Air Pump Won't Shut Off

[Kestas]

Unfortunately, Arthur Dalton and I were in the middle of diagnosing my problem with the air pump running continuously when my thread got lost in the server crash.

To recap, the air pump on my 95 E320 runs continuously. Per Arthur Dalton's advice, I learned that my car does not have a switchover relay in the relay compartment. The codes I got were:

5, 19, 21 (on-board diagnostic module)
13, 49 (Pin 8)

Arthur last wrote:
Unplug the pump and run the car a couple of days to
see which codes
clear by themself. specially the 02 code
Also, check that the purge valve is beating [Search]
How is harness?????????

Engine wiring harness was replaced two years ago.

I found the purge valve was plugged with granules (activated charcoal?). I cleaned it out and got it working with 12v and some tapping, then gave it some exercise with 12v and WD-40. Now it has a healthy heartbeat. But the air pump still runs.

After running the car with the pump unplugged I got only Code 13 (Pin 8). With the pump connected I get the same code, which is the only code I get after reading/clearing codes a few times. So now I'm looking at Code 13 as the clue to the problem.

Additional info: O2 sensor was replaced two weeks ago

[Arthur Dalton]

Was it a Universal 02 sensor??
Is it plugged in??
Are the wires correct??

Is the car going into Closed Loop??

[Kestas]

The O2 sensor is a Bosch 13725 spliced into the old wiring. It did get rid of a check engine light problem I had three weeks ago. But all the same error codes persisted. Splicing was done correctly.... black with black... gray with gray... both white wires with white. This was done on the bench with solder, shrink tubing, and tape wrap.

Closed loop?... we may be on to something. Lately, the temp gauge only goes to 80-85°C. Before it was always 90-95°C. I tought it was due to the cooler weather. This is my summer car. I've never driven it this late in the season. Bad thermostat?

[Arthur Dalton]

The concern would be if the ECU is seeing cycle cross counts from the 02 sensor.. the AIR pump will not run once the car has gone into closed loop
condition, which is dependant on 02 sensor output.
I would do an 02 sensor check with a DMM , unplugged from ECU, while taking the hose off fuel pressure regulator to see if 02 responds to rich and create a vac leak for lean...
it is also possible that you have a shorted driver circuit in the ECU AIR switching transistor, but I would check 02 output first..

also check ECT sensor values

[Kestas]

I get roughly 0.5v from the O2 sensor with my analog multimeter. This is with the engine warmed, O2 sensor disconnected from the rest of the car, and leads on the black and gray wires (at the connector, after the splices). Opening a vacuum line will cause the voltage to drop almost to zero momentarily. Disconnecting the vacuum line from the fuel pressure regulator also causes fluctuation.... rises to 0.55v and drops momentarily (probably from the interruption in vacuum).

I have yet to measure the ECT. What do I look for?... a certain resistance value, or is this an on/off switch?

[Arthur Dalton]

>>

You have to plug the vac line when doing an enrichment test.. the 02 should give a high V reading when disconnecting the fuel reg. That is what you are doing.. richening the mix..

Then see if you ghet cross counts when the 02 is connected back to ecu..

[Kestas]

I checked the ECT. I get 3800Ω when cold, 230Ω when hot.

I checked the O2 sensor again. I get movement that averages around 0.5v when idling. Breaking a vacuum line lowers the voltage to below 0. Pulling and plugging the vacuum to the fuel pressure regulator raises the voltage to 0.65v. High enough?

I don't know how to measure cross counts after plugging the O2 sensor back into the ecu. Have I reached the limit of my DIY capabilities?

[Arthur Dalton]

Cross count is every time the voltage crosses the .45v line .. looking for 20 in 10 secs.
Sensor should read higher .. try cuffing intake a bit with reg hose off...