93 DOHC Normaly Aspirated, Stealth scan data question

[buickwagon]

My son's Stealth is having some issues that I believe are O2 sensor related. This car has a single heated 02 sensor on the inlet of the cat. No codes are stored, CEL is off, but there is a distinct gasoline smell when running, power is reduced and fuel economy is off. No sign of any fuel leak.

In order to do some testing, I hooked it up to a scan tool (OTC 4000E with 1993 import cartridge and Mitsubishi adapter) to monitor the live data stream. Scan tool shows a steady near-0 volts from the 02 sensor, so ECU should think mixture is lean. However, I cannot find a fuel trim value in the available data stream. Does the ECU not report fuel trim settings or is this a bug in my scan tool's software? Similarly, I don't see anything indicating if the system is operating in closed loop or open loop. The only fuel-related data stream is the injector pulse width, which of course varies considerably according to other factors.

Oddly enough, bench testing of the O2 sensor with a propane torch suggests the sensor is functioning and responds quickly to changes. Applying 1.1v to the vehicle side O2 signal wire is immediately reflected accurately on the scan tool and the engine idle changes. (This suggests to me that the car is in closed loop at that time.) However, on removing the voltage it does NOT drop immediately to 0. It slowly falls over time like a capacitor is being gradually discharged. I'm scratching my head trying to figure out if this indicates I am chasing an ECU problem rather than an O2 sensor problem. It would appear that the PO has changed this sensor at east twice: There is a mismatch between the O2 sensor wire colours and the vehicle harness colours. And it appears a universal O2 sensor was subsequently spliced into that.

The heater circuit is only drawing 1/2 an amp, which seems a little low to me and I noted that the sensor barely warms when power is applied to the heater on the bench. I think that would delay closed loop operation, but shouldn't the exhaust gasses eventually raise the O2 sensor up to operating temperature?

I would appreciate any insights. Thankyou.

[buickwagon]

Update: I used an arbitrary waveform generator to inject a 1hz square wave. The data stream accurately reflected the frequency, but instead of the 0.2v to 0.8 I programmed, it was swinging from 0.3 to 1.5v.

[speedy25]

You have gotten more sophisticated than most people on here to test! At least I know you understand electricity.

My experience is that if the O2 sensor reponds its OK. That may or may not be correct from the info you have given. I think you have other problems than the sensor. Have you confirmed that the right wires were connected from the universal part?

I hope you already know that the 1st gen cars have capacitor problems in the ECU that cause a bunch or woes. rymer1280 kniows these well and is well thought of for repairs. The two of you can probably geek speak electronics.

I would look to the injectors for being the problem. I bet they are leaking.

-SP

[Bashere]

The ecu does report all fuel trims. I think your particular tool just doesn't work correctly with these cars. Siunds like a universal type.

If there is a way you can poke a specific memory address manually you can get the stored value and do the conversion. Though would be easier just to get mmcd running on a palm.

As far as the voltage swings go, seems really odd to me. It should not be rising over 1 volt. I would be pulling the case off the ecu looking for bad components. If you haven't already, go to 3sx.com and get a manual from the FAQs section. Check the output pins to the o2 from the ecu to make sure they are sending the correct signals.

Sent from my SAMSUNG-SGH-I717 using AutoGuide App

[buickwagon]

The wiring did send us down the garden path for a bit. The wire colours differ from those shown in the downloaded manual, which shows two blacks and two whites in the 93 MFI diagram of Volume 2. On the harness side, we have a red, a white with a silver stripe, a black and a black with a silver stripe. Contrary to the first few posts we found on this site, the two blacks are not the heater and the white is apparently not the ground. The Red is definitely the heater supply, with +12v. The white/silver and black/silver are apparently the signal circuit, with white/silver apparently being positive. That leaves the black as heater ground. If all that is correct, then the sensor is wired correctly. We verified the heater circuit in the sensor and the polarity of the signal wires.

The scan tool is not a Mitsubishi product. OTC made the Tech and Tech II scan tools for GM and the 4000E was their generic tool covering all marques through vehicle-specific software cartridges and adapters. OTC licensed the technology from each manufacturer and supposedly their cartridges offer OEM tool functionality. I've never run into any problems with other cars as far as missing data streams before, but occasionally there is a bug with commanding certain functions. EG: the 2001 Buick ABS motor re-home command doesn't work right. So this could be a software shortcoming. Or it could be the nomenclature -- OTC uses 3 or 4 letter shorthand to identify the streams. I was looking for something like "STFT" (Short Term Fuel Trim) or "BLM" ("Block Learn Mode"). I'll try again and if it I can't find it then I'll hunt down MMCD.

Ok, googled it. I have some palm pilots. I wonder if it will work with my Palm TX (Garnet OS) or if I have to dig out my old IIIC. I found two cable schematics -- one being a simple transistor and resistor interface, the other incorporating a full RS232 interface chip. Is there a preference for these cars? Or even better, a proper cable with a mating connector instead of plugging discreet wires into the OBD connector (I'm getting to that age when my arms are too short... )

As for the injectors, I haven't done a fuel system leak-down test yet (still looking for a test port, starting to suspect there isn't one) but I don't think that's the problem here -- if the injectors were leaking and the O2 circuit was working properly, the system should be pegged rich. This is pegged lean.

[Bashere]

The passive cable works just fine for hooking up to the serial. Just have to become a contortionist to get it plugged in. My cable is nothing more than a diode isolating the send and receive pins and a jumper to put the ecu into diagnostic mode.

The hard part is finding a palm that will support the odd data rate the ecu uses. I know the IIIc will work but newer is always nice if it will work.

Sent from my SAMSUNG-SGH-I717 using AutoGuide App

[buickwagon]

Look what I found:

Mitsubishi 12 Pin to 16 Pin Female OBD 2 Cable Diagnostic Connector Adapter Tool | eBay

For $6 I'll cut the cable to get the connector. I have a couple of old Palm IIIc's. Now I just need to dig out a serial cable for it. I know I have a couple of cradles, but the cable might be better for this.

[speedy25]

Originally Posted by buickwagon

if the injectors were leaking and the O2 circuit was working properly, the system should be pegged rich. This is pegged lean.

I understand, but you shouldnt be smelling fuel at any time. I assumed you smelled it out of the exhaust but maybe its a system leak?

Sounds like you are on your way to making it right. Good luck for an easy fix.

-SP

[buickwagon]

Swapped in a new 02 sensor this evening. This one does work and flips up and down -- but only with the engine revved up, not at idle. So something else is strange, but it was getting late and dark. I need to assemble the cable for the palm so I can see what's going on with the fuel trim. For fun, we unplugged the MAF to see what would happen and did get a code + CEL. I wonder why it won't throw a code/light with the O2 sensor disconnected?

Checking around the back with a combustible gas detector, we found some of the gas smell may be coming from the seal between the tank and the sender/pump cover. We'll have to take that off and check it out, but for now we just sprayed the nuts and studs with penetrating oil.

There appears to be two connectors hanging loose between the firewall and the battery. The somewhat rounded one has a green wire with a black stripe and a black wire. The square one has a green wire with a red stripe and a red wire with a green dashed stripe. Anyone care to hazard a guess what these are for?

[jake stealth]

Hi Buick – Couple of thoughts

I removed the entire exhaust system and ran around with dual straight pipes for a week and even with a disconnected and totally missing O2 sensor, there was no check engine light – so perhaps you don’t have a problem there. I did not look for a code so cannot be of any help regarding that.

About the rich mixture and the lazy O2 sensor activity – have you looked at the hose between the intake manifold/throttle body and the fuel pressure regulator? It’s about 3/16 inch in diameter and if it’s cracked or disconnected, it will cause the fuel pressure to increase which might (?) cause the thing to run a little rich. During one (? LOL) of my several intake manifold removals I forgot to connect this hose and the engine did idle faster (vacuum leak) and the throttle response was not as crisp as when it was connected.

Regarding connectors behind the battery – there is one which is used to ground the ECU when setting the initial ignition timing and another one is to hot wire the fuel pump to see if it operates. However, I believe these to be single wire connectors. Some advice: I did use this fuel pump connector to prime the system and then started the engine. Turned the ignition key to “off” and the engine kept running until the fuel pump jumper was unhooked. Seems the fuel pump jumper also powers the injectors and the ignition – sort of like it bypasses the MFI relay.

As I remember, there is at least one multi-connector behind the battery that seemingly has no mate so I never worried about it. I’m wintering in FL and my car is in upper MI so I can’t check it out to be sure.

Good luck.