P1133 DTC tough to solve, 2000 Outback MT

[avk]

There is a number of threads related to this code, "Front Air/Fuel Sensor Heater Circuit High Input" and I have looked through them. The problem I have is that the code is still there after replacing both the sensor and the ECU: first the sensor and then, when the code was still there, the ECU. This is on a friend's vehicle. If I erase the code, it comes back as soon as the engine is started, even though FSM says it's a two-trip code.

I did some electrical testing prescribed by the FSM and found the voltage on the ground side of the heater, which is controlled by the ECU, to be indeed out of spec.: 9.5 V to 11.5 V while it should be below 8 V.

The supply voltage is 14V with engine running.

This might look like something is shorted but poking around with a meter and shaking the harness doesn't point at any fault so far.

Anyone have any insights on this one? Thanks.

[OB99W]

Before going into this in depth, did you use an OEM Subaru sensor for the replacement?

[avk]

Oh yes. I was making use of my Subaru Bucks, unfortunately with no success so far. The part numbers are 22641AA00A for the sensor and

22611AF60ERB for the ECU (remanufactured, which is what the catalog calls for).

 

Add: another piece of data I have is the resistance of the heater element. It's about 1 Ohm at room temperature. I understand it's

a normal value for an A/F (wideband) sensor although regular type should read about 5 Ohm. The resistance is the same between old

and new sensors.

But if 1 Ohm were, in fact, too low, it would explain the short circuit but also imply that something caused both sensors to go bad,

perhaps the old ECU if it itself was bad. Then we'd just need to change the sensor one more time.

There's no specification for the heater element resistance in the FSM, other than it has to be below 10 Ohm. But Toyota was using

this type of sensor in late 1990s-early 2000s and their manuals give the figure of 0.8 to 1.4 Ohm which is what I measure.

If someone can measure the resistance of a known good A/F sensor, between the terminals connected to two black wires, it would be hugely helpful.

Edited May 5, 2009 by avk

[ron917]

Look for a poor ground connection in that circuit.

 

With electrical issues like this, the problem is often hard to find but easy to fix once you've found it.

Edited May 5, 2009 by ron917

[avk]

Thanks, we did check the ground to ECU. It's the only ground connection involved. The heater current is controlled through the ground side. There is a wiring diagram posted in one of the past threads.

Edited May 5, 2009 by avk

[OB99W]

[...]I did some electrical testing prescribed by the FSM and found the voltage on the ground side of the heater, which is controlled by the ECU, to be indeed out of spec.: 9.5 V to 11.5 V while it should be below 8 V.

The supply voltage is 14V with engine running.

Specifically where were the above voltages measured and to what ground reference?

[avk]

Specifically where were the above voltages measured and to what ground reference?

The voltages in the heater element circuit were measured relative to the body ground, as directed. Specifically, one of the meter probes was in contact with the bare-metal ECU bracket. The resistance between the bracket and the ground wires coming into ECU reads zero, so that location was as good as any. The two voltages we measured relative to the ground were (1) the constant power supply voltage, coming from the battery through the main relay, which read 14V and (2) the voltage on the other side of the heater element, in the line between the ECU driver circuit and the heater. That voltage was seen switching in the range of approximately 9.5 to 11.5 V. The driver is pulse-width modulated, so the pulsing is normal, but the number is too high according to FSM which instructs to look for a short circuit to the battery side, but we haven't found one yet, unless 1 Ohm for the heater is really out of spec.

Both voltages were measured at the corresponding pins on ECU connectors (the power supply for the sensors also comes into ECU).

Edited May 5, 2009 by avk

[OB99W]

As I see it, there are only two possible scenarios for the ''low'' side of the heater to be at too high a voltage:

 

1) The resistance across the heater is too low, due to the heater itself or wiring.

Since you've replaced the sensor, its probably unlikely that its heater is the cause. There could be a battery-side wiring short -- you should be able to check for that by disconnecting the ECU and A/F sensor, and measure resistance across the connector pins that would go to the heater.

--or--

2) The ECU is incapable of pulling the heater's low side close enough to ground.

Hopefully the problem isn't within the ECU itself, since it was also already replaced. A wideband A/F sensor's heater draws a lot of current (you noted the low resistance). Therefore, it's critical that the path from the heater's low side to ground isn't excessively resistive. I realize that you've checked ECU ground wiring resistance and found it to be ''zero'', but only a few tenths of an ohm can make the difference between the heater's low side being at what you've measured or the less-than-8-volts spec.

[Cougar]

After reading through the posts on this problem it does sound like the return side of the sensor is too high for some reason and the return side is isolated from true chassis ground. Since you have logically replaced both of the suspected components and still have the same problem this may mean something like the power source isn't right. Have you checked to see if the alternator is putting out any AC voltage across the battery while the engine is running? There should be less than .1 volt of AC across the battery if the alternator is ok. That might account for the voltage difference you are seeing. If that is ok then check the resistance between the return wire and the positive battery cable. Disconnect the positive battery connection and the connection to the ECU before measuring it. There should be no connection between the two points, an open circuit. If there is some resistance seen then you need to clear the unwanted connection to that return wire.

[avk]

Many thanks to both of you for taking your time to think about this problem. Let me begin with saying that it did go away after I opened up the harness trying to chase the short circuit, but I am not sure what could have been the cause. I also believe that FSM is incorrect on this one. More details are given in the comments below.

1) Since you've replaced the sensor, its probably unlikely that its heater is the cause. There could be a battery-side wiring short -- you should be able to check for that by disconnecting the ECU and A/F sensor, and measure resistance across the connector pins that would go to the heater.

That was done: we unplugged them and checked the resistances for shorts and continuity. The ground-side wire from the heater actually splits in two (there's a splice close to the firewall on the interior side) and connects to two pins on the ECU. This is probably because a single pin there is too small to handle the current. So we tested this wiring and didn't find anything. At one point it seemed like there was something wrong with the spliced connection and we decided to open up the harness. We unwrapped it both on the inside and on the outsdie. It was all in plain view except for the portions inside the grommet seal and another one where it runs up the firewall behind the AC box. Nothing was visibly amiss and voltage and resistance readings did not change when we were pulling and twisting the harness. To gain room for removing the tape in the engine compartment, we unplugged the ECU harness from the two big square connectors on the engine. Maybe that's what made a difference, but when we plugged everything back together to take readings with the harness unwrapped and started the engine, the CEL was gone.

--or--

2) The ECU is incapable of pulling the heater's low side close enough to ground.

Hopefully the problem isn't within the ECU itself, since it was also already replaced. A wideband A/F sensor's heater draws a lot of current (you noted the low resistance). Therefore, it's critical that the path from the heater's low side to ground isn't excessively resistive. I realize that you've checked ECU ground wiring resistance and found it to be ''zero'', but only a few tenths of an ohm can make the difference between the heater's low side being at what you've measured or the less-than-8-volts spec.

This indeed might have to do with the root of the problem because the ECU is grounded through ground connections on the engine. That is, the ground wires in the ECU run to those square engine connectors which we unplugged and then plugged back together. If there was a bad ground, it could have fixed it. I didn't expect these plugs to be a trouble spot, but the car had head gaskets replaced under extended warranty some years back and perhaps the connectors weren't plugged back all the way. Can't be sure about it, but that's all I can come up with.

 

After reading through the posts on this problem it does sound like the return side of the sensor is too high for some reason and the return side is isolated from true chassis ground.

The book says 8V is the limit but the only way we can get a reading below 8V is with the sensor unplugged and key on, engine not running. Now with CEL off, the readings are the same as before if not higher, so I don't even know what the correct diagnostic procedure should be. Any comments from the pros on this one?

No surprise people have been having such trouble with this code.

Did I say EJ25 sucks?

Have you checked to see if the alternator is putting out any AC voltage across the battery while the engine is running? There should be less than .1 volt of AC across the battery if the alternator is ok. That might account for the voltage difference you are seeing. If that is ok then check the resistance between the return wire and the positive battery cable. Disconnect the positive battery connection and the connection to the ECU before measuring it. There should be no connection between the two points, an open circuit. If there is some resistance seen then you need to clear the unwanted connection to that return wire.

I didn' measure the AC pulse but will do for a good measure. The rest has been addressed above. Thanks again and if the problem should come back or I have any more ideas, I will post an update.

What's funny is that there's a strong chance that both the sensor and the ECU were good to begin with. I wonder why the replacement ECU is only $200: maybe they want us not to feel too upset about it.

Edited May 5, 2009 by avk

[Cougar]

You're welcome for the help.

 

I suspect that the trouble may have been at the SMJ connector(super major junction, as Subaru likes to call them). Hard to say now that it is working. You may have been picking up some extra noise on the return wire due to a slightly bad connection. Check the power for any AC ripple with the engine reving.

 

The return wire isn't really chassis ground since it is so high above the DC reference ground.

[OB99W]

Many thanks to both of you for taking your time to think about this problem. [...]

You're welcome, glad to offer theories.

 

 

[...] the ECU is grounded through ground connections on the engine. That is, the ground wires in the ECU run to those square engine connectors which we unplugged and then plugged back together. If there was a bad ground, it could have fixed it. I didn't expect these plugs to be a trouble spot, but the car had head gaskets replaced under extended warranty some years back and perhaps the connectors weren't plugged back all the way. Can't be sure about it, but that's all I can come up with.

It wouldn't surprise me if the problem was related to those connectors, since very little ''extra'' resistance will limit the current the heater can draw. Poor contact between mating pins is certainly a possible cause. I've also run into a marginal connection between a wire and a crimped-on pin or splice. In some cases I've extracted pins from connectors and soldered the crimps.

 

 

The book says 8V is the limit but the only way we can get a reading below 8V is with the sensor unplugged and key on, engine not running. Now with CEL off, the readings are the same as before if not higher, so I don't even know what the correct diagnostic procedure should be. [...]

I could be wrong, but I suspect that you used a DVM (digital voltmeter) for the measurements. In the case of a voltage that's rapidly switched (PWM or whatever), a DVM's reading will be dependent on the circuit's switching rate versus the DVM's sampling frequency. If you have an analog voltmeter available, you might see what the reading is using that.

 

I find it's often easier to locate poor connections that handle significant current by checking for voltage drop across the connection. For example (should the problem recur), I'd try measuring voltage between the ECU ground pins and a known-good engine ground point -- any voltage seen between them should be very low.

[avk]

I did use a DMM as shown in FSM, if it can still be trusted. The pulse frequency isn't great, something like two per second. It makes sense, because it takes time for the sensor to heat up or cool down, that is there's a reaction time involved. But even if the readings aren't quite accurate, it's still puzzling that they didn't change much at all. Of course it would be interesting to see the actual waveform.

Adding to the confusion, I found that the code P1133 is defined differently for AT models. With MT, P1133 is "high input" but for AT it's listed as "low input". The other code is P1132, also with opposite meanings for MT and AT. I even looked through the available corrections, now free at techinfo, and didn't find any changes to this.

But apparently, this particular problem affects mostly MT vehicles from model years 2000 and 2001 (in the Legacy line at least) so the confusion with code definitions does not come into play.

Edited May 6, 2009 by avk

[Cougar]

Have you checked for AC voltage across the battery while the engine is running, yet?

[avk]

I'll do it as soon as I can get my hands on that car.

[avk]

I measured the ripple, or tried to measure I should say. With the meter set for AC, there was no meaningful reading. Same thing happened when I did it for comparison on my minivan. I suppose one needs a true RMS meter to get a good reading. It would jump between 0 and 35 V, so it's not even the DC bias that throws it off. Same on both cars. On DC voltage setting, there was fluctuation in the range of few tenth of a Volt.

The CEL stays off for now.

[ron917]

If you can borrow an oscilloscope, that would tell you a lot more than a DVM.

[avk]

It just doesn't look like there's a problem with the alternator.

[Cougar]

I assume you were getting the 35 volts when you first set the meter probes across the battery and I could understand that. It does sound like the alternator is ok to me also since you didn't continue to see the voltage. Hopefully the code will stay away.