Having worked in engine controls for some time now, let me tell you my
understanding on how O2 sensors work. I'm on the periphery of this, not
directly involved in engine management system design. But I've been
designing electrical systems for engine powered machines for several years
now.
The sensor oscillates. It's output is part of a tank circuit. It does not
put out a DC voltage, rather a ac voltage. The computer reads the amplitude
and frequency and interprets this as the CO level.
So there is no way to gang or parallel sensors (simply anyway) and keep the
ECU happy.
Now if the ECU were set up to read both sensors to begin with, that is a
different story.
Terry.
<jon@dakota-truck.net> wrote in message
news:atejo3$5ed$2@bent.twistedbits.net...
>
>
> How about a good old fashioned DML engine theory discussion? :-)
>
>
> The Magnum engines all have a Y pipe with the oxygen sensor placed
> right after the Y. This is obviously a good place since the sensor can
> sample both banks of the engine, and it is close enough to the engine
> to be able to compensate for A/F ratio changes quickly.
>
> It becomes a problem when changing to a dual exhaust setup. The
> generally accepted theories on what to do with the exhaust sensor in
> a dual exhaust are to put it in (or after) one of the downtubes (in
> which case you pretty much assume (hope) that both banks of the engine
> are running at the same A/F ratio), or to put it in the H pipe or X pipe.
> Some people have said they haven't been able to get a good reading from
the
> H pipe, though I imagine right in the middle of an X pipe would not
> have that problem.
>
> The problem with the first solution is that you are only sampling one
> side of the engine. This could cause a problem if for some reason the
> side of the engine with the O2 sensor in it was rich but the other
> side was lean. (Perhaps due to the crossover tube on the fuel rail
> being restricted?) The computer would see the engine as rich and lean it
> out some more. Not a good thing for the already lean side of the engine.
> (This is actually how my Barracuda is currently set up, as this is what
> most people do...)
>
> The problem with the H/X pipe solution is that it only works if
> you have an H/X pipe. I was lying under my Ram today and had some ideas
> for the exhaust but an X pipe did not look like a viable solution.
> (In order to have enough room for an X pipe with what I want to do,
> the exhaust would need to hang under the frame where it would be
> subject to being smashed up (4x4). Also, the H or X pipe is often
> back a ways - this might be too far from the engine for the computer
> to respond to A/F events in time, or might cause the computer to be
> constantly overcorrecting (porpoising).
>
>
> It seems to me that if it were possible to use 2 oxygen sensors,
> both of the above problems would be solved - the catch is how to pull
> it off? One idea would be to just take the O2 sensor connector and
> Y each wire to create two connectors. I would guess this would work
> OK for the ground wire and the two heating element wires, but what about
> the signal wire? If both O2 sensors were reading 0.5v, what would be the
> reading be at the computer? 1v? 0.5v? none of the above? I have not
> tried this so I'm not sure - what happens when two voltages come together?
> Do they add together or try to equalize at the same voltage or what? I
> know that when you wire two batteries in parallel, the strong battery
> discharges and the weak charges until they equalize at the same voltage;
> would the same thing happen with the oxygen sensors?
>
> The above idea seems so simple that if it actually worked, everyone
> would be doing it. So, I'm assuming it won't work to just tie the
signals
> of 2 oxygen sensors together. What about a circuit (either analog or
> digital) which would generate a true average of the signals?
>
> If I remember right, 0v would be full lean and 1v would be full
> rich. If that is the case, what about a variation of the above idea
> where a circuit would pass through the lowest signal to the computer?
> (For example, if the left O2 sensor was reading .6v and the right was
> reading .7v, the .6v signal would be allowed to pass through to the
> computer.) The idea there of course being to send the leanest signal
> of the two to the computer to keep the engine safe (too rich is better
> than too lean).
>
> Another variation of the above idea is a circuit which would
constantly
> shift between the two sensors. It would essentially be a switch which
would
> pass through the left O2 signal, then after X amount of time, allow the
> right signal through instead, then after X amount of time, back to the
> left sensor. The X amount of time would need to be determined, I would
> suspect somewhere in the 1-5 second range?
>
>
> Some of the above ideas I am not sure would work, but some of them
> I am fairly certain would work (it would just be a matter of building
> it right). So... Does anybody have any experience, ideas, opinions,
> etc. on any of the above?
>
>
> --
>
> -Jon-
>
> .---- Jon Steiger ------ jon@dakota-truck.net or
jon@jonsteiger.com ------.
> | I'm the: AOPA, DoD, EAA, NMA, NRA, SPA, USUA. Rec & UL Pilot - SEL
|
> | 70 Cuda, 90 Dak 'vert, 92 Ram 4x4, 96 Dak, 96 Intruder 1400, 96 FireFly
|
> `------------------------------------------
http://www.jonsteiger.com ----'
This archive was generated by hypermail 2b29 : Fri Feb 06 2004 - 11:48:14 EST