Building a simple fuel controller using wideband input

Thread Starter

MykkH

Joined Feb 1, 2016
4
Hey Guys, I've got an idea in my head I'd like to execute if possible. Wishing to do away with the MAF sensor on my OBD1 Bosch Motronic 3.3 in my '95 bmw 540i. I've got an Innovate wideband O2 sensor and gauge installed in the car and wondering if I could build a circuit that would send MAF voltage to the vehicles computer based on wideband gauge analog output. Essentially it would be a variable 0V-5V output that would would adjust in order to maintain a 2.5v input. I'm starting with 2.5v for 14.7AFR as the test to see if it's even possible before altering input voltages for different AFR's.

As input voltage to the circuit from the wideband increases (leaner) the output voltage of the circuit would also need to increase to richen the mixture and stabilize AFR and thus bring the input voltage down. It would do the opposite to lean the mixture, as input voltage decreases (richer) the circuit lowers voltage to take fuel away. The output would be constantly changing as engine RPM, load, driving conditions change. At idle it would be around 1V to maintain a 2.5v input, driving and light throttle cruise would be 1.3v-3v range.Acceleration and higher RPM would be 3v-4.8v...all in the name of maintaining 2.5v.
Like how a voltage regulator takes a higher and fluctuating input voltage and converts to a solid output voltage, this would be the opposite.

Is such a creature even possible? Thanks in advance for any help.

ys,
 

blocco a spirale

Joined Jun 18, 2008
1,546
This can be done using a micro-controller with an analogue input and output. You would use a look-up table to convert the output from the O2 sensor to whatever the ECU would expect to see from the MAF under the same conditions. The difficulty is understanding what the relationship is between the output from the MAF and the O2 sensor and you would need to know this to create the table.
 

Thread Starter

MykkH

Joined Feb 1, 2016
4
How about the use of an Error Amplifier or Shunt Regulator. Like the TL431?



Maybe add a Cap on the output to vehicle computer to dampen voltage switching?
 

mcgyvr

Joined Oct 15, 2009
5,394
Maybe there is a darn good reason they include an O2 and MAF sensor on vehicles..
Knowing the auto industry is always about "low cost/bean counters,etc..." they wouldn't include both if both weren't needed to properly regulate the engine..

One can remove a MAF sensor (due to its restriction/throttle response) but then you would need to switch to a MAP sensor.

I don't expect the MAF to track O2 perfectly at all.. You need to know speed density information somewhere..
 

Dr.killjoy

Joined Apr 28, 2013
1,196
I wouldn't unless you know what you are doing cause you will cause damage to the engine..
I remember a while ago someone used a resistor or something to mod output signal of maf to get it to run lean or rich... Personal I would get a profession tune or performance upgrades..
 

Thread Starter

MykkH

Joined Feb 1, 2016
4
I wouldn't unless you know what you are doing cause you will cause damage to the engine..
Am I by no means an authoritative on the subject but I did graduate from a vocational college as a master automotive tech. About 10 years ago I started my own PCM tuning business for OBDII GM's as a performance upgrade and I'm brought in as a guest lecturer to the local college to talk about fuel injection computer systems.

Is this similar to a filter? means to control the effect of emissions, like a filter does?
It's not a filter, yet it would have an effect on emissions. But that is not the primary goal.[/QUOTE]
 

cork_ie

Joined Oct 8, 2011
428
1) You are going to have huge swings with the set-up you are proposing, i.e. mix goes rich , "signal voltage" drops , ECU corrects , mix goes lean "signal voltage " rises ECU corrects again. There is going to be quite a lag in between. The car will also be undriveable cold until the sensor actually starts to work.
2) Another problem is when you suddenly snap open the throttle you will not get the normal acceleration enrichment.
3) How are you going to measure the "Signal Voltage" indicating A/F ratio from the wideband O2 sensor? My understanding is that it is pump cell current that is the measure of A/F ratio on wideband O2 sensors and this is extrapolated by the ECU and ouputted as a voltage.There is no direct lambda voltage as in a narrowband sensor
 

Thread Starter

MykkH

Joined Feb 1, 2016
4
1) You are going to have huge swings with the set-up you are proposing, i.e. mix goes rich , "signal voltage" drops , ECU corrects , mix goes lean "signal voltage " rises ECU corrects again. There is going to be quite a lag in between.
I was concerned about that. That would be the major set back. The TL431 is quoted as "lightning fast switching" perhaps experimenting with capacitor sizes could fine tune the swings?

2) Another problem is when you suddenly snap open the throttle you will not get the normal acceleration enrichment.
Acceleration enrichment is throttle position sensor based. As you open the TPS the computer compensates by lengthening injector pulse width momentarily. I was actually counting on this alone to compensate for the time it takes the combustion process to reach the O2 sensor for compensation. The "Accelerator shot" as if it were a carb, is tune-able by millisecond and decay time in the ECU.

3) How are you going to measure the "Signal Voltage" indicating A/F ratio from the wideband O2 sensor? My understanding is that it is pump cell current that is the measure of A/F ratio on wideband O2 sensors and this is extrapolated by the ECU and ouputted as a voltage.There is no direct lambda voltage as in a narrowband sensor
The actual wideband gauge itself is acting as an interface between AFR and voltage. The Innovate has been regarded as having the fastest wideband O2 response time with <100ms reaction time by the time the AFR mix hit's the sensor.

The output voltage from the AFR gauge is also tune-able:



The car will also be undriveable cold until the sensor actually starts to work.
The sensor/gauge has a built in 25 second warm up period before it will display AFR's as it's heating the sensor. This would delay engine starting 25 seconds before fire-up. I was willing to live with that.

The other "maintenance" side of my idea is that the sensor and gauge itself need frequent calibration. All this means is that the sensor is removed from the exhaust stream and then calibrated to free air and replaced back in the O2 bung in the exhaust.
 

slackguy

Joined Feb 11, 2016
76
today's cars, most cars made, all deliver fuel ratio by the ammout of air sucked in

however moisture and density matter, and the MAF accounts for that (its' just 10 cents of nichrome wire, though they may charge $600 for it)

so. humidity sensors are not easy it's not just the O2 to worry about. how accurate is the O2 at widely varying temp and humidity. sure it reacts quickly in a muffler - but that's a different environment (and requires heat to work). is the sensor suceptible to give differing accuracy depending on environment changes? the MAF takes care of that.

i wonder what you are thinking here - ok maybe change the cheap sensor for something higher tech: but O2 sensor is likely not what you are looking for

your not going to get efficiency here as much as if you'd change your chinese injectors for some better atomization and forced air

as for diesel, which many german cars are (yours?), that shouldn't matter as to this topic

better yet take your foot off the gas pedal you'll ruin a nice german car
 

schpenxel

Joined Mar 7, 2016
1
Does that vehicle not already have oxygen sensors? I don't quite understand what you would be trying to accomplish if it does. They should already be driving fueling (via fuel trims) to an AFR of 14.7:1, or, more correctly, to a lambda of 1, whatever that is for the fuel you're using (E10 is more like 14:1)

I know quite a bit about how engine controllers work, but nothing about that specific vehicle.

I've thought about the same before, but only because factory narrowband O2 sensors are only useful for staying around lambda=1. I had an idea to try to use a wideband to have fueling control while richer than that (i.e. at WOT) but never bothered trying to implement it.
 
Top