New O2 Sensor Changed Idle Speed

[Latitude48]

Background:
I used the graphing capability of my Ftech9 DME/OBD+ to verify that my NBO2 sensor wasn't causing the DME to cycle between RICH and LEAN to maintain stoichiometric combustion in closed-loop mode.

So, while the turbo was in the shop for clutch master, slave, etc., a new NBO2 sensor was also installed. Now the DME cycles the fuel correctly, but my idle speed has mysteriously jumped up to around 1350 rpm.

Can anyone explain why this happened, and how to fix it?

Thanks.

[Tom]

Is this on the turbo? The idle speed is mostly a function of how much air is getting into the motor. The idle stabilizer valve (ISV) is a closed loop system that constantly adjusts the amount of air it is letting in (bypassing the throttle), and the DME targets 840. When the RPM's start creeping above 840, the ISV closes down a little to let less air in; when the RPMs drop below 840, the ISV opens up a bit to let a little more air in; and it does this constantly as a close-loop feedback system. When you set the base idle on these cars, you are setting the idle in a way that gives the ISV room to adjust the idle up or down by a few hundred RPMs as may be needed for whatever reason in order to keep it around 840. I'd start by re-setting your base idle (see Clark's) and seeing it that fixes the issue. There's a chance your idle is just too high after these repairs for the ISV to adjust it down enough. If setting the base idle doesn't fix it, then it's most likely a vacuum leak somewhere and/or the ISV is not working right for some reason (they get dirty and stick open/shut, and the connectors fray inside, and the driver inside DME's have gone bad occasionally). You can pull the connector off the ISV and plug/black one of the hoses to see if that's where the air and RPMs is coming from too.


Re vacuum leaks, yours wouldn't be the first shop to knock a vacuum fitting loose.


http://www.clarks-garage.com/shop-manual/fuel-08.htm

[Latitude48]

Yes, Tom, this is on the turbo (as stated ). And the DME is brand new, so that's not an issue. I agree that it's entirely possible that while the clutch master, slave and connecting hose/tubing were replaced a vacuum line could have been disturbed. I'll see what I can find. I'll also run the engine again and graph the ISV % Open. If it's closing and still can't control the idle, a vacuum leak is the most likely culprit.

Re: vacuum leaks. When the S2 had a recent major service the intake manifold was pulled and all vacuum lines were replaced, as well as a new hose for the air-oil separator, which unknown to me had a major hole that someone years ago "fixed" with RTV. Over the years this repaired hose leaked again as the idle speed on the S2 has always been problematic and erratic. Now that all these little leaks are fixed the idle is rock solid.

[michael2e]

Timely discussion. Sorting out an idling issue on my S2. Any of you guys know the difference between the ISV on these cars vs the turbos? I recently completed a full run through to solve misfiring on my 87 951 and as I'm digging into the S2 it appears that the ISV may function differently on these cars. When closed the 951 ISV (which has 3 electrical terminals) does not allow any air to pass through it whereas my S2 ISV (with 2 terminals) doesnt seem to seal closed to that degree. Is this right? Thanks for any insight

[cda951]

I 100% agree to start by checking for vacuum leaks.

You should also check that the throttle is closing all the way, and that the throttle position sensor is adjusted properly so the microswitch is engaged to tell the DME to operate the idle control valve. It could be that your base idle is set too high (1350 RPM) and a misadjusted or malfunctioning TPS or wiring is causing the high idle.

I always set the base idle a bit high (maybe 1000-1050 RPM) to compensate for wear in the idle control valve, which are often quite old in these cars. If everything is functioning properly, the ICV will bring the idle down to 850-900.

[Latitude48]

I 100% agree to start by checking for vacuum leaks.

You should also check that the throttle is closing all the way, and that the throttle position sensor is adjusted properly so the microswitch is engaged to tell the DME to operate the idle control valve. It could be that your base idle is set too high (1350 RPM) and a misadjusted or malfunctioning TPS or wiring is causing the high idle.

I always set the base idle a bit high (maybe 1000-1050 RPM) to compensate for wear in the idle control valve, which are often quite old in these cars. If everything is functioning properly, the ICV will bring the idle down to 850-900.

One of the parameters displayed by the Ftech9 DME OBD+ is the Idle Control Valve (ICV) % Open. Another parameter is the Throttle Switch Mode. With the engine warmed up and idling in the 1200-1300 rpm range per the tach... roughly 1160 ish displayed by the OBD+, the ICV % open is 87.7. The Throttle Switch Mode is "Full Load," meaning the throttle is closed. I've asked Joe about this as this nomenclature is confusing to me because I would think that the throttle switch in the idle position would be a mode like "Closed" or "No Load." Increasing the throttle above idle changes the Mode to "Above Idle." So, it does seem like the throttle position sensor is working correctly.

And the ICV does move more towards closed when the rpm is increased, so I know it's "trying to do something," I just don't know why it doesn't try to close and regulate the idle to a lower setting when the throttle is closed. But I guess if the base idle setting is off the ICV may think this is the correct idle position. The base idle setting hasn't been adjusted in a long time, so it may have been set when the O2 sensor was bad.

BTW, before the O2 sensor was replaced when I first started the engine it ran a bit rough and wouldn't idle smoothly when cold. It runs much better now when it's cold... so, something has changed for the better.

[usury]

Does the throttle body on the turbo have an idle adjustment screw?

[cda951]

The Throttle Switch Mode is "Full Load," meaning the throttle is closed. I've asked Joe about this as this nomenclature is confusing to me because I would think that the throttle switch in the idle position would be a mode like "Closed" or "No Load." Increasing the throttle above idle changes the Mode to "Above Idle." So, it does seem like the throttle position sensor is working correctly.

That is weird. The 951 DME and KLR setup is unique in that the two control units "see" different inputs from the TPS. The DME control unit receives the input from the idle contact and thus controls the ICV. The TPS potentiometer and the 66%/full load signal are conveyed directly to the KLR and then from the KLR to the DME via pin 3 (which signals the DME to ignore the oxygen sensor input and enable full load enrichment to protect the engine), so I don't think it is normal for the reading to be "full load" at idle.

Of course Joe @ Ftech9 knows a lot more about these control units than I do, but I did install one of these in a customer's '89 951 a few years ago and I was using the diagnostic software to pinpoint a drivability issue. I am almost certain that the TPS input readings corresponded with reality, as in "idle" or "above idle" and "full load." I would have immediately flagged a weird discrepancy as you describe above.

[Tom]

The Throttle Switch Mode is "Full Load," meaning the throttle is closed. I've asked Joe about this as this nomenclature is confusing to me because I would think that the throttle switch in the idle position would be a mode like "Closed" or "No Load." Increasing the throttle above idle changes the Mode to "Above Idle." So, it does seem like the throttle position sensor is working correctly.

That is weird. The 951 DME and KLR setup is unique in that the two control units "see" different inputs from the TPS. The DME control unit receives the input from the idle contact and thus controls the ICV. The TPS potentiometer and the 66%/full load signal are conveyed directly to the KLR and then from the KLR to the DME via pin 3 (which signals the DME to ignore the oxygen sensor input and enable full load enrichment to protect the engine), so I don't think it is normal for the reading to be "full load" at idle.

Of course Joe @ Ftech9 knows a lot more about these control units than I do, but I did install one of these in a customer's '89 951 a few years ago and I was using the diagnostic software to pinpoint a drivability issue. I am almost certain that the TPS input readings corresponded with reality, as in "idle" or "above idle" and "full load." I would have immediately flagged a weird discrepancy as you describe above.

I read Latidue48's comment to mean the ftech9 logger uses "full load" as just a label to mean closed throttle. If so, he should really update that label as it is, at best, super confusing and, at worst, just flat out wrong. Making it worse is that the KLR has a separate output for "full load" which tells the DME to switch over to its wide-open throttle maps once the throttle opens more then ~65% open. When that happens, KLR pin 18 goes to ground (0 volts) and that signal is picked up on DME pin 3 to tell the DME to switch to its WOT maps. Does the ftech9 allow you to monitor the status of DME pin 3 -- i.e., the actual 'full load' signal generated by the KLR? As long as that is 5 volts at idle and part throttle, then the system is not truly in full load mode. You can also test the voltage at terminals B and C of the diagnostic port under the hood. You should see 5 volts at idle and part throttle.

When Crazy Eddie's KLR had a cracked solder joint, it would sent the car into WOT/full throttle mode at idle randomly. The engine speed actually decreased and ran poorly at idle, because it was running so rich off the WOT maps. Not sure what to make of the ISV open 87% thing though. One could imagine the DME opening the ISV like that if it thought your were flooring the throttle. I'd make sure DME pin 3 is getting 5 volts at idle. Assuming it is, I'd go back to vacuum leaks and capping off the ISV, etc.

[Latitude48]

Tom, OBD+ is working correctly and displaying what it "thinks" the throttle switch position is. The correct mode for idle is "Idle." I just confirmed this by hooking up my laptop to the DME and without the engine running and the ignition ON, "Idle" was displayed for the mode. Moving the throttle to any other position changed the mode to "Full Load." However (!), turning the ignition OFF and ON again the idle position mode was now "Full Load" and the off-idle position was "Above Idle." I tried numerous times to get it back to "Idle" and "Full Load" but could not repeat the initial test.

Something weird is happening.

All the OBD+ readings and behavior I'm seeing now explain why the NBO2 Digital and Voltage signals behave the way they do. At idle NBO2 is always RICH and the NBO2 voltage is high because the DME thinks the throttle is at Full Load, so it ignores the NBO2 signal and richens the mixture. At any throttle position off-idle (just barely crack the throttle) the NBO2 Digital signal switches between RICH and LEAN like it should, with corresponding changes in NBO2 Voltage as the DME attempts to control the mixture to stoichiometric. This never occurred at any rpm before I changed my NBO2 sensor, so this does confirm the old sensor was faulty.

These results seem to point towards a faulty Throttle Switch (TPS), or an issue with the KLR. According to my records the TPS was changed in 2002, or 58K miles ago. OBD+ does not have an output for DME Pin 3, it just displays the input voltage to the DME, which in my case was 13.1x.

Tom