Time to close the O2 loop

[Richard123]

Does anybody have any idea what all causes the ECU to go closed loop and if it can be manipulated??

My 1994 S4 used to close the loop in about 45 seconds but is now waiting about 4 minutes for a coolant temperature of 56.8C to be achieved.  I put in a new O2 sensor (13030) and checked the 12V heater circuit, no codes thrown, drives fine though it idles up (say 2000 rpm) when rolling clutch in, then idles back down (1200 rpm) when the car stops, until warm.  The O2 sensor still hits approximately 0.9V by 45 seconds.  I tried the battery disconnect to no avail.  I will try resetting the BLM's via Freescan, but figured maybe this was a familiar thing?  (The car has never had the exhaust backpressure valve since I've owned it, and I've blanked off the throttle jack, so neither function.) 

[drdoom]

From what I have read O2 closed loop with cold engine is altogether down to balancing emissions control with driveability. Catalyst requires light up time, though more like your familiar 45 seconds when cat is sound. Sounds like some sensor is spoofing the ECU with readings outside optimal, so causing the symptoms. Worth considering by what means a temperature sensor could be offering a false reading, or possibly something systemically out of sorts rendering a legitimate reading out of range. Eg: something in the cooling system not altogether right, as with an air pocket impairing the sensor function. Just an amateur 2p's worth.  

[RobinB5]

I'd also like to know the exact conditions that lead to the ECM deciding to go closed loop. I believe that the Zirconium O2 sensor needs to warm up to ~300degC to provide readings indicative of O2 levels. Its internal heater should do this, and heat up would be accelerated by the hot exhaust gases flowing over it. What I don't know is whether the ECU assumes that the sensor has reached the required temperature, or whether it senses it, and whether this is the only condition the ECU decides to go closed loop.

From what I've seen of Espritmon logs, when all is well, as soon as it goes closed loop, the O2 sensor voltage wobbles all over the place. Prior to that its voltage varies much more slowly. Maybe because before it gets up to its activation temperature, it's simply a (temperature sensitive) resistor, after it heats up enough, the fancy sensor materials do their thing and start to report O2 (partial pressure) levels via their voltage (under I assume a constant current condition?).

Does the ECU know the O2 sensor is 'ready' prior to going closed loop? Or does it wait until other sensed values are at a level where it assumes the O2 sensor is hot enough to accurately report O2? Why does the signal wobble all over the place (good O2 readings) at the exact point when it goes closed loop? I've no idea!   @sailorbob might!

[Richard123]

2 hours ago, RobinB5 said:

I'd also like to know the exact conditions that lead to the ECM deciding to go closed loop. I believe that the Zirconium O2 sensor needs to warm up to ~300degC to provide readings indicative of O2 levels. Its internal heater should do this, and heat up would be accelerated by the hot exhaust gases flowing over it. What I don't know is whether the ECU assumes that the sensor has reached the required temperature, or whether it senses it, and whether this is the only condition the ECU decides to go closed loop.

From what I've seen of Espritmon logs, when all is well, as soon as it goes closed loop, the O2 sensor voltage wobbles all over the place. Prior to that its voltage varies much more slowly. Maybe because before it gets up to its activation temperature, it's simply a (temperature sensitive) resistor, after it heats up enough, the fancy sensor materials do their thing and start to report O2 (partial pressure) levels via their voltage (under I assume a constant current condition?).

Does the ECU know the O2 sensor is 'ready' prior to going closed loop? Or does it wait until other sensed values are at a level where it assumes the O2 sensor is hot enough to accurately report O2? Why does the signal wobble all over the place (good O2 readings) at the exact point when it goes closed loop? I've no idea!   @sailorbob might!

I think I know this one.  The O2 signal climbs from about 0.45 volts when cold up to about 0.9 volts when hot.  The high reading is because the voltage output is higher when there is no oxygen (rich).  Once it goes closed loop the ECU leans the mixture (less gas) until voltage drops (it senses oxygen), then immediately richens mixture until the voltage increases, then leans it, etc.  The signal wobble is produced by leaning out every time the O2 reads high, and richening every time the O2 reads low.  It's best to watch on an analogue voltmeter, the needle acts like a metronome!  

My issue is that I have a nice hot O2 sensor putting out a solid 0.9V and an ECU that is reluctant to close the loop, so certainly something else is at play.  I'm pretty sure engine temps plays a part in my car, but not in others' Esprits and I don't know why.

[Barrykearley]

If the o2 trace ain’t wobbling - the sensor is knackered. 

@RobinB5 I’m sure closed loop is based on coolant temp

[drdoom]

Agree with the lads points above, would add that the ECU is programmed with quite a number of maps from which it correlates things like O2, coolant, throttle position, manifold absolute pressure, intake air temperature and other sensor inputs all for the resolving of when to fire spark and what size squirt of fuel to deliver for a given cycle. Some maps are baseline, the one for fuel being set as suitable for a fresh engine working at normal running temperature as read by the coolant sensor. There will be maps for compensating both fuel and spark for conditions other than that, and on it goes. The O2 sensor is central to keeping air fuel ratio in the narrow range amenable to maximum catalyst effectiveness, once the ECU refers to its input so going into "closed loop". The oxygen present in the exhaust produces the voltage at the O2 sensor used to inform the ECU on fuel compensation and there is a dithering either side of ideal target as successive,  ever so lean or rich combustion events result in varying exhaust gas compositions . Regulators permit the open loop state for conditions such as full throttle and, more pertinently here, a cold engine. For all the considerable processing power inherent in the ECU one must remember that the thing is dead blind, absent inputs from the sensors. If there's nothing fundamentally wrong with the engine then it's routinely a case of faulty input(s) misinforming the ECU when strange operating conditions manifest.

Cheers

[RobinB5]

I get that once it goes closed loop the ECU takes all sensor inputs, coupled with its own maps, to achieve the stochiometric afr of 14.6, including the ongoing process of 'learning' so as to make adjustments to account for local conditions, drift of characteristics of various parts etc.  When cold no 'real' O2 measurements are available and so it starts off in open mode. Going back to @Richard123's question, what actually triggers going closed loop? The following is a log from mine, it took about 8 minutes to go closed loop (was a chilly day though)! When it did, the O2 voltage showed the wobbly behaviour you'd expect. Coolant had just hit 70, rpm went past 3000, maybe there's a map in the ECU that determines when to go closed loop? And why does the O2 signal wobble about (accurately) at exactly the point it goes closed loop? (and if it doesn't, then on mine I get subsequent 'issues').

@Barrykearley's point about the O2 sensor being knackered if it doesn't wobble... from cold up to a temperature where it's capable of recording O2 levels, it's just a dumb temperature dependent resistor, thus the smooth blue O2 line, with an increasing voltage. Sure, if it never wobbles even when up to temperature, it's fubared.

[Richard123]

Robin,

From the graphs, it seems that your issue is the same as mine.  Your O2 sensor was hot (near max output) well before the ECU closed the loop.  After the loop closed, your O2 sensor "wobbles" just fine, lol.  So I don't suspect a bad O2 sensor.

Yours is a 1995, so it came factory with the exhaust back pressure valve and throttle jack which were intended to allow the cars to go closed loop quicker, correct?  And can I assume they have been removed?  Mine are removed from my 1994 S4, and I am wondering if the 1200 rpm cold idle isn't throwing the ECU off, as it is expecting the throttle jack to produce a much higher cold "warm up" idle??  Might that anomaly cause the ECU to abandon the standard map and decide to wait for a specific coolant temperature before closing the loop?  

Just a theory, but if your EPBV and throttle jack have been removed, it may make some sense.  I might test it by hooking up the throttle jack again and going with a 2400 rpm (or such) cold idle just to see what it will do.

    

Edited June 26, 2020 by Bibs

[RobinB5]

Richard, actually the above log is when all works well  It is the log in the below thread that is when it doesn't. In that case, when it goes closed loop, the O2 sensor does not wobble around like it should, the BLM and Integrator values drop right down (trying to get rid of a phantom over rich condition), then the car stalls. With little experience and too much assumption, I'm convinced it's the O2 sensor maybe not being heated properly. Anyways, new one is on order from PNM, will see if it fixes things!

 

[Richard123]

Ah, fair enough.  I thought your issue was the long time for loop closing.  For interest though, has your exhaust back pressure valve been removed?    

I had one instance where the car would not run closed loop, occurring after I washed the engine seeking the source of an oil leak.  It would sit and idle, but not take any fuel.  Threatened to stall every time I touched the throttle.  I let it cool to get it home, and replaced the O2 sensor with one of those Bosch 13030's.  My O2 4-wire connector is long gone and a couple of the single connectors were more than dodgy, so I suspect getting the wires wet caused the issue though I can't confirm as the issue never reappeared.  I'm betting your new O2 sensor cleans up your issue.

Edited June 26, 2020 by Bibs

[RobinB5]

35 minutes ago, Richard123 said:

Ah, fair enough.  I thought your issue was the long time for loop closing.  For interest though, has your exhaust back pressure valve been removed?    

 

Removed!

[drdoom]

The long 8 minutes taken for Robin's car to go into closed loop strikes me as problematic, expecting that a '95 era car would have been required to get there more quickly though I can't speak to that with certainty. An O2 sensor is more than the sort of heat varied resistor affair that is typical of coolant temp sensors, though it must attain a significant level of heat to function it works via an ion exchange from the surface of some rare earth metal within. Such sensors are quite susceptible to failure by corruption of the core surface due to leaded fuel, anti-wear oil constituents ( eg. zinc compounds ) and even water native to the exhaust stream  in cases where the sensor boss is  wrongly located where condensation will accumulate.

It is reasonable to assume the ECU will have failsafe protocols for when any engine sensor delivers readings out of range, all else being normal, and reversion to open loop state is among the likeliest of outcomes in such cases.