How would a twisted pair effect a signal

[Solakian]

I'm working on a 1990 Ford electronic engine control (EEC IV). Compared to today's vehicles, the ECU (electronic control unit) is fairly simple. I've been monitoring the O2 (oxygen) sensor signal and determined that the engine has been running lean. There are no diagnostic codes. The inputs to the ECU are O2 sensor, Manifold pressure, Atmospheric pressure, Coolant temp, Intake air temp, Throttle position & RPM. I planned to modify the O2 sensor signal to change the fuel injection command.

With the battery disconnected, I cut the signal wire from the O2 sensor at the ECU. I ran a pair of wire through the fire wall connecting one end of each to each side of the cut signal wire (from the sensor / to the ECU). I intended to add a circuit to modify the signal inside the cab. The pair of wires I used (for no particular reason other than they were available) happen to be a twisted pair of 16ga primary wire, typically used for datalink. Since disconnecting the battery causes the ECU to loose it's learned fuel map, I wanted to let the ECU re-learn before adding my circuit. I simply connected the pair together in the cab. At this point, all I had was an additional 3' of wire length added to the O2 sensor signal wire, however the two 18" wires that made up the 3' extension were twisted together.

The result has been surprising. The engine no longer runs lean. I have run it long enough to allow the ECU to re-learn the fuel map and it now runs stoich (ideal). The only other change was the removal and reconnecting of the ECU connector.

An O2 signal typically swings from .2 to .9 volts DC (.2 is the presence of oxygen / .9 is the absence of oxygen) at a relatively slow frequency of about 40 to 110 Hz. The ECU adjusts the fuel pulse to maintain a constant dithering. Does the twisted pair somehow change this input signal?

The only other possibility I considered was re-seating the ECU connector may have cleaned some connections, changing the signal values of some or all of the signals, most of which are simple analog inputs.

 

[Ylli]

Or perhaps the ECU simply needed to recreate the fuel map.

 

[Solakian]

I suppose that's another possibility

 

[dave9]

The result using twisted pair wires is just coincidence. The system is set up to not have susceptibility to noise introduced from the environment, for example mainly the ignition, and can run straight non-twisted, non-shielded wire. Any noise rejection that wasn't needed, is anecdotal. That's a method used on low level digital signals, not a simple analog voltage feedback.

Is there some very unusual situation with this vehicle, trying to run the wrong ECU for the engine?

Normally what you are attempting is the wrong way to do this, and potentially illegal to circumvent the emissions controls.

It is probable that disconnecting the battery lost the volatile memory and it must go through a set of driving cycles to relearn parameters, then it will run lean again. Idling for a long time may not be enough.

You should not be trying to alter the O2 signal. Fix the problem instead. Typically running lean is from one of a few factors:

- Bad sensor

- Fuel delivery problem (pump, power to the pump, or strainer, or filter, or fuel pressure regulator (or vac line to it), which you could determine by low pressure at the fuel rail, or evenly clogging fuel injectors not bad enough to cause a misfire yet (or does it misfire?)

- Vac line leak

- Intake manifold or injector o-ring leak, or anywhere on the intake (typically after the first sensor)

- Severe valve carbon buildup (if it only happens initially after start, till the carbon is saturated with fuel)

- Unintended EGR flow at idle

- More. Lean is the most common fault out there on old vehicles.

Remember that just getting it to the right mix idling in your garage, is not an assurance it will continue to do so at all engine loads and temperatures, and if you stay too rich you may end up fouling your catalytic converter and O2 sensors and then have multiple problems to fix, as well as fairly dismal fuel economy.

It is possible you just had a wiring or connector fault and *disturbing* the wires and connector has resolved it. Certainly this is not uncommon on vehicles so old, but generally speaking there are many more likely causes and it wouldn't be surprising if this lean condition resurfaces... just as old age can cause wiring or connector issues, it can also, more often, end up causing poor fuel delivery or unmetered air getting in.

 

[Solakian]

Thank you for you response. This vehicle is a 1990 model, during which time a lot of effort went into intentionally running lean. The thought process was if you have less fuel in the mixture, you will gain mpg. It didn't turn out to be true. I understand the implications of changing the fuel map, and I understand that my "experiment" is not the normal way to accomplish my goal. The vehicle is extremely well maintained and all sensors, vacuum lines, emission controls, etc are in good working order. I'm working on a theory, and this is what I choose to test it on. The result I got prior to making any modification (other than the addition of the O2 signal wire extension) surprised me.

I'm just wondering if anyone can confirm (or refute) that twisting the wires on my extension could alter the signal, and if so, how would it alter the signal. It doesn't seem like the twisted pair should have an effect, but it if does, I need to understand it before I change the circuit.

 

[Audioguru]

You gain only a little more maximum power by polluting the atmosphere by running rich.
How often or do you ever drive at maximum power??

 

[Solakian]

Thank you for your response. I'm really looking for someone with knowledge about twisted pairs.

 

[Alec_t]

I ran a pair of wire through the fire wall connecting one end of each to each side of the cut signal wire (from the sensor / to the ECU).

For AC/radio-frequency circuits a twisted pair can be advantageous as it enables cancellation/reduction of some undesirable effects. However, there is no such advantage when, if I understand you correctly, the two wires of the twisted pair are in parallel. The additional resistance presented by the extra 3ft of wire is unlikely to have any (significant) effect on the sensor voltage measurement, given the likely high impedance of the ECU input.
I agree with the others, that a poor connection somewhere may have been improved by unplugging/plugging things.

 

[hevans1944]

I doubt whether the twisted pair of 16 ga. wires has any causal effect on what you are seeing. Poor connection continuity, temporarily remedied by disconnecting and re-connecting the affected connector or connectors, is a more likely cause. Why not test this theory by removing the twisted pair and replacing the pair with a single wire of appropriate gauge and length, say 22 or 24 ga. and 36 inches?

If there is no change, you can stuff this new wire through the firewall and proceed with your "experiment" of farkling with the ECU O₂ feedback input. If the O₂ sensor is a typical Bosch sensor, it requires warming up to a certain operating temperature provided by the exhaust manifold before it produces a reliable, low-level, output signal. After that, if the system is operating in a normal "closed loop" fashion, the O₂ sensor output will cycle between rich and lean as the ECU varies the fuel-air ratio in an attempt to maintain a stoichiometric ratio. The rate at which this cycling occurs depends on thermal time constants of the O₂ sensor and you have no control over that, but it is slow enough that passing the signal through a twisted pair of wires will have no effect on the signal. However, resistance in the connections is likely to have a huge effect because the signal is so small.

 

[Martaine2005]

It doesn't seem like the twisted pair should have an effect, but it if does

The process of twisting pairs is to limit 'crosstalk' on data lines and limit EMI (noise).
It will have no effect on what you are doing as all the members here have already told you.
Try a pair, try a twisted pair.. You will see no difference. Now, high resistance is another issue. But again, as already explained to you, bad connections cause immeasurable problems.
My advice would be to separate all connectors, spray a little contact cleaner and connect/re-connect a few times. Then see what happens. You will probably find that all answers given in this thread are correct.
Try it out for yourself.

Martin

 

[Solakian]

Alec T, Heavens1944, and Martain2005, Thank you very much for your comments (I had to look up "farkling). I suspected that the twisted pair had no effect on this circuit, but needed some expertise to confirm. I shall move forward with my farkling.

 

[Martaine2005]

I too had to look up 'farkle'. @hevans1944 always does this in most of his posts.
I have read the dictionary more times than 'The Art Of Electronics' thanks to his posts...
He is stalling me.....

Martin

 

[hevans1944]

I have been using farkle for many years to mean any activity based on obtaining random results through experimentation, without any underlying basis of understanding of what is cause and what is effect. Fun way to pass idle time, I suppose.

IIRC, I first heard it used in 1960s, way before I ever rode a motorcycle or ever heard about adding "bling" to a motorcycle, which apparently is how the word is commonly used today. It just sounded like a "kool" word whose meaning could be inferred from the context of its use. Might even have picked it up while serving as an enlisted man in the U.S. Air Force. Us troops liked to pickup and use new-fangled words like farkle and fubar and snafu.

My apologies to Martin and Solakian for any confusion that may have resulted.

 

[Martaine2005]

Hops , you are incredible.
your knowledge with the English language and your knowledge of electronics....I read every post you write.
You never confuse, just takes a while to 'Google' what you mean..

Martin

 

[Audioguru]

A twisted pair is used for telephone lines to cancel common-mode interference. The twisted pair is completely balanced with a signal on one wire and the inverted signal on the other wire without either wire grounded like it would be on a car.
On a car a twisted pair would produce only a little amount of shielding unless you make both wires have a balanced signal.

 

[Martaine2005]

There is definitely a twisted pair or two in my car tonight....Geeshh....Go farkle.

 

[hevans1944]

On a car a twisted pair would produce only a little amount of shielding unless you make both wires have a balanced signal.

...Which is not a bloody likely scenario in automotive applications, where all signals are referenced to the almighty chassis ground. Still, where it IS effective, twisted-pair is an economical solution to common-mode noise. As @Audioguru points out, Western Electric and AT&T have been using it for years, with one hundred pairs of (mostly) non-interfering twisted-pair lines in a single trunk cable.

And then along came CAT 5 twisted-pair flat cable to replace bulky and hard-to-terminate coaxial 10 MHz ethernet cables, followed a little late and eventually by CAT 6 shielded twisted pair flat cable. The phone company is not exactly known for quickly embracing the latest technology. Were it not for the Princess telephone, we would still be using round telephone cords with screw terminations.

Sound techs and their engineers have long used shielded, twisted-pair, heavily rubber insulated and rugged microphone cables for reliable, low-hum, stage connections, along with Amphenol XLR three-pin connectors. I like this system although it is a real PITA to get it soldered up correctly. In my shop, I generally use twisted-pair red and black wires for power supply connections when I have the right gauge and color available. Not sure why I continue to do this. though... the power supply output is low-impedance (generally) so not much chance of picking up significant interference on the connecting wires. Maybe I continue to build this way out of habit, or because I've accumulated a helluva lot of spools of wire over the years, or because it looks pretty neat.

 

[dave9]

Thank you for you response. This vehicle is a 1990 model, during which time a lot of effort went into intentionally running lean.

False. Computer control was just the beginning of more fuel efficiency. They not only did not run too lean, there were many events in which they could have run leaner, these events being refined and compensated for further over the next couple decades in firmware along with mechanical changes.

I've owned that era vehicle and did not find them running too lean in a properly maintained state. It is as if you are trying to pretend the entire era had people with constant too lean conditions from the factory when this is certainly not the case.

The thought process was if you have less fuel in the mixture, you will gain mpg. It didn't turn out to be true.

Yes, it was in fact true. How you can dismiss the gradual increases in fuel economy is beyond me. Have you no memory of what vehicles achieved in MPG before this era of computer controlled ignition and injectors? Average fuel economy went up at least 60% for similar vehicles, compared to carbureted *dumb* predecessor vehicles.

I understand the implications of changing the fuel map, and I understand that my "experiment" is not the normal way to accomplish my goal.

You have not stated your goal, only a false theory with no goal. Never has it been stated what you hope to gain, except that you seem to think you must correct a problem that doesn't exist in a properly working vehicle of that era.

Do you realize that these especially lean parameters are usually sacrificed when at full throttle? They are not your enemy except if your primary problem is too fat a wallet from not paying enough to fuel it. Granted, you have not mentioned particulars of the vehicle make/model/engine, I am only generalizing and specific vehicles could have their own quirks and flaws.

The vehicle is extremely well maintained and all sensors, vacuum lines, emission controls, etc are in good working order.

This does not make sense. It is as if you declared it instead of checking everything, while never declaring any problem you seek to solve. If it is running too lean, you do in fact have a problem with one of the above.

If you are trying to turn a 30 year old vehicle into a race car, modifying the O2 signal alone is just not worth the bother and not even needed because a new tune firmware would take care of that and is needed anyway.

I'm working on a theory, and this is what I choose to test it on. The result I got prior to making any modification (other than the addition of the O2 signal wire extension) surprised me.

I can appreciate testing a theory, but only if you follow the time tested expert advice formulated over the decades of people repairing these vehicles. This is like a re-hash of existing knowledge, not some new frontier.

I'm just wondering if anyone can confirm (or refute) that twisting the wires on my extension could alter the signal, and if so, how would it alter the signal. It doesn't seem like the twisted pair should have an effect, but it if does, I need to understand it before I change the circuit.

I fully realize this, yet you should be told that you're far deep into a rabbit hole that has no other exit than backtracking the way you came and finding out another way to get to whatever result you seek.

That era of vehicle did NOT have problems running too lean (in general). Common ways to improve power were then, as now, a turbocharger, or supercharger, improving air intake (and exhaust), a more aggressive cam and ECM tune, removing things belt driven, removing weight, etc, etc.

You might have some unique vehicle design flaw, but what you do not have, is some generalized era-rampant issue that is fixed by defeating the emissions controls with a fixed resistance or circuit break when a dynamic feedback to the ECM is needed.

Again, you're just going to waste fuel and potentially damage your O2 sensors and cat. Fix the real problem instead. It does NOT run too lean by design. It is always going to be some part degradation. No manufacturer can sanely market a vehicle that is undriveable new from the factory. Whatever has changed since new, there you go, something is not "in good working order".

 

[Audioguru]

Before old computerized cars were made, most drivers did not bother replacing plugged air filters and therefore drove around running rich making black exhaust air pollution. Therefore maybe the computer was programmed to make the cars run lean to "fix" that problem.
Maybe in those days they did not know that lean made nitrogen oxide pollution?
I remember when car commercials mentioned "lean burn" to reduce air pollution and increase fuel economy. Maybe they lied.