Open Loop After Start Up.

Status
Not open for further replies.
Joined
May 10, 2005
Messages
2,731
Location
Toronto, Canada
After starting and driving away on a cold engine, the O2 sensors start switching in about 15 secs but the loop status does not change for over 3 minutes.

The photos show the display on my Tech2. The frames are at about 5 frames a second. The first photo shows the sensors start switching. The solid thick line in the second photo shows the loop status change.

The O2 sensors will only switch when the ECM adjusts fuel delivery based on feedback from the O2 sensors, pretty much the definition of closed loop. So it is surprising the ECM does not declare closed loop for over 3 minutes.

100_1273.jpg


100_1271.jpg
 
Doesn't closed loop begin when an engine achieves operating temperature (or a number close to it)? My xB has had weird O2 sensor faults that show up on a cold start, when still in open loop, but upon checking sensor voltages all is normal, and the faults go away when they are cleared.
 
Originally Posted by George7941
The O2 sensors will only switch when the ECM adjusts fuel delivery based on feedback from the O2 sensors, pretty much the definition of closed loop.

Yes. Apparently the definition of closed loop has slipped. Perhaps it means fully closed, where long term trims are being adjusted, etc.
 
My truck engine is nowhere near operating temperature at three minutes after cold startup. It is at about 40C / 100F when the ECM changes the Loop Status to Closed.
 
My BMW will run the Secondary air pump for up to about 90 seconds after startup, maybe that needs to complete before the ECU goes to "closed Loop".

Does the engine adjust the Air Fuel ratio on a cold start. I know some vehicles will do that and will not get to closed loop until the richer mixture decays out over time.

Another vehicle I have has a temperature limit before the ECU actually uses the O2 sensor output. The O2 sensors on that vehicle are not heated.
( the closed loop limit is set to a head temp of 149 F, the engine normally operates at around 350 F head temp. It is an air cooled engine
smile.gif
)
 
The only way the O2 sensor will switch is if the ECM adjusts fuel delivery based on feedback from the O2 sensors and the A/F ratio is maintained close to stoichometric, and it is doing that about fifteen seconds after startup. The O2 sensors are heated, which is why they start working so quickly. No air pump on this engine.
 
Last edited:
Engine management (albeit very primitive) in my Barchetta doesn't consider the O2 sensor information till coolant temp reaches (I think) 40°C. Also the variable intake system is deactivated till coolant reaches a certain temp.

I guess the O2 sensors are "ready" way before that, perhaps the same thing is happening in your car, as others have said.
 
I believe it starts in open to heat everything up using a set fuel map. Then I think in most ECU's the ECU goes from open to closed loop based on several parameters like catalyst temp, RPM fluctuations (settling down) and throttle position. Plus I don't think it'll go into closed loop until the cat is fully heated. Until then it stays in open loop. How long it takes for everything to get heated up to where the ECU goes into closed loop can be effected by outside air temp, altitude, humidty. Your O2 sensors can be bouncing around but the ECU is ignoring them until the catalyst heats up, at which point it goes into closed loop using feedback from the O2 sensors and not set fuel map.

Again, that's just my understanding and I could be off/wrong so it's worth double checking.
 
Last edited:
The graph of the O2 sensor shows that it is switching exactly the way it does when the engine is at operating temperature. It is not bouncing around, it is reacting to changes in the mixture.

Can the ECM control fuel delivery enough to accomplish switching without feedback from the O2 sensors? It uses fuel maps and the learned STFT and LTFT values to produce a lean mixture one second and richen it the next second without feedback? Maybe or maybe not and the way to find out is to artificially enrichen the mixture with external propane or lean the mixture by disconnecting a PCV hose. That will be my next experiment. I will post the results if I can get it done soon enough before this thread becomes stale.
 
Originally Posted by George7941
The graph of the O2 sensor shows that it is switching exactly the way it does when the engine is at operating temperature. It is not bouncing around, it is reacting to changes in the mixture.

Can the ECM control fuel delivery enough to accomplish switching without feedback from the O2 sensors? It uses fuel maps and the learned STFT and LTFT values to produce a lean mixture one second and richen it the next second without feedback? Maybe or maybe not and the way to find out is to artificially enrichen the mixture with external propane or lean the mixture by disconnecting a PCV hose. That will be my next experiment. I will post the results if I can get it done soon enough before this thread becomes stale.


Yes, unplug your O2 sensors and watch it do just that. You wouldn't want your engine to just shut down in the event one of your O2 sensors fails. That's because the ECU will go into open loop and pick up where it left off and begin to manage the fuel trims using the fuel map and stored adaption table (self learning). The ECU's goal is to operate in closed loop but the ECU will go into open loop during certain situations like heavy load.(like towing or acceleration) because in open loop it doesn't have to "wait" around for the O2 voltage (lambda) readings. This is the compromise between emissions standards and mfgs performance requirements. The ECU generally operates in closed loop when things are steady like cruising on the hwy (steady rpms) and/or the throttle position is less than a certain %.

Fwiw, I believe emissions standards require a ECU to get into closed loop upon start up, within a certain time. I don't know what that time limit is...
 
Last edited:
To give a specific answer to YOUR controller (PCM's) behavior which is a Delphi P59.....

It has what we call Open Loop STFT control. As soon as the O2 are driven rich (Producing their own voltage) which indicates they are hot enough. The PCM will use the last updated STFT base fueling tables/maps/cells to begin Proportional Fueling.
Proportional Fueling is the "Switch/Oscillation" function that drives the O2's Rich-Lean without Over/Under shooting it too much.

Your able to see this more so on engines that don't warm-up very fast, The 4.3L in particular as it's base architecture is so old, But the PCM running the show is relatively modern. The first year for the Delphi P59 Controller was 2003. The only reason it wasn't used past 2007 was because it didn't quite have memory for AFM control.

Minimum Closed Loop enable Coolant Temps range from 100°F to 131°F on GMT800 platform's (P01 & P59 Controllers)
 
I haven't started my Camaro in a while, So I decided to data log a cold start......I'm using a 2005 GMT800 Delphi P59 to control a "Cammed" 2008 L92 (6.2L) that's converted to 24X/1X Crank/Cam reluctors

Drops into closed loop at the 2:48 mark, The time it took ECT to reach 132°F.
79°F coolant start-up temp, Ambient in my garage was about 85°F which to IAT reflected.

*if you watch the O2-mV....They heat-up & drive full Rich within 8 SECONDS!!!!. Granted I had the key ON for @30 seconds loading the data logger. My O2 heaters are powered by ignition HOT & Case ground. Not by the PCM. A stock P59 equipped vehicle will control the O2 heaters via Duty Cycle/PWM.

*At 28 seconds....Open Loop STFT begins by driving the fueling lean. Then O2 voltage oscillation is logged....Followed by STFT activity .

*While not technically in closed loop.....It IS using O2 feedback to control fueling!!!

Turn your speaker volume down!, I was listening to music on my laptop while filming the datalog. While my speaker were at a low volume.....My phone was at max volume
 
Very interesting! Thanks, clinebarger. 132F temp is reached at about 19.17.05 and the graphs at the bottom right show absolutely no change in behaviour at this point.
 
Now that we are discussing turbine rpm and input shaft rpm,isn't the turbine half of the torque converter locked (splined) on to the input shaft, just like the pump half is locked on to the engine flywheel? So why the two pids, input shaft rpm and turbine rpm, when they will always be the same?
 
Originally Posted by FlyNavyP3
Clinebarger,

Can you explain why the Trans turbine speed is faster than engine RPM?


Good question!!!!
grin.gif


The datalog was taken in Park, So the Turbine & Turbine Shaft (Along with the O/D Carrier, Overrun Housing, & Forward Clutch Drum in a 4L80E) are free to freewheel with the only drag being the mass of the parts along with the Forward & 4th Clutch Frictions overruning against their steels.

With no load on the Turbine inside the converter & with a little help from relatively cold ATF......The Turbine actually spins faster than the Impeller. Put another way.....The converter efficiency is over 100%. But it's not doing any work.

If HP-Tuners had a valid TCC Slip RPM PID......It would show Negative RPM under these conditions & the math works out perfectly everytime.....
Turbine RPM minus Engine RPM equals TCC Slip RPM.

Put the trans in gear (Brakes applied/Car stationary).....& the TCC Slip RPM PID will mirror/match Engine RPM as the Turbine Shaft is now being held stationary by the Forward Clutch in the Drive ranges & The Forward Drum is held stationary by the Direct Clutch & Lo/Reverse Band in the Reverse range.


I cycled in & out of drive in the 1st video, It's only good to the 1 minute mark as I turned my phone.
2nd video shows shome revs, Tracks linear through the RPM range.
3rd video is the TCC Slip RPM PID in Park & in Gear.....My wife startles me at the very end
lol.gif
The car was reverberating through the house.
 
Originally Posted by George7941
Now that we are discussing turbine rpm and input shaft rpm,isn't the turbine half of the torque converter locked (splined) on to the input shaft, just like the pump half is locked on to the engine flywheel? So why the two pids, input shaft rpm and turbine rpm, when they will always be the same?


Another good question!

On this particular unit (4L80E).....The ISS Sensor reads directly off the Forward Drum RPM. The Turbine Shaft RPM is actually Calculated as there is no physical speed sensor that reads Turbine Shaft speed.

The Forward Drum RPM & Turbine Shaft/Input Shaft RPM are the same under the following conditions.....
*The O/D Roller Clutch is Holding.....Decelerating/coasting down a hill in D4 will cause the roller clutch to Overrun. (engine return to idle speed). D1/D2/D3 ranges have the Overrun Clutch come-on & prevent the O/D Roller Clutch from Overruning thus providing engine braking.
*The trans is in ANY gear besides 4th/Overdrive.

In 4th gear/overdrive......The Forward Drum is Overdriven by 25%, So the ISS RPM PID will read 25% faster than the Calculated Turbine Shaft RPM PID.
 
Status
Not open for further replies.
Back
Top