Adding an O2 sensor?

Again, depends on your local laws. In BC, if the car came from the manufacturer
with a cat, then it must always have a cat. In most of the province, nobody
will usually notice or care, but legally it must still be there; in the
Vancouver area, where we also have emissions testing, the car won't pass if it
originally had a cat that's now missing, regardless of the year.

 

Fair'nuff... it was more of a generic response to Jim's comment that a cat is
"not essential".

I'd suspect it's more likely the various parts were left out for cost savings,
with no regard for emissions OR mileage.

 

Ayon kay Matt Ion:

Oh, no slight meant there, Matt. I was just making it clear that I'm
not contemplating anything illegal

Anyway, I'm somewhat confused by TeGGer's response over at
rec.autos.makers.honda, where he says that the O2 sensor is simply
there for emissions. He's a resident guru there and I'm inclined to
believe him, but OTOH when someone asks about poor mileage, group
posters usually point to the O2 sensor as the culprit. Also, most
references on the net say that the sensor is vital for normal running.

Anyone have any idea on how this ECU was tweaked to run without an
oxygen sensor? How does running in open loop affect things?

 

I'd guess that the Honda programmers would only edit/alter their ECU
program enough to ignore a missing sensor rather then write another
completely different one for mileage.

Some adventurous soul here could disconnect their O2 sensor and see if the
ECU then flashes the "check engine" light. A brief test.

 

Open loop makes the fuel-air mix richer than optimal(for emissions) ALL the
time,just as it does until the engine warms up to operating temperature.You
have to avoid a too-lean condition in -all- operating conditions,to avoid
damaging the engine.

Closed loop optimizes the fuel-air mix for the leanest possible mix -
without- risking a too-lean condition that could burn a piston or make the
motor run too hot.It minimizes the extra hydrocarbons (rich exhaust)that
the cat-converter has to burn up,for best emissions.

You have a open-loop at the start because the catalytic has not warmed up
yet,nor has the motor.(coolant,plugs)

 

Jim Yanik said:

This makes a lot of sense.. So a sensor retrofit would be useful then?

If only some kind soul would try this... But I think the ECU would
just throw a CEL. I'll look at the ECU pinouts today, and post back.

 

I doubt it would be worth the expense.
Considering you would have to change the ECU in addition. That's big bucks.
If you're looking for better gas mileage,changing your driving habits would
be far more productive,keeping your tires inflated properly.

It's supposed to;it's a failure of the O2 sensor(missing = "open"
sensor),and the control loop is no longer closed,emissions not controlled.

Your car should have a *different ECU* if it does not have an O2 sensor.
Honda kept the same wiring harness for cost and simplicity,but the ECU
would have to have different programming to ignore the missing O2 sensor
and not flash the CEL.
Since Honda solders in their PROMs,ECUs are different.

 

Thanks, jim. You make a good point.

I'm really curious about this "open" vs "closed" loop mode. Are we
greatly disadvantaged by always running in open loop? Or are we "lucky"
as TeGGer says? Can't an ECU run in closed loop using just MAP, TPS,
TA, etc sensors?

 

no, it needs to know "lambda", the actual ratio of fuel to air the
engine is experiencing. the only way to do that is with a sensor.
older sensors only detected what's called "stoichiometry", i.e. when the
mix is perfect. newer sensors are "broad band" and detect the range
typically sought, i.e. lambda from 0.8 to 1.2. [the former is lean, the
latter rich - and peak torque.] either way, a sensor allows the ecu to
tune the engine perfectly for its running conditions. closed loop is
the way to go.

if you want to retrofit, you could look at this as a science experiment
and buy a used ecu online from the states. they're typically in the $30
to $100 range for the standard models.

 

Well, it's worked for over 100 years... your average classic muscle car has
virtually nothing in the way of "engine management".

 

Often,those "classic muscle cars" would not run well in other than optimal
weather or driving conditions.

 

"closed loop" is just using feedback from the O2 sensor to continually
adjust input fuel/air mix to optimum. Fine tuning on the run.

"open loop" is just "dump enough gas in to make it run" for a wide range of
conditions. There's no correctional feedback.

Without the sensor measuring at the END of the combustion process,it's not
"closed loop".

the "loop" is; input..combustion..output..*measure output*..make CORRECTION
to input;repeat.
Without the correction,there's no "loop",it's an "open-ended" process.
The closure is the feedback corrections.

Think of the AC and thermostat in your home; you set a temp,and the AC
runs,cools the room,and the TS measures what the room temp is,and when it
reaches your setpoint,it shuts off the AC,and cycles. That's "closed
loop".

"Open loop" would be you turning on the AC and running it for 10 minutes
every hour,say on a timer.
The room MAY be close to where you want it,perhaps not.That would depend on
how hot the room was to begin,and how how it is outside.
There's no measurement and feedback control.

 

Thanks. I think I see now. That since the main output of the ECU is the
injector duration, then the main input (to have a true feedback
"closed" loop) is a sensor that measures the effects of combustion (O2
sensor).

So I take it you mean that based on the other sensors (MAP, TA, TPS,
PA), the ECU does no fine tuning based on real-time combustion result,
but rather only bases the injector duration on predefined "maps" in its
ROM. So, the ECU does vary the injector duration and spark timing, but
only based on what it "thinks" is best for a given situation, but would
still not be "closed loop". You really need to somehow measure the
exhaust. Is my understanding correct?

This makes sense, but isn't the O2 sensor just another input for the
ECU? Is it any different from the other sensors, that the ECU uses to
vary timing and injection? What makes O2 so special, if the responses
to it are also based on a ROM "map"?

TIA

 

Closed loop is the way to optimize for low emissions and good fuel
economy. Open loop certainly can work. Most cars manufactured before
1980 were open loop systems be they carbs or early fuel injection.

For those countries where leaded fuel continued to be available the open
loop style carried on for a long time.

John

 

based on the maps and readings of the sensors, including lambda, the ecu
can then compute in real time the precise amount of fuel to be injected.

if you're interested, there's an open source injector project out there
called [charmingly] "megasquirt". once googled, you can read the whole
nine yards about how injection works, even source code an ecu runs on.

 

Well,the desired end result is complete combustion(a clean exhaust),and
exhaust O2 content is an indication of how well that's happening.
They are real-time corrections,and measure ACTUAL exhaust output.
(if a plug misfires,or a input-condition sensor is inaccurate,this feedback
corrects for it.it also allows for individual cylinders' injectors to be
corrected,compensating for differing air flow thru the cylinders.They are
not necessarily all equal!)


I believe there may be two 'maps".
One for open-loop,and one for when everything warms to operating temp.
The 2nd map would be much simpler,and smaller in data size,or just a
refinement on (or corrections for) the primary map.
It's like the ECU first sets injector duration for the TPS,IAT,MAP and RPM
via the map,and -then- looks at O2S data and fine-tunes if needed (at op
temp) for optimal exhaust O2 content. It also minimizes the work the
catalytic has to do,-and makes it last longer.

It does take some amount of time to balance the loop once a variable is
changed.(like throttle position or intake air temp)
Loop response cannot be instant,there's always some hysteresis,but it's
inconsequential,until the O2 sensor gets "slow" in it's response(degrades).

OBD-II systems supposedly measure this O2S response time and set a trouble
code (and CEL)if it gets too long(or fails),and also uses a 2nd O2 sensor
to back up the first sensor and monitor the catalytic converter's
performance to ensure low emissions.

 

To add more;

Yes.
The TPS,MAP,IAT,and RPM sensors are all input-side variables,the O2S is the
ONLY output measurement.

If I said this in the first post,apologies.
(it's then just a "senior moment"!)

 

Wow, Jim, kudos on all the info, you really hit the spot. Thanks!

I'll take a look at "megasquirt", it sounds really interesting!

 

Uh, hold on there. Few if any modern fuel injection systems monitor
each cylinder and adjust them independently. At best you have one O2
sensor per bank on a multi-head engine. The ECU has no information
about individual cylinders.

I know that some years ago Saab was working on a system which used an
ionization current passed through the spark plug during the exhaust
cycle to try and do the individual cylinder optimization you talk about,
but I don't know if such a thing ever got into production. Certainly
that is not the scheme in use on the majority of vehicles today.

Closed loop fuel injection also does not correct for ignition problems
like a misfiring spark plug.

John

 

Well, I can't claim to know exactly how these systems DO work, but I can
hypothesize that it wouldn't be all that difficult for the ECU to get an idea
what's going on with each cylinder individually. Given timing input, it will
know which cylinder is firing when, and a subsequent change in the O2 sensor
reading is pretty easy to correlate to that cylinder.

As O2 sensor input is split up to separate groups of cylinders, as has already
been discussed here, such calculations become even easier...