How do I go about making a circuit board for averaging out 2 Oxygen sensors (input) into one output?

Thread Starter

judgedredd

Joined May 21, 2024
8
I have a BMW E36 320i, of which there are also 4 other engine capacities for the same 6 cylinder engine (323i, 325i, 328i and M3). The 318i is a four cylinder engine and the exhaust is unique to it.

The 320i and 323i have a single pipe from the 2 downpipes of the exhaust headers (via a 2-into-1 junction) all the way to the muffler in the rear. These two models have 1 oxygen sensor (also called a lambda sensor).

The rest of the models have two pipes - one for each exhaust header (each exhaust header has 3 exhaust ports per header and all models from the 320i up have two exhaust headers) that go all the way to the rear muffler.

There are 4 oxygen sensors on these double pipe models - one for each header and one for each catalytic converter (that screws in just before the CC).

I have the possibility of getting a used exhaust system that used to be on a 328i which has 4 oxygen sensors. I would not be using the exhaust headers from the 328i just the exhaust system from the down pipes back, which includes the 2 catalytic converters. So I would only have 2 oxygen sensors to deal with.

My thinking is that somehow I can use the 2 oxygen sensors that are plugged into a box that has a circuit that averages out the two oxygen sensor readings (the inputs) and the output would then be passed on to the car's computer via the regular plug that the sensor normally plugs into (I'm surprised that there isn't a ready-made setup as this would be of benefit to lots and lots of car enthusiasts).

I haven't researched yet how many wires the oxygen sensors from the 328i have or, for that matter, the 320i. But just wanting to know how simple/feasible this would be and whether my thinking is sound on this. I understand that the flow rates of exhaust gases may be different for the different engine capacities but the sensors measure the oxygen content via voltage fluctuations. Hence my post here.
 
Last edited:

Thread Starter

judgedredd

Joined May 21, 2024
8
Having done some more research, the oxygen sensors for the 320i and the 325i, 328i and M3 all have sensors that have 3 wires.
 

MrChips

Joined Oct 2, 2009
31,059
This is not my area of expertise.
Why not just use one oxygen sensor?
You could bridge the two sensors into one using two resistors (like they do when converting stereo audio into mono).
 

Jerry-Hat-Trick

Joined Aug 31, 2022
588
My understanding of 3 wire lambda sensors is that one wire is the sensor output, one is the sensor ground and one is the heater which uses the body of the sensor for ground. These sensors are presumably binary which means that their output swings sharply from just under 1.0 volts when the mixture is rich to just over zero volts when the mixture is lean. The ECU tries to keep the average voltage to 0.5 volts

You could possibly add two together with an op amp and halving the output but I don’t think you will get a better result than simply using one of the sensors and ignoring the other
 

Thread Starter

judgedredd

Joined May 21, 2024
8
Thanks for your replies Jerry-Hat-Trick and MrChips.
I have thought of using just the one and might even do that, though reading various threads on car forums it doesn't seem to be as easy as that. Hence this post.

One other thing... I have not done electronics at all even though I have fixed things like a TV and a Sony Walkman-like device (handheld tape player/recorder by Akai, back in the 80s).

I did do some research via the 'net and found formulas for averaging circuits, which were diagrams, but nothing on how to build or make one. I do know of a device that is put in-between the Oxygen sensor and the sensor plug that keeps the voltage constant at 0.5v. So maybe I will try this route too. Though knowing myself, I would want to do it properly and take two readings, average them out and send the one output to the computer.

So according to the posts here there are three ways of doing this:
1. Use 1 oxygen sensor and blank off the other hole
2. Bridge the 2 sensors
3. Use the item that I know about

The cheapest and most convenient option is #1, though I have no idea if this would work. $0 cost (I would use the O2 sensor I already have)
The next easiest option is #3. This costs $28
Option 2 is where I have to learn how to make one of these items, learn and understand the nomenclature and source the parts and understand how to read a wiring/component diagram. Definitely the most challenging option.

Thank you all for your input.
 

LowQCab

Joined Nov 6, 2012
4,290
These are the reasons that this plan will fail-miserably ................

#1)
You don't know why this is a bad idea.
This means that You don't have enough working knowledge of your System
to make any changes to it that won't adversely affect the operation of the Engine.

More reasons ..........

These are most likely "Narrow-Band" Oxygen-Sensors, in fact I would bet money on it.

The Output-Waveforms, Frequencies, and Voltages, are unique to "Narrow-Band" Oxygen-Sensors.

Multiple Narrow-Band-Oxygen-Sensors can NOT be mixed together under any circumstances.

The Outputs are a roughly "soft" Square-Wave which also varies over a rather narrow Voltage-range,
with the Frequency varying by Temperature, Engine RPM, and other Engine running conditions.

The Engine-Computer is set-up to interpret these odd Waveforms and Voltages in very particular ways.
Mixing 2 of these Signals, while it is quite simple to do,
would only serve to trash your Engine's driveability, and give You a Check-Engine-Warning-Light.

You don't need, or want, 2 Narrow-Band-Oxygen-Sensors.

The only exception to this would be when installing an Aftermarket-Engine-Computer
which has multiple Inputs designed specifically
for much more useful WIDE-BAND-OXYGEN-SENSORS.

I personally would recommend an Aftermarket-Engine-Computer,
just because You are getting into swapping important parts that can easily trash your Factory-Tune,
BUT,
there is a rather steep-learning-curve to learning to Tune an Engine,
and having somebody else install it and tune it for You can be brutally expensive.
Expect your Car to be down for at least a week if You do this.
Tuning for "Part-Throttle-Driveability" can easily take over a month of tweaking.
.
.
.
 
Last edited:

Jerry-Hat-Trick

Joined Aug 31, 2022
588
Reading again your original post perhaps I can ask to clarify? You say that the 320i has only one lambda sensor but surely there is a second one that is positioned downstream of the catalytic converter? And I’d expect the twin pipe exhaust system to have two sensors for each pipe, one in each exhaust header, one after each catalytic converter, not before the CC?

Also, perhaps you could post a link to the device that outputs 0.5V? I’m not understanding the purpose of it
 

Thread Starter

judgedredd

Joined May 21, 2024
8
Jerry-Hat-Trick,

On the 320i there is only ONE sensor before the CC. See this link for more info: https://www.realoem.com/bmw/enUS/partgrp?id=CB13-USA-05-1993-E36-BMW-320i&mg=18 - this is a 1993 vehicle, just after most car companies switched over to computer controlled fuel injection.

The device I mentioned is for those who are seeking to maximize distance travelled per gallon. I have been trained to run an engine on any liquid including water but with computer systems and oxygen sensors now being used in fuel injected engines (as opposed to carburetors) their real purpose is to make it extremely difficult to do that. Hence the following:
https://eagle-research.com/product/electronic-fuel-injection-enhancer-efie-device/
 

LowQCab

Joined Nov 6, 2012
4,290
There is only one thing that will substantially increase the Fuel-Mileage,
increased Compression-Ratio.
( and to a lesser extent, Combustion-Chamber design, shape, and clearances )

This is why modern Cars are coming from the factory with as much as a ~13 to 1 Compression-Ratio.
There are a ton of tricks used to prevent Engine-destroying DETONATION at these high CRs,
one of the biggest tricks is Direct-Injection of fuel into the Combustion-Chamber.

You will need an aftermarket Fuel-Injection-Computer to maximize Fuel-Mileage,
especially because Narrow-Band-O2-Sensors can NOT accurately detect Lean-Fuel-Mixtures.
.
.
.
 
On the 320i there is only ONE sensor before the CC - this is a 1993 vehicle, just after most car companies switched over to computer controlled fuel injection.
Thanks for coming back to me. I understand that the E36 320i has one sensor fefore the CC, usually .close to the engine, but I'm surprised if there isn't one after the CC. I think the two pipe system will have a total of 4 sensors as you describe, one for each pipe near the engine and one for each pipe after the CC.

The device you have in mind is actually doing something along the lines of your suggested adding of two sensor signals. The way that binary (or narrowband) sensors work is that they will swing to about 0.8V output when the mixture is rich and 0.1V when the mixture is lean and the ECU is constantly adjusting the mixture to achieve an average of 0.45V. If the sensor and the ECU are working correctly the switching between the two states should happen every 100 to 200 milliseconds. It seems that the device simply adds the sensor signal to a small voltage generated by the device to trick the ECU into leaning the mixture to maintain that average of 0.45V. They suggest adding 0.35V for the initial set up which will probably make the mixture far to lean. I certainly wouldn't use one on a car I care about.

If you plan to go ahead with the twin pipe exhaust then I'd first try connecting only the sensors from one pipe to the ECU and checking the emissions from that pipe which should be within specification and then check the emissions from the other pipe - you may find that they are also within the limits! Alternatively, connect the sensor near the engine from one pipe and the sensor after the CC to the ECU and see how that affects emissions.

Are the ECUs different for the single pipe and twin pipe exhaust? There are proably experts on BMW forums that will know more about this.
 

Thread Starter

judgedredd

Joined May 21, 2024
8
The system that I was trained on is called GEET (Global Environmental Energy Technology) by Paul Pantone (deceased). But this topic is not for here.
 
I may get shot down in flames by those who have far more expertise than me, but I do think you could try averaging the value of two narrowband sensors using a dual rail to rail to rail op amp like the MCP6002. Essentially, buffer each sensor output with one op amp each and tap the join of two equal resistors each connected to the op amp outputs. You'll need something like an LM805 linear regulator to supply the op amps with 5.0V
 

geekoftheweek

Joined Oct 6, 2013
1,250
I don't know European law or anywhere else for that matter, but in the US most cars built before 1996 only had a sensor before the catalytic converter. In 1996 the OBD2 standard was enacted which required one after the converter.

If you only have one sensor you only need one sensor. Plug the other holes. The early injection systems weren't as precise as they are now and were designed to run off more of an average than individual cylinder readings.

Trying to combine multiple signals and such is both pointless and going to be more of a headache than it is worth!!
 

LowQCab

Joined Nov 6, 2012
4,290
The Computer continuously monitors the frequency of the "Cross-Counts" and the PWM%.
If You mix 2 Signals of this type, the Computer will not be able
to determine what the conditions are that it is supposed to be continuously measuring.

There is no getting around this fact,
that's the way that Narrow-Band O2-Sensors work.

When tuning for maximum Fuel-Mileage, You MUST have Wide-Band-O2-Sensors.

If poor Driveability is not a concern,
in other words, this is a "on-the-road" contest with
other specially prepared Cars that are definitely NOT "Daily-Drivers".

There are cheezy ways to cheat at the game.
1) Disconnect the O2-Sensor and Ignore the resulting Check-Engine-Light.

2) Record the Fuel-Rail-Pressure with the Vacuum-Line disconnected from the stock Regulator.

3) Purchase an Aftermarket Adjustable-Hot-Rod-Fuel-Pressure-Regulator,
replace the stock Fuel-Pressure-Regulator with the new adjustable-Regulator.

4) Keep reducing the Fuel-Rail-Pressure until the Engine starts to have noticeable Driveability-Problems.
Don't go any lower in pressure than this, "You are there" as far as the "ideal-mixture" is concerned.
If You have Bosch K-Jetronic Mechanical-Fuel-Injection, ( I don't remember what Year they dropped it ),
You can adjust the overall Fuel-Mixture almost anywhere that You want it.

5) Strip the Car, be brutal about it.
If it's not absolutely necessary to making the Car go down the road, then REMOVE IT.

6) Do a special Front-End-Alignment-Job with "Zero-Toe-In"

7) Run ~60-PSI of Air in the Tires, ( regardless of the warnings on the Sidewalls of the Tires ).

8) Do an Oil-Change and refill the Engine with only 3-Quarts of Mobile-1 "0w8" Full-Synthetic-Oil.
( 0w16 is also acceptable if You can't find 0w8 )

9) Flush the Cooling-System with straight Coolant, NO WATER, NO "Pre-Diluted" Coolant.
After this, You should run the Car for about a week WITH THE PRESSURE-CAP REMOVED,
to let all the residual Water evaporate.
You will have to periodically top-off the Recovery-Bottle with more straight-Coolant until all Water is gone.
( even straight-Coolant is around ~30% Water, You want zero-Water ).
Cycle the Engine from Cold to Hot, Hot to Cold, Cold to Hot, as many times as is practical.
The Pressure-Cap should NOT be replaced, leave the Recovery-Bottle "open".
After this procedure the Engine can be safely operated, ( at low-Loads only ),
at up to ~300 degrees F with no fear of damage or Detonation.

After this procedure the Cooling-Fan should be removed.
Watch the Coolant-Temperature if the Engine will be
Idling for more than a few minutes after it's thoroughly warmed-up.
A manually-controlled-Electric-Cooling-Fan would be a good idea, but is not absolutely required.
Do not run the Electric-Cooling-Fan except when absolutely necessary,
it puts a substantial Load on the Alternator, and therefore, the Engine,
but in any case, it will be far more efficient than an Engine-Driven-Fan.

10) Get the Car thoroughly warmed-up right before any contest or official-measurements.

11) Acceleration is the enemy, the Throttle should limited by some kind of an adjustable "stop"
to prevent it from opening more than a few degrees.
It should only be allowed to open enough to very-slowly, and finally,
reach the top-speed required by the Contest-Rules.

12) Find a Gas-Station that sells "Ethanol-Free-Gasoline", and use this type of Fuel exclusively.

These are all of the things that You can do without
disassembling the Engine and making substantial internal-modifications,
and installing an Aftermarket-Engine-Computer.
Being able to alter the Ignition-Timing-Map is vital to getting the most efficiency out of the Engine.
Learning exactly how to do that takes a serious amount of experience, and preferably, a Dynamometer.

If tests will be conducted at speeds exceeding roughly ~45-MPH,
a smooth Plastic-Belly-Pan under the entire Car, plus, flush-fitting, flat, Wheel-Covers will help.
Also remove the Rear-View-Mirrors, and keep all Windows rolled-up.

All of these "tricks", when done together, "might" result in a ~15% to 20% Fuel-Mileage increase, maybe.
( Nothing can keep your Foot off of the Gas-Pedal except You ).
But, simply raising the Compression-Ratio of the Engine to around ~14-to-1
will result in a solid, and easily noticeable, ~15% improvement in Fuel-Mileage,
all by it's self,
and with no other changes.
It will also result in approx. ~5% Power increase, evenly, at all RPMs.

Special-Spark-Plugs,
Fuel-Additives,
"Turbo-Air-Spinners",
Special Air-Cleaner-Elements,
Oil-Additives,
are all 100% BS when it comes to improving Mileage.

If your Engine has a poorly designed Combustion-Chamber,
a 500-Volt-DC "Plasma-type" Ignition-Box "may" make a very small improvement, but don't bet on it.
.
.
.
 

geekoftheweek

Joined Oct 6, 2013
1,250
To expand a bit on my previous reply...

Unless you have head gasket issues, or something seriously wrong that causes one side of the engine to run different than the other having a separate sensor for each side isn't going to do any good. There won't be enough of a difference from one side of the engine to the other to matter. I highly doubt there is any real difference even today... multiple O2 sensors is probably more of an emission standard requirement than performance requirement.
 

Thread Starter

judgedredd

Joined May 21, 2024
8
Thanks geekoftheweek and LowQCab for the responses. LowQCab, you seem to have an awful lot of knowledge around cars! Thanks for the extensive info. As to that exhaust that was available, that has now gone, but the info gleaned from this thread is great. More than enough to know what to do next time.
 
Top