Code 21 throttle position sensor high voltage, how do I fix this?

Tiny
GILBRO
  • MEMBER
  • 1990 CHEVROLET TRUCK
  • 5.0L
  • V8
  • 2WD
  • MANUAL
  • 150,000 MILES
Got a code 21 throttle position sensor high voltage. How do I fix this?
Monday, September 18th, 2023 AT 5:26 PM

13 Replies

Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
How is your knowledge of electrical theory, and do you know how to use a digital voltmeter? If you do not, I can help with that. This circuit is relatively simple, but I need to know where to start so we don't overlook anything. Here's the diagram we'll be using.
Was this
answer
helpful?
Yes
No
Monday, September 18th, 2023 AT 6:07 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
My electrical knowledge is not too good. But I do know how to use a voltmeter. The diagrams look like they fit my truck.
Was this
answer
helpful?
Yes
No
Friday, September 29th, 2023 AT 8:37 PM
Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
Dandy. The first thing to be aware of, and this is especially important in this type of circuit, voltage readings are only valid when taken with everything connected and plugged in. This is a little less critical on the 5.0-volt supply wire and the ground wire, but you WILL get a wrong reading on the signal wire. I'll explain that in better detail shortly. If you look at the back of the throttle position sensor's connector, there's a rubber weather seal around each wire. In some cases there's an extra cover that must be removed, but the idea is to stick a voltmeter probe through that seal, alongside the wire. If the probe is too fat, use a stretched-out paper clip to reach the back of the terminal.

First let me explain the fast way to test this circuit. That's to measure the voltage on the signal wire. That's the dark blue wire. If you get the proper results, you're done. That can only happen when the 5.0-volt feed and the ground wires are okay, and the sensor is okay.

The better way for now is to look at each wire. The first one is the 5.0-volt feed from the Engine Computer. The ignition switch must be in the "run" position. Ground the meter's negative probe to a paint-free point on the engine, body, or the battery's negative post. Back-probe the gray wire. You should find 5.0 volts. Note on the diagram they show that wire also feeding the MAP sensor, so if there's a problem with that wire, you'd have a MAP sensor code too.

Next, measure on the black, ground wire. As a point of interest, it doesn't actually go directly to ground. It goes back to the computer, through some monitoring circuitry, then to ground. That way the computer can monitor that circuit. Due to that circuitry, it is common to find just a little voltage, typically 0.2 volts, on that wire. Don't be thrown if you see that. This ground wire is also used for the coolant temperature sensor. If there's no fault code set for him, that ground wire is most likely okay too.

The secret to setting the code you have is on the signal wire, the dark blue one. This sensor is exactly like the volume control in older tvs and car radios. It's a carbon strip with a movable contact that picks a spot along that strip. The big difference though is this sensor has mechanical stops inside it that limit how far that movable contact can go. It's picking a place somewhere between 0 and 5 volts, but those stops limit it to picking a signal voltage of between 0.5 to 4.5 volts. Those numbers are for training and explanation purposes. In actual practice, you may find 0.42 volts, for example, and 4.37 volts. Those two values are not critical, but the important point is they can't go all the way to 0.0 or 5.0 volts. Those are the two values that trigger fault codes.

You should find something close to 0.5 volts with the throttle closed. Watch the meter as you slowly run the throttle to wide-open-throttle. As you do, the signal voltage should rise smoothly to around 4.5 volts. If that is what you find, the fault code likely set due to the computer seeing a "dropout" or glitch that occurs much too quickly for us to see on a meter. Even readings on expensive scanners don't respond fast enough in most cases. This fault code will also set if you unplug the sensor while the ignition switch is on. When that happens, in most cases the Check Engine light will turn off the next time the ignition switch is turned on after the plug is reconnected, but the fault code will remain in memory. On Chrysler products, if the defect doesn't occur again, the code will self-erase after 50 engine starts. I suspect something similar will occur on GM models. If it doesn't, disconnect the negative battery cable for a couple of minutes.

If you find 5.0 volts on this signal wire, there's either a break between the mating terminals in that connector, or a break inside the sensor. You can test the sensor by doing continuity checks between its three terminals. A faster check is to pop in a different sensor. These have a very low failure rate, so I'd be happy to harvest one from a salvage yard.

If there is a break in that signal wire, logic says you should have 0.0 volts on it, which would set a fault code for "TPS voltage too low". This is where I mentioned you WILL get a wrong reading if you unplug the connector. With a break in that wire / circuit, there's a whole bunch of interconnected circuitry inside the computer that could let some stray voltage "float" to some random value. If that random voltage were to remain between 0.5 and 4.5 volts, the computer will accept it and try to run on it. Engine performance will not be right, but there will be no fault code to tell you where to start the diagnosis. To prevent that, all computer sensor circuits use either a "pull-down" or a "pull-up" resistor. Many import models use pull-down resistors. Those are connected between the signal wire and ground. They are so big electrically that under normal conditions, they have no effect on circuit operation. It's only when there's a break in the signal wire that they drag the voltage down to 0.0 volts. That's outside the acceptable range of 0.5 to 4.5 volts, so a fault code is set.

Most domestic models use pull-up resistors. Those are connected between the signal wire and the internal 5.0-volt supply. Again, they're so high in resistance, they don't affect the circuit until there's a break in the signal wire, then they put 5.0 volts on it to force a defective condition that can be detected. You get the fault code, "TPS voltage too high".

If you get this TPS high code repeatedly or it won't erase, this signal wire is most likely where you'll find the cause. The other possible cause is a break in the sensor's ground wire or a break inside the sensor itself. Remember, the coolant temperature sensor shares this ground circuit, so if it isn't setting a code, the break has to be in the sensor, between its mating pair of terminals in its connector, or in the wire leading up to where the two ground wires are spliced together. All of those are rather rare, so we'll cover that if testing shows that's where we need to go.

There's one last thing we can look at but it requires a scanner or a code reader that can display live data. Some of those are very inexpensive but they usually only work on '96 and newer models. This test is simply to see if your voltmeter shows the same voltage on the signal wire as the scanner shows for TPS voltage. If the voltmeter shows the correct values, between 0.5 and 4.5 volts, but the scanner says the computer is seeing 5.0 volts, there's a break in the signal wire. In this case I'll let you unplug the connector to the TPS, then you should find 0.0 volts on the signal wire. The 5.0 volts the scanner is showing is due to the pull-up resistor, but that 5.0 volts can't make it past the break to get to the meter.

When there is a break in a wire, it is common for whatever caused the damage to affect multiple wires in that area, and there would be additional fault codes or dead circuits. With just the one fault code, the best suspects are the sensor and the connector terminals. Terminals can often be identified by wiggling the connector and seeing the voltage change on the signal wire. Often you can use a pick to bend the terminals to make a tighter connection.

Start with those three voltages and tell me what you find.
Was this
answer
helpful?
Yes
No
Friday, September 29th, 2023 AT 11:00 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
Okay. So, I back probed the gray wire using neg on battery with ign on and got 5.04 v. I also traced the wire n according to the image u sent it also goes to MAP sensor, mine goes straight to computer. Measured the black wire got 0 volts n checked for continuity to battery neg and temperature sensor there is continuity. Measured blue wire with throttle closed n got 4.17 volts. Opened throttle slowly and voltage came down to 0.3v wide open. I did resistance test with it unplugged my numbers are, between gray and black (closed) is 1.8 resistance and open 5.3 Between gray and blue 4.8 n didn't move when I opened throttle. Blue and black got 5.0 closed and 1.3 resistance fully open.
Was this
answer
helpful?
Yes
No
Saturday, September 30th, 2023 AT 9:42 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
I should also mention when I start the truck idles high for a while then when it finally calms down anywhere from 1 mon to 5 minutes the service engine light comes on. It stutters when I'm driving and just not running right.
Was this
answer
helpful?
Yes
No
Saturday, September 30th, 2023 AT 9:48 PM
Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
The 0 and 5 volts are correct. The 4.17 and 0.3 volts on the signal wire would be fine except they're backward. The voltage must be higher at wide-open-throttle.

Without knowing the history of the truck, I would guess this is the wrong throttle position sensor or the black and gray wires are switched in the connector. I looked as far as a '95 model, which is the same, but there are no drawings given to show the connector view. They did point out this sensor is on the right side of the throttle body assembly, which is not normal. That means it's likely the sensor rotates backward compared to most of them. I could see where the wrong sensor was able to be installed.

By reading backward, there's another symptom I would not be surprised to hear about. That's hard starting or a long crank time. The high voltage you're finding at closed throttle is "clear-flood" mode to the computer. It's turns off the injectors during that time. If you press the accelerator pedal and that makes the TPS signal voltage go down, that would get it out of clear-flood mode. Once the engine is running, the TPS has very little to do with fuel metering calculations, but it will cause symptoms related to direction and rate of change, meaning things like surging or stumbling.
Was this
answer
helpful?
Yes
No
Sunday, October 1st, 2023 AT 6:30 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
What would happen if I did change the wiring around? Anything major in a bad way? My TPS is on the right side if you are sitting in the truck. Just checking before I start my process of elimination. Thanks for your time and knowledge it really is appreciated.
Was this
answer
helpful?
Yes
No
Sunday, October 1st, 2023 AT 9:58 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
I do get a long crank before it starts.
Was this
answer
helpful?
Yes
No
Monday, October 2nd, 2023 AT 8:46 AM
Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
You won't hurt anything by switching the 5.0-volt and ground wires. What I would question though, first, is if it's the wrong sensor. It has a spring-loaded lever that is pushed by a tab on the throttle blade shaft. If there is such a thing as a mirror image version, I would suspect that lever won't get pushed by the throttle blade or the sensor can't even be bolted on without force. If I had to take a guess, I'd say a better suspect is someone replaced the connector and attached the wrong wires to the new ones on the connector. We know the lever is getting pushed by the changing voltage you found on the signal wire as you work the throttle.

I have to defer here to my knowledge of Chryslers, but they do most things the same way as GM does. That is, if you were to do anything to cause the 5.0-volt feed wire to become grounded, for any reason, the Engine Computer will shut that power supply down to protect it. The computer does not get damaged. GM had all kinds of computer problems through the '80s and early '90s, which is why still to this day so many people incorrectly jump on a computer failure as their first suspect, but shorting the 5.0-volt supply isn't one of those failures. What you will have to do, once the short is removed, is to turn the ignition switch off, then back on, to reset the 5.0-volt supply.

By any chance, to you have access to a scanner or to a code reader that can display live data? If you do, I'd love to know what it shows for TPS voltage at idle and at wide-open-throttle. It should agree with your voltmeter readings.
Was this
answer
helpful?
Yes
No
Monday, October 2nd, 2023 AT 4:50 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
Well, you were right. I had the wires crossed. Again, thank you. Very much appreciated.
Was this
answer
helpful?
Yes
No
Wednesday, October 4th, 2023 AT 10:21 AM
Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
Dandy. So is the fault code gone? Is the engine running properly?
Was this
answer
helpful?
Yes
No
Wednesday, October 4th, 2023 AT 6:50 PM
Tiny
GILBRO
  • MEMBER
  • 12 POSTS
Yes, on both questions.
Was this
answer
helpful?
Yes
No
Wednesday, October 4th, 2023 AT 8:16 PM
Tiny
CARADIODOC
  • MECHANIC
  • 33,752 POSTS
Very happy to hear you solved this. Please come back to see us with your next insurmountable problem.
Was this
answer
helpful?
Yes
No
Thursday, October 5th, 2023 AT 11:37 AM

Please login or register to post a reply.

Sponsored links