ramco

I know that most of you guys are not concerned with gas mileage, but I was wondering if anybody has done this: I want to install a variable resistor in line with my O2 sensor, so I could fine tune my cruise speed mixture strength for optimizing fuel economy? I would like to install a vacuum gauge and TPS volt meter, so I have a reference point when making adjustments with the O2 resistance. I know there are data loggers that can do that, but that adds to much complexity and cost for me. A 0 to 5 volt meter that mounts on the dash, would be perfect for the TPS volts, if they are available. (this is for the B16B CTR motor in my Mini)..... Sugestions?

DCFIVER

Huh?This makes no sense. Buy a wide band and tune accordingly. The leaner you can run at cruise, the better your fuel mileage will be. A Zirc O2 is too limited for this.

ddd4114

Your factory oxygen sensor isn't linear, so that won't work as well as you think. The ECU just looks for the signal's inflection point about the stoichiometric ratio of the fuel.

Here is a rough example:

ramco

I understand that, so Stoich is between 200 and 750 mv. Im thinking I can dial in, say another 100 mv resistance, so when the O2 is signaling stoich, the computer will still think it is rich and pull out more fuel....no?
Re the wide band use, My OBD1 computer won't read a wide band for closed-loop, and shops don't know how to tune for cruise in open-loop. All they seem to know is idle and full throttle tuning, with a couple points in between if you're lucky. Dyno shop tuner guys don't have a clue how to tell what throttle angle I am at while cruising at 65 on the Freeway, and even if they did, couldn't simulate it on a chassis dyno. if I had a volt meter to record the throttle angle, that, along with the RPM would be a starting point.

ddd4114

No, that curve changes with temperature. The inflection point is 0.45 V. Plus, wouldn't a resistor shift the curve down and make it read leaner than it really is?

If your tuning shops don't understand how to tune partial throttle, avoid them. I'm sure you can find somebody in CA that can tune the whole map, but you might pay a little more for it.

Either do it right and get a wideband oxygen sensor, or don't do it at all. You have little - if anything - to gain with fooling the factory oxygen sensor like that.

DCFIVER

Stoich is 450mv. Period. But because of the nature of the ICE and the need for catalyst feed gases, the o2 will never remain at 450mv. If you add resitnce to the O2 via a resistor you will flag a code and likely have drivability problems.

I seem to remember a thread here a few moths ago about installing a wideband on an OBD 1 vehicle. Search Innovate i think. As for tuning at cruise, that is pretty much the norm nowdays. Any tuner that cant give a part throttle tune shouldn't be tuning a vehicle. Find a competent tuner.

DCFIVER

It does not change with temperature. Temperature is only needed to excite the noble metals at the tip of the O2. It has no bearing on the sensors reading once fully warmed up. The graph you posted only shows the extreme peak to peak point of the sensor. It does not represent the sensors normal behavior.

GagnarTheUnruly

iTrader: (3)

Is $150 too much to spend? For that you can get an MTX-L which is an all-in-one AFR gauge, sensor, and controller, and it can easily send a narroband mimicing signal to the ECU.

ddd4114

http://wbo2.com/lsu/lsuworks.htm

ddd4114

DCFIVER, what's with the deleted post? I thought you were "on a completely different level ... in terms of information", so I would like you to explain why I am wrong. Since you are obviously the smartest person here, can you please share your wisdom with the rest of us simpletons?

DCFIVER

I reread the info posted in the link and realized it is indeed correct. However my statement about temperature still stands. Your link reinforces the notion. It is needed to get the sensor functioning, (heat up the zirc,Platnium, and not mentioned in the link,cerium) which is the reason O2's have heaters. But normal operating temp will not affect the sensor and it is only at the extreme heated range will the values be skewed. You inferred that temp will affect the switchrate,and that is only true of a cold sensor. Also the diagram you posted is only the peak to peak excursions and does not represent the normal oscillation of an O2. Yes I am on a completely different level, but I surmised that was pompous and chose to delete that statement as well. If ,however, you would like to challenge me on the notion,then Im here for you son, swing for the fences.....

ddd4114

Oh. I didn't mean to imply that it would change the switching point, just that the shape of the curve would be affected (and only in extreme conditions). Like you said, the stoich voltage output will be the same.

For what it's worth, I didn't make that post because I thought you were wrong. I made it mock your arrogance. Deleting it was the right thing to do, and I probably should have ignored it in retrospect (now that I'm fully awake).

ramco

Well, You both seem to know more about the details than I do. I am a Mechanical Engineer, not Electrical. All I know is that the ECU adds or subtracts fuel, based on the voltage coming back from the O2 sensor. Common sense tells me if I reduce the voltage that the ECU gets, it will take out fuel, as it thinks the mixture is still fat. The sensor will then experience a lean condition and put out more voltage, until there is enough volts to get .45 volts through the resistor, satisfying the ECU, that the mixture has richened up to Stoich, even though the actual mixture is leaner, .......which is what I am looking for. If this is flawed thinking, please explain to me, in non-geek speak, why.
Regarding the simulated narrow band output from the wide band electronics mentioned above; I would still have to controll that output, in order to experiment with leaner fuel; mixtures while driving. It would be nice to see the wide band data on a gauge, but I think the ECU would give better results reading an actual narrow band than getting a "simulated" signal from some questionable electronics device. So, as much as I hate adding exhaust restrictions, I would rather just punch another hole in the collector, for a wide-band sensor and the gauge. Does anyone know why the sensors are typically installed so far into the exhaust pipe? Wouldn't they read the O2 levels just as well, with the tip just flush with the Pipe ID? It's no problem to make a longer bung.
I saw a referance to the cat......I don't have one, My Mini is old enough to be exempt.
Thanks.

DCFIVER

I already did.
Fluid dynamics. There would be little to no flow "up" to the O2 sensor were it to be sitting flush with the pipe.

ramco

Sorry, but I do not understand your explanation about why the variable resistor won't work; all I got was that you didn't think it would.
On the O2 flush mount question, I don't see why there needs to be a lot of flow past the tip, as long as there is some. The lag time from combustion chamber to to sensor location, is going to be less than .01 seconds at even low RPM. if you doubled that, it wouldn't matter. Further, the gasses are going to be rather uniformly distributed through out the pipe cross section, so "up" realy has no bearing on the question; further, my sensor is not mounted at the top of the pipe anyway, clearances dictated that I mount it on the side, angled up slightly so not to trap water. Have you actually tried flush mounting a sensor before or are you speculating?..because I can speculate as well as the next guy. I am looking for a real reason why it would not work well, flush mounted, or maybe mounted just slightly proud of the wall ID. I'm thinking it might have more to do with heating the sensor, but that could be addressed with insulation around the pipe and sensor mount. I know there is probably a good reason why the factory engineers stick the sensor way into the exhaust flow, at the expense of increased back pressure, and I would like to find out what it is.
Could it just be "tradition"?.. "That's the way we've always done it"?..... It happens!

kp

It has to be immersed in the exhaust flow to get an accurate reading. If it was flush mounted to the id of the pipe, the volume of gases around the sensor would not be "replaced" with new gases quickly enough to accurately run the engine.

I usually assume that every decision Honda has ever made is strongly backed by extensive r&d and scientific evidence.

99.9% of the time messing with Honda engineering just creates problems.

ddd4114

There isn't a huge impact on flow from the oxygen sensor. Yeah, it's measurable, but it's not going to be a night and day difference. It especially won't make much difference while cruising.

If you're that concerned, I'd monitor the oxygen sensor voltage during a few transient conditions and see what happens. Then, shim the oxygen sensor until you notice a delay in the reading or some other inaccuracy. Just keep in mind that the boundary layer thickness is going to change with load, speed, and temperature, and you don't want to have it on the edge of giving you an inaccurate reading at a different loading condition.

Regarding the resistor idea, I still don't get how that would benefit you. With a narrowband oxygen sensor, a high voltage indicates a rich mixture, so if you're reducing the voltage measured by the ECU, wouldn't it enrich more than normal?

DCFIVER

Try it. I already have. Years ago. Not for the same reason as you, but it didnt work none the less. The O2 will code. When the O2 coded on me, I set to find out why.
Your understanding of gas flow is wrong.I cannot begin to explain computational fluid dynmics to you as my understanding of the subject is vague. What I do know is how it applies in the case of gas flow through a pipe. Moving the O2 out of the direct path of the exhaust stream and sitting it flush with the ID of the pipe will move it into a zero/low pressure zone. You will get very little to no exhaust gas there.
If i were speculating, you would at least have the benefit of 15 years experience to support the speculation. And some research into the phenomena of gas flow. But I am not. Twice in my career I had to diagnose a slow O2 response code on brand new O2 sensors. In both instances the Original O2 threads were damaged and a DIY'er tried to weld their own bung in a different location and on top of the exhaust pipe rather than in. Replacing the pipes with OE fixed both problems with out replacing the O2's.
I gave it to you. You may also want to look into O2 sensor defoulers and how they work. Same principle. No. That was true of very early O2 sensor systems, but for the last 20 years or so O2 heaters take care of that job.

ramco

ddd4114, I thought it was the other way around; but if you're right with the high voltage-is-rich relationship, then my idea will not work using a resistor.
I understand that the exhaust gasses flow faster through the middle of the pipe than near the perifery, and normally I wouldn't be concerned with mounting the sensor an inch into the pipe, but I will be adding a wide band right next to the narrow band, or in line with it, which increases the restriction. If I was using 2.5" tubing, there would be no issue, but our pipe is only 2", so it could make a signifigant difference. Also, consider that the smaller diameter also increases the gas flow speed, so when comparing to the 2.5" pipes, that a lot of tuners run, I would bet that the flow past the tip of a sensor, inserted say 1/4" into the smaller 2" pipe, would equal the rate of flow around a sensor fully inserted into a 2.5" pipe. I am, and have been assuming, that the reactive part of the sensor is at the very tip, and not along the sides of the stem.

ddd4114

Yeah, the Nernst potential is highest when the exhaust stream is devoid of oxygen (full rich). It's lowest in free air (full lean).

ramco

Thanks for your comments. When I get around to buying/installing the wide band, I think I will try pulling the sensors out considerably, and see what happens. I will post results, one way or the other.