New O2 Sensor Changed Idle Speed

[danmartinic]

What happens to the idle when you disconnect the O2 sensor (the plug back middle of engine)?

[Tom]

Something indeed sounds weird. I would probably rely on traditional multimeter testing unless/until you can firm what the ftech9 is reporting is what's actually happening. It's super easy to check the full load signal on the diagnostic port -- I'd definitely do that. Also, the KLR will not trigger its 'full load' signal -- i.e., it will not switch its pin 18 to ground -- unless the motor is running (according to the KLR test plan). While anything is possible if the KLR is faulty, I'd be pretty suspicious of the 'full load' indicator on the ftech9 read-out unless/until you confirm it manually with a multimeter (at least suspicious of it as an indicator of the DME/KLR switching into WOT maps.)

I'd say to test the full load signal at the diagnostic port. If that seems ok, I'd reset the base idle and see if the idle goes down to normal. cda's point about the throttle plate screw and TPS adjustment are good points, so I'd check both as part of setting the base idle. If you still can't get it down to 840ish, I'd pull the ISV hose off the intake manifold and block off the intake and the hose, to see if that's where the extra air was coming from. If it's still idling way up there with the ISV blocked off and the idle screw cranked down, it almost has to be a vacuum leak. It can't rev without air, and if the air isn't coming from the ISV or TB, it's coming from a leak....

[Latitude48]

Not sure why, but I've written a couple of updates that never got posted. I must be getting logged out again.

Anyway, Joe from Ftech9 responded to my comments in the OBD+ thread of RL. He said it is likely the idle contacts are not electrically making a connection when throttle is closed. I concur with that comment as it would explain why the ICV doesn't act to control the idle. However, it doesn't explain why with the engine running and the throttle in the idle position the mode is "Full Load." For the DME to think the throttle position is at full load the throttle position must be "Above Idle" (signal from the DME) AND in the WOT position (signal from the KLR when greater than 65 deg open). The first condition is true if the idle contacts are not working, but I don't know how the second condition can be true in the idle position. Maybe the KLR is sending a faulty signal?

I need to find my multimeter and do some testing. I'm not sure where I stored it as my garage is in a bit of a disaray as I've got thousands of my kid's (ha, ha, he's 47) vinyl records stored, so I've moved a lot of things around.

In any event, I ordered a new Bosch 0 280 120 400 sensor from an eBay vendor ($255).

It's interesting that per Clark's Garage a couple of the symptoms of a bad throttle switch are: 1) Poor Idle - particularly a high idle, and 2) Intermittent Boost. I have both of those. The latter condition has been an issue for years... some days you get boost, some you don't.

[Latitude48]

I ran some simple tests today:

Test 1 - Cold Engine w/Engine OFF, Ignition ON
Idle Position Throttle Switch (TS) Mode - Idle; barely move throttle, Mode changes to Full Load (should be Off Idle); back off throttle, Mode changes back to Idle. Repeat, repeat, repeat... same result.

Test 2 - Engine OFF, Switch Ignition OFF, then ON again
Idle Position TS Mode - Full Load (should be Idle); barely move throttle, Mode changes to Off Idle, back off throttle, Mode changes back to Full Load. Repeat, repeat, repeat... same result.

Test 3 - If I repeat Test 2 (switching the ignition OFF and ON) a few times I can get the same results as Test 1, but it seems the longer I keep the ignition ON while testing, eventually I can't get the mode to indicate Idle with the throttle closed, indicating something in the DME or KLR that changes after power is applied for a period of time.

Test 4 - Engine RUNNING
Idle Position TS Mode - Full Load (should be Idle) barely move throttle, Mode changes to Off Idle, back off throttle, Mode changes back to Full Load. Repeat, repeat, repeat... same result. These are the operating modes I always get with the engine running. I can never get Idle mode with the throttle closed.

One item to note: the ICV % Open started out around 70% and very gradually increased to around 85% open during the period the engine was running.

Test 5 - Engine OFF, KLR Disconnected, Ignition ON
Same results as Tests 2 or 4.

I really don't know how to interpret these test results. On one hand the KLR might be at fault since it appears to be sending a Full Load signal with the throttle in the idle position. But on the other hand, if the KLR is disconnected the OBD+ software (SW) still thinks the mode is Full Load in the idle position, which doesn't physically seem possible unless the SW latches the previous value when the engine was running.

I ordered a new DVM and will run through the various troubleshooting tests in the Turbo WSM when I receive it. To me it currently seems like the "TPS" may be functioning correctly since there are tests when the mode is Idle with the throttle closed, and other times the mode is Off Idle with the throttle advanced, both of which would indicate normal operation. However, I can't get them to display in that order, which points to OBD+ software and/or KLR (most likely problem?) issues.

If anyone else has performed any of these tests, please post your results. Thanks.

[Tom]

I think it is wise that you ordered a multimeter. Your tests offer pretty convincing evidence that the ftech9 software is reporting ‘full load’ when, in fact, the KLR is not producing an actual ‘full load’ signal.

The TPS has two things inside. An on/off switch, that tells the DME if the throttle is fully closed or not fully closed. For a high idle, that is the signal that is most relevant. The DME needs to know when the throttle is fully closed so that it can command the ISV to open/close as needed to get to 840 rpms. If the idle switch is not working and adjusted correctly, it can cause a high idle. Does your system report on the idle switch?

The second thing inside the TPS is a potentiometer (variable resistor) used exclusively by the KLR to determine when to trigger the full load signal. If the full load signal is falsely triggering, it can create idle problems, but typically not a fast idle. And as described below, I don’t believe that the ‘full load’ signal reported by your system corresponds to an actual ‘full load’ signal from the KLR.

Re Test 1: See the video below, but the bottom line is that the KLR will not produce a full load signal unless the motor is running. The KLR test plan says the motor has to be running to test the signal, but in an effort to post first-hand information, I ran the test myself with and without the motor running (ignition on for both), and can confirm the KLR simply does not produce the full load signal unless the motor is running.

Re Test 2/3: I discovered doing my own tests that the full load signal from the KLR is not a meaningful indicator of anything when the engine is not running. If you turn the key to off and then back to ignition on, the full load signal produces random outputs, typically about 1 volt. This would seem to be just an artifact of the electrical design and not an intentional output. After a key-out hard reset of the KLR, you can get 5 volts again – which probably explains the different results between your test #1 and #2 and your observations in test #3. Unless there is a separate idle switch contact status indicator, my suspicion is that ftech9 likely uses both the TPS idle contact switch and the TPS potentiometer to create a software-based throttle status indicator. Either way, it can’t work right without the motor running because the KLR itself is producing random/meaningless voltages on the full load signal line when the motor isn’t running.

Re Test #4: Hard to know what’s going on there, but reason enough to break out the multimeter to check the TPS idle contact and the full load signal.

Re Test #5: If you are seeing ‘full load’ with the KLR disconnected, that’s just further evidence that the ‘full load’ reported by the ftech9 software is not an accurate representation of the KLR’s full load signal, for whatever reason.

My general sense is that the ftech9 readings are complicating things for now. Tests 1-3 are probably red herrings since we know the KLR is not designed to work right with the motor off. So, even if the ftech9 software is working perfectly, it can’t be expected to work when the KLR isn’t. Test #4 is the only one that ‘should’ work, and its results are illogical.
Sorry to go on and on, but the bottom line is I’d test the TPS idle contact with a multimeter (all the way to the DME) and go from there.

[Latitude48]

The TPS has two things inside. An on/off switch, that tells the DME if the throttle is fully closed or not fully closed. For a high idle, that is the signal that is most relevant. The DME needs to know when the throttle is fully closed so that it can command the ISV to open/close as needed to get to 840 rpms. If the idle switch is not working and adjusted correctly, it can cause a high idle. Does your system report on the idle switch?

According to Joe, for the DME to enter Full Load it must sense BOTH an Off Idle condition and WOT. The DME may be working correctly but the status values reported by the OBD+ may not be true. The OBD+ doesn't report the status of the microswitch by itself, just what it thinks is the throttle position: Idle, Off Idle, or Full Load. If the SW thinks the throttle status is Full Load at idle but the engine idles fine, just too high rpm, it seems that the voltage the OBD+ is sensing to determine Full Load may need to be adjusted to avoid what looks like false positives due to spurious voltage output by the KLR.

The second thing inside the TPS is a potentiometer (variable resistor) used exclusively by the KLR to determine when to trigger the full load signal. If the full load signal is falsely triggering, it can create idle problems, but typically not a fast idle. And as described below, I don’t believe that the ‘full load’ signal reported by your system corresponds to an actual ‘full load’ signal from the KLR.

I agree that I do not believe the DME is getting a true Full Load signal at idle since the idle itself is fine, it's just too high.

Unless there is a separate idle switch contact status indicator, my suspicion is that ftech9 likely uses both the TPS idle contact switch and the TPS potentiometer to create a software-based throttle status indicator.

Correct, as stated above.

Tests 1-3 are probably red herrings since we know the KLR is not designed to work right with the motor off.

Agreed, except for the times I do see a throttle status of Idle, which "might" indicate the microswitch inside the throttle switch is working... at least some of the time. I do have a new throttle switch, but won't change it until I can run the tests in the WSM in the next week or so.

Thanks much for your input, Tom

[Tom]

According to Joe, for the DME to enter Full Load it must sense BOTH an Off Idle condition and WOT. The DME may be working correctly but the status values reported by the OBD+ may not be true. The OBD+ doesn't report the status of the microswitch by itself, just what it thinks is the throttle position: Idle, Off Idle, or Full Load. If the SW thinks the throttle status is Full Load at idle but the engine idles fine, just too high rpm, it seems that the voltage the OBD+ is sensing to determine Full Load may need to be adjusted to avoid what looks like false positives due to spurious voltage output by the KLR.

That all makes sense. As a focus group of 1, for diagnostic purposes, I would think it's more useful to have the system report the actual electronics-driven values of the TPC idle switch and potentiometer, and KLR full load signal, rather than trying to generate/replicate those values via software. My system does that and it's extremely useful -- it's like having a multimeter hooked up to all the relevant signals all the time. For the full load signal, it sounds like this OBD+ software is reporting when it thinks the KLR SHOULD be triggering its full load signal rather than a report as to when it actually IS triggering the signal. Maybe that's all dictated by the OBD+ format and can't be changed though (?), I don't know.

The other question I'd have is whether that OBD+ status is solely for user information, or if the DME is using it for any engine management purpose. If the DME 'thinks' its in full load mode when it really isn't, that's a much bigger issue than just wonky data logs.... (I'd assume it's solely user information....)

Just to clarify your last point, the KLR full load signal was a clear logic-level 5 volts at idle and part throttle, and a clear logic-level ground when it triggers full load. No spurious voltages when the motor is running. The only time I got spurious voltages is when I tried it without the motor running, to replicate your tests. But since that is not a valid operating or test condition, the OBD+ can't be expected to make any sense of that signal when the motor is off.

I agree that I do not believe the DME is getting a true Full Load signal at idle since the idle itself is fine, it's just too high.

Agreed. If anything, it's not really in idle mode if the idle contact in the TPS isn't clicking/working when the throttle is closed. (Have you confirmed it clicks when at idle, and that the throttle stop screw isn't holding the throttle open too much? All part of setting the base idle, which is free and easy while waiting for the multimeter....)


[/quote]

[Latitude48]

That all makes sense. As a focus group of 1, for diagnostic purposes, I would think it's more useful to have the system report the actual electronics-driven values of the TPC idle switch and potentiometer, and KLR full load signal, rather than trying to generate/replicate those values via software.

Joe posted a special test version of the software today that also displays the DME voltage. The DME does not actually read the two digital inputs for idle and full-throttle directly. These two inputs are processed by a resistor network and combined into a three-state voltage level. That voltage level is then fed to the DME’s ADC and is read out like any other analog sensor. Idle voltage should be ~ 3 V, Off Idle ~ 5 V and Full Load ~ 3.8 V.

Suggested tests with KLR & TPS connectors disconnected:
1) Ignition ON; this should yield Off Idle result since terminals 4 & 6 are an open circuit.
2) Short terminals 4 & 6 on the TPS connector; this should yield Idle result.

Results: test #1 should yield ~ 5 V; result V = 4.96. PASS.
I haven't performed test #2 yet as I need to find some fine gauge wire to short the connector terminals or get some small test leads.

The other question I'd have is whether that OBD+ status is solely for user information, or if the DME is using it for any engine management purpose.

The displayed parameters such as the Throttle Switch modes are for information.

(Have you confirmed it clicks when at idle, and that the throttle stop screw isn't holding the throttle open too much?

The switch clicks when closing the throttle but does not click when opening. I haven't adjusted or looked at the throttle plate adjustment screw as I doubt it has changed.

So, it looks like the DME is working correctly and the TPS may be faulty, but there are still issues with the throttle switch modes displayed that need to be understood.

[Tom]

I just looked at the DME schematic and see the resistor network you mention. Never looked at that before. I assume Joe's system is trying to replicate that output somehow, rather than monitoring that actual value? Otherwise, I'm not sure how they could ever differ the way yours seem to be showing. Joe should join Carpokes. We're nice here, the water is fine. He can run a free banner ad.

I think that second test (shorting the TPS connector to simulate the close idle contact) is a critical test to run. I just went out and tried mine, and it definitely clicks as you open the throttle and again when closing the throttle.... If yours does not click when opening, it might just need to be adjusted a wee bit, though it's a bit suspicious that it clicks in one direction but not the other....

If bridging the TPS connector proves inconclusive, I'd test it both at the TPS and at the 35-pin connector in the footwell. Here's is the carpokes guide for doing that....

viewtopic.php?f=19&t=525

[Latitude48]

Joe should join Carpokes. We're nice here, the water is fine. He can run a free banner ad.

If bridging the TPS connector proves inconclusive, I'd test it both at the TPS and at the 35-pin connector in the footwell. Here's is the carpokes guide for doing that....

viewtopic.php?f=19&t=525

Wow, nice test procedure

I've asked Joe to join Carpokes, guess I need to keep bugging him.