I'm looking for any more ideas. I got this code last week - I pulled the AIR hose by the one way valve on the driver side bank. When I started the car I could tell the pump was running from the air pressure in the line. I did not check the passenger side bank. Since the pump seemed OK, I reset the PCM by pulling the fuse (hoping for an anomoly). Today - second driving cycle - the light is on again. Can anyone offer anything else to check?

...and it does look like the update from the TSB has been done on this car. The hose routing looks like the diagram from Shoebox's website.

 

The diagnostic looks at the O2 sensors in open loop, to see if they are reading below 400mV, to confirm air flow from the pump. Then when the PCM enters closed loop, it turns the AIR pump on and looks for a lean condition on the O2 sensors and the appropriate response from the short term fuel corrections. In effect, the air that the pump is putting out must be reaching the O2 sensors. All hoses must be leak free, not plugged up, and the check valves must be opening up when the AIR pump is pushing air through the system. Your pump is running, but something may be preventing the air from entering the exhaust system.

 

This is my understanding - please correct me where necessary.

Open loop is after startup - until certain condition(s) are met - among them, operating temperature. After which the system enters closed loop. I thought that the AIR pump was only operational during the open loop - and where this diagnostic would fail. But as I understand now that you are saying it operates during closed loop as well?

Also, I just disconnected the passenger side rubber hose from the pipe going down under the motor (to the pump I guess)- when I run the car I have suction - not positive pressure. Is that right? I thought I shoud have positive pressure on the steel pipe to force air through the valve and into the manifold. What am I missing? By the way, the car was warm when I did this test.

 

Correct.

The AIR pump operates only during closed loop startup, to preheat the cats. While its running, the PCM checks to see if the excess air being pumped into the exhaust is causing the O2 sensor voltage to drop below 400mV, indicating the lean mixture that results from the secondary air injection. This check is only made once per trip.

When the PCM enters closed loop, it is no longer necessary for the AIR pump to run. BUT.... to check the operation of the AIR pump, the PCM runs a diagnostic routine when the PCM enters closed loop. The pump is not being turned on for any reason other than to verify that it is functional. The test is run up to 3 times per trip, and takes 3 seconds to run. The PCM looks for the O2 sensors to drop below 222mV for a period of 1.5 seconds, or for the fuel integrator (short term correction) to change by 14 counts when the pump is turned on. The AIR pump is run in closed loop ONLY for the diagnostic test, not for any other purpose.

The AIR pump will pulll a suction on the rubber hose from the air inlet ducting to the inlet nozzle on the pump. Everything after the pump should be under pressure when the pump is operating. That includes all the plumbing from the pump discharge to the exhaust manifolds. There should not be a vacuum in the line, UNLESS the check valves are not functioning correctly.

If the AIR pump is off, and the engine is running, the passage of exhaust across the connection for the AIR line can cause a negative pressure.... Bernouli's principle.... and the can educt fresh air into the exhaust system, throwing off your O2 sensors and causing the engine to run exessively rich. Just like an exhaust leak. In an exhaust system with high back-pressure, its possible for leaking check valves to pressure the AIR system. I believe the AIR pump recall components included a purge valve to prevent any exhaust gasses from flowing into the AIR pump, and causing corrosion.

 

Very detailed explanation. Thanks.

From what I gather, there are two check valves - one on each bank above the valve covers. I understand your explanation that the suction would occur on the exhaust manifold side of the check valve - my observation was that the suction was on the pump side of the check valve, not the exhaust manifold side.
Come to think of it - if I understand your reference to Bernouli's principle, wouldn't a volume of air enter the exhaust at all times? The check valves operate to prevent reverse air from the exhaust back to the pump under pressure, but admit air freely into the exhaust manifold through the pump plumbing. Even if the pump is off, air must be drawn through - or is there another valve that closes when the pump shuts off?

Where exactly is the pump intake?

 

The pump pulls air from the 5/8" rubber hose connected to the engine's air intake ducting, between the filter box and the MAF sensor. Consider that if you have a dirty air filter, your air intake ducting could be at a negative pressure, and also providing some vacuum to the AIR system.

If your check valves are working, they spring closed and take the AIR pump pressure to open them, not the small vacuum experienced due to flowing exhaust.

You need to run the pump and verify that you have air flow to both banks of the engine, and that this air flow is getting past the check valves.

 

Injuneer, you are an absolute fountain of knowledge. Just one more Q and I will leave you alone, promise .
I am now wondering whether when I checked the DS bank the other day - I might have been feeling the same negative pressure instead of positive pressure . Is there any (relatively easy) way to jump the AIR pump from a 12v source to confirm it is operational?

 

If you go to Shoebox's website, you can find the wiring diagrams for the 1995 PCM. On page 4 of 7, at the top is the wiring for the AIR pump. I'm assuming that your 96 is similar. The pump is turned on by the PCM supplying a ground to the E2 pin on the relay. Disconnect connector C100 and apply a ground to that wire. Alternatively, you could apply +12V to pin E4.

http://shbox.com/1/4th_gen_tech2.html#1995_pcm_wiring