P2138 code and intermittent limp mode!

[Lando217]

Hey guys my goblin has the 2008 LNF ss doner.
recently when I have been driving around, my goblin has been intermittently going into limp mode. It can be a 2 min drive or a 15 min drive when it kicks in. I turn on the car and turn it back on, then it’s fine again for another 2min-15min.

It’s throwing a p2138 code-“Throttle/Pedal Position Sensor/Switch D/E Voltage Correlation”

which in my understanding is that the accelerator pedal and throttle body sensors are not synced up? Iv replaced the pedal because many people said it was a common problem on the cobalt. I replace the throttle body with the built in TPS cause it was also throwing up a TPS code that is no longer there.
Iv checked the connections on both ends and Every thing looks good with no corrosion.
It happens in random times, accelerating hard or constant speed doesn’t make a difference. One time it seems to kick into limp mode then back out.
before I do diving completely into the wiring to the PCM I was seeing if anyone any advice?

It’s hard to tell If this was always a problem since the process of the build, because before recently completely finishing my goblin I have only drove it around the block a few times and to my parents once (about 3miles) with no problems at all.
Thanks!

 

[jamesm]

If you replaced the throttle body, you will most likely have to allow the ECM do "relearn" the TB. I would also check the output of the pedal sensor and check all the grounds for the pedal, ECM, and engine harness.

 

[Rttoys]

Yes, will probably need a relearn. A tech2 or good generic scanner like a snap-on can do this. HP Tuners might be able to, too.

 

[Lando217]

My buddy owns a shop with those expensive scanners, I’ll bring it to him and see if can do a relearn.

Never a bad idea to check grounds again lol

I don’t know if anyone else has experienced limp mode but I wanna make sure that’s what my car is doing.
Limp mode won’t let me boost and driving with engine power only?
Also on my TC I noticed the harness to the pedal had a 4in long two connector pig tail that seems like both ends are the same. What’s the propose of that? To extend the harness 4”?

 

[Ark :D]

My buddy owns a shop with those expensive scanners, I’ll bring it to him and see if can do a relearn.

Never a bad idea to check grounds again lol

I don’t know if anyone else has experienced limp mode but I wanna make sure that’s what my car is doing.
Limp mode won’t let me boost and driving with engine power only?
Also on my TC I noticed the harness to the pedal had a 4in long two connector pig tail that seems like both ends are the same. What’s the propose of that? To extend the harness 4”?

That's a good question. It seems to be agreed upon that all it does is adapt one connector to a different one, ie, with the wires reversed. It's probable that GM has multiple throttle pedal assemblies and that adapter is just used when needed.

 

[Chubbs]

Reviving another zombie thread here, and hoping maybe someone has some more information.

I'm chasing check engine lights. Very recently I finally defeated a p2229 code, but right around that time this p2138 code popped up. The throttle body was replaced when I built the car, so I have no reason to think that's the problem. I just replaced the pedal assembly, And that didn't kill the code either. The entire car got the old wire jiggle test when I replaced the shifter and had access to the center tunnel. No difference. Have re-seated the throttle body connection, and of course the pedal connection when I replaced it.

It only seems to pop up when the car is idling, but I can't quite replicate it. It's seemingly random, but it never comes up while the car is under power. Only while idling in the yard, or a couple of times while decelerating for a long off ramp.

I've read about the relearn, but it seems like the relearn procedure is simply letting the car idle for a while, and driving it, which I've done. Is there an actual relearn process?

 

[Joebob]

I had this code as well and I did as you did as well. I first replaced my TB because my donor's was hit in the wreck and replaced with a cheap non GM one and though that might be the issue. After a bunch of research it is common in the Cobalts due to windshield leakage corooding the plug and wire at the accelerator pedal. From that I guessed it is most likely a wiring problem. No amount of jiggling or reinstall worked for me.

Next thought was we extend those lines in the tunnel so for the pedal connector's 6 wires, we have 12 splices of which any of them could have increased resistance and screw up the sensor readings causing a fault. There are two redundant sensors at the pedal and the TB so that an errant sensor doesn't floor the car into a wall. if they two do not match whithin a narrow window, the car throws it's hands up and goes into limp mode.

The solution for me after testing the pedal system was to buy a replacement plug with pigtails, and a 10' length of shielded 6 conductor cable and rewired that system back up to the ECM plug to eliminate the previous old wires, connector, and any bad splice job I might have done. Fresh solder joints 6" before the ECU plug and I have not had any issues from it in about a year.

Good luck.

Joe

 

[Joebob]

Oh, and since the tunnel is open at the front and back, you can get a cheap fiberglass wire fishing tool and pull the new wire through without removing the tunnel top (also why I went with a 6 conductor wire bundle).

18/6 shielded cable

fishing tool

pedal plug for 2008 SS/NA

 

[Chubbs]

Thanks for the info, I fear that's the direction I'm heading as well. A little foresight would have saved me a bunch of time....I just had the tunnel open to replace the shifter . Would have been easy to run new wires a week ago, but now it's all riveted back together.

I did some prelim wire probing yesterday. The two "inner" wires on the pedal connector are both ground wires, with good continuity to ground. None of the other wires are shorted to ground, but I didn't yet check resistance from one end of the wires to another.

 

[Gtstorey]

If you run new wires, add a pull string for next time.

 

[Metal Mech]

Thanks for the info, I fear that's the direction I'm heading as well. A little foresight would have saved me a bunch of time....I just had the tunnel open to replace the shifter . Would have been easy to run new wires a week ago, but now it's all riveted back together.

I did some prelim wire probing yesterday. The two "inner" wires on the pedal connector are both ground wires, with good continuity to ground. None of the other wires are shorted to ground, but I didn't yet check resistance from one end of the wires to another.

This is why I love threaded rivets. This kind of thing always happens to me. That or I finish up something and second guess myself and want to double check.

 

[Robinjo]

I had this code as well and I did as you did as well. I first replaced my TB because my donor's was hit in the wreck and replaced with a cheap non GM one and though that might be the issue. After a bunch of research it is common in the Cobalts due to windshield leakage corooding the plug and wire at the accelerator pedal. From that I guessed it is most likely a wiring problem. No amount of jiggling or reinstall worked for me.

Next thought was we extend those lines in the tunnel so for the pedal connector's 6 wires, we have 12 splices of which any of them could have increased resistance and screw up the sensor readings causing a fault. There are two redundant sensors at the pedal and the TB so that an errant sensor doesn't floor the car into a wall. if they two do not match whithin a narrow window, the car throws it's hands up and goes into limp mode.

The solution for me after testing the pedal system was to buy a replacement plug with pigtails, and a 10' length of shielded 6 conductor cable and rewired that system back up to the ECM plug to eliminate the previous old wires, connector, and any bad splice job I might have done. Fresh solder joints 6" before the ECU plug and I have not had any issues from it in about a year.

Good luck.

Joe

I'm currently fighting the same P2138. It started off as a Accelerator Pedal Low Voltage and replaced with a new Accelerator Pedal Assembly. Then it popped up as a Throttle Body Low Voltage MIL and replaced the TB. Finally came up as P2138 so a new Accelerator Pedal Jumper and I'm still chasing the P2138 code. Don't bother removing the P2138 from HP Tuners, it still gives 'Reduced Power Mode' but won't show a code. I guess my next step is to rewire like you have done. It's really, really random. I can go for days without getting the code and then get it a bunch of times in a row. I've even gotten my hands on a new to me LSJ engine harness since I read that the connectors at the TB can be an issue. I'll try your solution first as I can just resell my engine harness for what I have in it.

This is why I love threaded rivets. This kind of thing always happens to me. That or I finish up something and second guess myself and want to double check.

My tunnel and rear firewall are held on with these M4 rivnuts and screws. I knew I'd be back into the tunnel at some point and here we are..... The part I'm most dreading out of the whole ordeal is pulling the seats. I have the stock SS seats and they are HEAVY.

 

[Chubbs]

I thought maybe I found the issue...maybe.

I disconnected both the PCM connector and the pedal connector, and checked continuity between them. At this point, all 6 wires to the pedal should be completely isolated from everything....one end of the wire at the pedal, and the other end at the PCM. All wires had continuity (less than .6 ohms resistance). Then I checked continuity to ground, and surprisingly, they weren't completely infinite. 3 of them were; the two ground references (remember they were disconnected, so they were no longer "ground") and the dark blue signal wire, were all OL on my meter. The other three, however, had very high resistances to ground. The white/blk 5 volt reference was 1.5M ohms. The Tan 5 volt reference was more than 5 M ohms. And the light blue signal wire as 0.9 Mohms. All very high....but not infinite.

Talking with my local electrical expert, those values should be as good as open for a 5 volt system. The chance of any electrons escaping through a 900k ohm connection at 5 volts is almost none. But maybe?

I yanked on the wires a bit, and eventually got all of them to go open loop on my meter. Then drove the car and....the light came back immediately. So I guess that's not the issue?

Don't bother removing the P2138 from HP Tuners, it still gives 'Reduced Power Mode' but won't show a code.

This saddens me greatly, because that was my next step. Not that I would leave it like that, but it would be nice to at least drive it for a bit. So frustrating.

 

[Robinjo]

This saddens me greatly, because that was my next step. Not that I would leave it like that, but it would be nice to at least drive it for a bit. So frustrating.

I was also hoping to temporarily eliminate my problem until I had something more concrete to try. I guess that more concrete is now, I needed to get into the tunnel anyways, so this gives me an good reason to go in. Also, now that I have it all hooked up and running decently I can reorganize my wire layout. I'm going to start adding my experience here as well, so it'll be in one place on the things you and I try. Maybe one of us will get something solid.

On the diagram listed in AllData for the APP and TPS, does the number (0.35 & 0.5 in this case) mean thickness in mm?
0.35mm --> which is about 22 AWG
0.50mm --> which is about 20 AWG
Since the drawing lists both, which size should I go with. Is there really that big a difference between 22 and 20 AWG? I'm a novice at best with wiring, so I do not know all these nuances in it, what is ok/NOK and what not.

 

[Gtstorey]

C201 is an intermediate connector (probably removed during the rewire, that is why it show both wire sizes. It never hurts to go up in wire size but in this case the smaller size would work. I would use 20awg or larger just because 22awg is small enough to be hard to work.

If the code is now constant, the diagnostic steps in the manual should lead you somewhere. They aren't always 100%, especially some of the older ones, but they are probably 90%. They aren't nearly as good when the problem is intermittent, especially if you can't pull freeze frame data.

 

[Robinjo]

C201 is an intermediate connector (probably removed during the rewire, that is why it show both wire sizes. It never hurts to go up in wire size but in this case the smaller size would work. I would use 20awg or larger just because 22awg is small enough to be hard to work.

If the code is now constant, the diagnostic steps in the manual should lead you somewhere. They aren't always 100%, especially some of the older ones, but they are probably 90%. They aren't nearly as good when the problem is intermittent, especially if you can't pull freeze frame data.

Thanks for the information. Would there be a problem jumping up to 18-AWG as Joebob did?

The problem is intermittent but annoying. It can occur a bunch in a day and then go for days or weeks without issue. It's very random, unfortunately. It'll also come on then remove itself sometimes.

Also, since I'm rewiring the whole thing should I look to remove the jumper harness from the ACC pedal? All it does if inverse the wires to the pedal. I could easily do that and cut out one junction in the wiring harness.

 

[Gtstorey]

No problem with jumping up in wire size to 18awg.

I don't know what you mean by "jumper harness".

Everything I've read about the APP and TAC trouble shooting indicates it can be hard to track down since everything (and there are several things it is checking) has to match very closely, so the ECM and TAC doesn't run away at the wrong time.

 

[Robinjo]

APP Jumper Harness. It connects the wiring harness to the Accelerator Pedal. I assume the wring was generic between Cobalts (and other GM models) as was the ACC Pedal. Then, all the needed was a 'jumper harness' to correct the wiring outputs. Everything basically makes an X to the other side. These were used in at least LSJ vehicles, do the other models not require this?

 

[Joebob]

One issue I had but not related to the P1238 code but also put me in limp mode was loose pins on the BCM plug. @Desert Sasqwatch refered to a drag test on the plugs. I found that I had 2 pins on the BCM plug that were on the tan/tan+black network lines that would cause all sort of TC< (auto) comminucation and cruise control codes. If the wires all check out, make sure the ECU plugs are good too. I found some paper clips that were the same diameter as the ECU pins and checked for resistance to insert and disconnect. If some are loose, you will get the intermittent issue even if the plug itself is properly seated. I went to a junk yard and pulled a bunch of plugs to repin after replacing any loose connectors. Wouldn't hurt to check before pulling the tunnel. Don't forget that you have a clear straight path to fish a new cable through without pulling the tunnel if that is the next step. I was able to do it without opening the tunnel, just had to work under the front, push the fishing tool through and pull the cable from the engine back to the front.

Joe

 

[Joebob]

It seems other models do not have the jumper harness. Mine as a SS/NA was straight to the wire harness. Have you checked the jumper for wiring issues?

Joe