New O2 Sensor Changed Idle Speed

Talk and Tech about turbocharged 924/944/968 cars
User avatar
Latitude48
Posts: 194
Joined: Fri Jul 16, 2021 7:08 pm
Location: Seattle Area
Has thanked: 56 times
Been thanked: 51 times
Contact:
Well Tom, when I run at idle with "Full Load" displayed for the mode, my NBO2 sensor drops out of closed loop (as I've stated previously) and the signal is always Rich. If I barely move the throttle to change the mode to Off Idle it goes closed loop and the mixture oscillates between Rich and Lean, with corresponding changes in NBO2 voltage.

The test that Joe wants me to run, with the KLR out of the loop and the throttle switch jumped to fix the mode to Idle, should show the NBO2 in closed loop with the throttle closed.


Tom Pultz
- 1989 944 Turbo - Guards Red/Linen
- 1990 944 S2 - Guards Red/Black
- 2003 Audi 1.8TQ - Denim Blue/Black
- 2003 Honda Civic Si - Vivid Blue/Black
- 2023 VW Golf R Base - Lapiz Blue/Titan Black

#31

User avatar
Tom
Site Admin
Posts: 5126
Joined: Fri Jun 25, 2021 2:04 pm
Location: Silicon Valley, CA
Has thanked: 477 times
Been thanked: 2240 times
Contact:
Latitude48 wrote: Sat Oct 29, 2022 10:10 am Well Tom, when I run at idle with "Full Load" displayed for the mode, my NBO2 sensor drops out of closed loop (as I've stated previously) and the signal is always Rich. If I barely move the throttle to change the mode to Off Idle it goes closed loop and the mixture oscillates between Rich and Lean, with corresponding changes in NBO2 voltage.

The test that Joe wants me to run, with the KLR out of the loop and the throttle switch jumped to fix the mode to Idle, should show the NBO2 in closed loop with the throttle closed.
Yeah, that would at least suggest the OBD+ and DME are both in Full Load mode. If I have time, I'll play around with my car and see if a bad idle contact might trick the DME into full load mode -- maybe with a potentiometer bridging the TPS idle contacts. I've never seen or heard of that, but I doubt most people think to check the full load signal when the TPS is acting up -- they just get a bad idle, change the TPS, and motor on.


#32

michael2e
Posts: 14
Joined: Mon Jan 31, 2022 9:53 am
Has thanked: 1 time
Been thanked: 3 times
“… like a moth to a flame! “. Wow guys super interesting thread. Following along. Makes me want to run out and get one of those F9t DME‘s with the OBD capabilities considering all the information you’re able to analyze.


1987 Porsche 951 "Blacky"
1989 Porsche 944 S2
2003 Harley Davidson Softail Deuce

#33

User avatar
Latitude48
Posts: 194
Joined: Fri Jul 16, 2021 7:08 pm
Location: Seattle Area
Has thanked: 56 times
Been thanked: 51 times
Contact:
Quick update. I ran a couple of tests:
1) KLR disconnected, DME in NA mode, throttle switch disconnected, and terminals 4 & 6 jumped to fix in Idle mode.
Result: engine started and idled normally; throttle switch mode displayed was Idle and the voltage was 2.95, both as expected. Idle speed was in the 8xx range. I'm not sure if this test proved or disproved anything, but I did run the engine until it was fully warmed up and the DME was in closed loop mode as seen by the cycling NBO2 sensor. I sent the log file to Joe for his analysis.

2) KLR reconnected, DME in Turbo mode, throttle switch reconnected, i.e., everything "normal."
Result: surprisingly(!), the engine started and idled normally; throttle switch mode displayed was Idle and the voltage was 2.95. Idle speed was in the 8xx range. Previously, the engine would have been operating in Full Load mode at idle.

As I saw in a previous test, the ICV % Open was 115.5 during both tests, which is not normal and needs some further analysis by Joe.

Apparently, disconnecting the throttle switch and reconnecting it must have resulted in a better electrical connection, which "fixed" the issue I was having with a Full Load condition at idle. Or maybe moving the switch around changed something in the harness. I just don't know at this point. Hopefully, the various resistance measurements I'll be performing this coming week might provide a clue.


Tom Pultz
- 1989 944 Turbo - Guards Red/Linen
- 1990 944 S2 - Guards Red/Black
- 2003 Audi 1.8TQ - Denim Blue/Black
- 2003 Honda Civic Si - Vivid Blue/Black
- 2023 VW Golf R Base - Lapiz Blue/Titan Black

#34

User avatar
Tom
Site Admin
Posts: 5126
Joined: Fri Jun 25, 2021 2:04 pm
Location: Silicon Valley, CA
Has thanked: 477 times
Been thanked: 2240 times
Contact:
Re-seating connectors can improve the connection by scraping away corrosion, etc. But you'd think with all the tests you've run, if that were the root problem, you'd have 'fixed' it by now. You can try rattling the harness with the motor running to see if anything changes. Intermittent electrical issues are the worst. It took a long time for Ed and I to find the cracked solder joints in his KLR. We'd test and all would be well, but every now and then the motor would idle terribly. We finally realized the full load signal was triggering at idle, and after testing everything else, we opened up the KLR and the cracked solder joints were pretty clear under a magnifying glass. 5 minutes with a soldering iron was all it needed, but getting to that point was pretty frustrating for Ed.


#35

User avatar
Latitude48
Posts: 194
Joined: Fri Jul 16, 2021 7:08 pm
Location: Seattle Area
Has thanked: 56 times
Been thanked: 51 times
Contact:
Joe provided feedback on the odd ICV open percentages (> 100%) I've been seeing with this Alpha build, which is a developmental build. He said that it's possible some of the changes he's been making changed something in the ICV calculations. He suggested I uninstall that build and install the latest Beta build, which I did today.

I ran the engine again and everything looked normal, including the ICV values. After the engine coolant temp was at 190, I recorded 20 sec of data. The average idle speed was 818, which is in the spec range of 840 +/- 40. The ICV % Open averaged 85.9.

So, a lot of experimenting and testing to diagnose a weird idle condition that was fixed (knock on wood) by simply disconnecting and reconnecting the throttle switch. FWIW, the 951 has been parked in the garage for about 3 years, and before that my son had it stored in a rental garage for a year, so I guess enough corrosion built-up on the contacts to change the idle position resistance enough to trigger Full Load mode.


Tom Pultz
- 1989 944 Turbo - Guards Red/Linen
- 1990 944 S2 - Guards Red/Black
- 2003 Audi 1.8TQ - Denim Blue/Black
- 2003 Honda Civic Si - Vivid Blue/Black
- 2023 VW Golf R Base - Lapiz Blue/Titan Black

#36

User avatar
Tom
Site Admin
Posts: 5126
Joined: Fri Jun 25, 2021 2:04 pm
Location: Silicon Valley, CA
Has thanked: 477 times
Been thanked: 2240 times
Contact:
Latitude48 wrote: Mon Oct 31, 2022 4:53 pm Joe provided feedback on the odd ICV open percentages (> 100%) I've been seeing with this Alpha build, which is a developmental build. He said that it's possible some of the changes he's been making changed something in the ICV calculations. He suggested I uninstall that build and install the latest Beta build, which I did today.

I ran the engine again and everything looked normal, including the ICV values. After the engine coolant temp was at 190, I recorded 20 sec of data. The average idle speed was 818, which is in the spec range of 840 +/- 40. The ICV % Open averaged 85.9.

So, a lot of experimenting and testing to diagnose a weird idle condition that was fixed (knock on wood) by simply disconnecting and reconnecting the throttle switch. FWIW, the 951 has been parked in the garage for about 3 years, and before that my son had it stored in a rental garage for a year, so I guess enough corrosion built-up on the contacts to change the idle position resistance enough to trigger Full Load mode.

Glad to hear it seems fixed, and that Joe is making progress toward a production version. I haven't had a chance to check joe's TPS resistance theory yet on my car, but it's on my list and will let you know when I do that.


#37

User avatar
Latitude48
Posts: 194
Joined: Fri Jul 16, 2021 7:08 pm
Location: Seattle Area
Has thanked: 56 times
Been thanked: 51 times
Contact:
There is an earlier build that's the current production build. The build I installed today is a Beta version that includes the ability to readout and record the ignition advance.


Tom Pultz
- 1989 944 Turbo - Guards Red/Linen
- 1990 944 S2 - Guards Red/Black
- 2003 Audi 1.8TQ - Denim Blue/Black
- 2003 Honda Civic Si - Vivid Blue/Black
- 2023 VW Golf R Base - Lapiz Blue/Titan Black

#38

User avatar
Tom
Site Admin
Posts: 5126
Joined: Fri Jun 25, 2021 2:04 pm
Location: Silicon Valley, CA
Has thanked: 477 times
Been thanked: 2240 times
Contact:
Latitude48 wrote: Mon Oct 31, 2022 5:24 pm There is an earlier build that's the current production build. The build I installed today is a Beta version that includes the ability to readout and record the ignition advance.
That alone is worth the price of admission, especially if running aftermarket chips with unknown timing curves. I was so determined to get that information at one point, I hacked into the DME board to calculate the actual timing advance off the speed/ref and ignition pulses...



#39

User avatar
Latitude48
Posts: 194
Joined: Fri Jul 16, 2021 7:08 pm
Location: Seattle Area
Has thanked: 56 times
Been thanked: 51 times
Contact:
Implementing the ignition advance feature wasn't easy as TDC had to be determined, which was a bit of a chore to do in code since the "Mark" sensor outputs a signal way in advance of TDC. This is still Beta code as it was to be verified by comparing to the ignition advance of a client's car with a standalone system on the dyno, and I don't know if that happened yet.

FWIW, the average advance of my 951 at idle in the recent run I recorded was 5.38, which IIRC is within the spec range of 5 +/- 3.


Tom Pultz
- 1989 944 Turbo - Guards Red/Linen
- 1990 944 S2 - Guards Red/Black
- 2003 Audi 1.8TQ - Denim Blue/Black
- 2003 Honda Civic Si - Vivid Blue/Black
- 2023 VW Golf R Base - Lapiz Blue/Titan Black

#40

Post Reply