this will like to buy new wideband sensors
a narrowband sensor with no heater costs 10$
http://cgi.ebay.ca/O2-OXYGEN-SENSOR-STD-1WF-SG12-HONDA-CRX-91-84-/200335656679?cmd=ViewItem&pt=Motors_Car_Truck_Parts_Accessories&hash=item2ea4ef86e7
there's another kind that has a uncontrolled heater that is still narrowband but gives accurate measurement faster
you can get a wideband for about 40$
this is a Bosch LSU 4.2
it's a good sensor but it doesn't like high pressures
a Bosch LSU 4.9 costs about 60$ I'm not sure if it's better a getting high pressure gas blown on it, there's another one called LC-1 that apparently you can put right after the turbo for the best response, I think they cost 140$ I can't find a link right now
there's a company called innovate motorsports and they make overpriced aftermarket automotives gauges and supposedly they use LC-1 sensors
the thing with wideband O² sensors is that to be accurate they have to be temperature controlled so that have a builting heater that you have to control and some sort of temperature, probably a K type thermocouple
to get an accurate reading you either keep the temperature at a stable temperature (I haven't found the celcius value yet) and the reading you get you run through a formula that takes into account the sensor response curve and that gives you the air/fuel ratio, but sometime the heater is not powerful enough to keep the temperature off so you have to compensate for it in the formula, also if the engine is heavily loaded the sensor temperature can go higher than the desired temperature even with the heater completely off. you have to compensate for that too
there is more information on how to make a wideband o² controller
http://www.megamanual.com/PWC/index.htm
but crucial details have probably been left out to impede DIY development of wideband o² controllers so information from the megasquirt project should be considered suspect until proven otherwise (megasquirt is a partially open ECU system that is getting more secretive as the community is no longer needed by the development team except for marketing purposes)
but still, there is a lot of good information on the megasquirt forum so that's a good source, just be wary
oh and I thought maybe it would be a good thing to have a break point between the datalogging and sensor interface backend and the display and user interface front end
maybe the backend could be a good start for making a truly open ECU
and a break point like this would allow someone to develop a separate front end , maybe some user want to re-use their cluster gauge, or use stepper motor dial gauges or even 7-segment, or those nice 3" cell phone graphical LCD but don't want to mess with the back end ?
what do you think ? makes sense ?