NefMoto

Been having a play with craigs S4 today, and decided to read off the fault codes, which highlighted a fault code for the coolant sensor:

17664 - Engine Coolant Temp Sensor (G62): Open or Short to Plus

However, the coolant sensor isnt that old, and appears to work perfectly

Viewing the measuring blocks on the ECU showed the sensor reading 19c (which was within 1c of the outside ambient temp), and upon starting the engine it sensibly rose to 92c over a few minutes of driving then fluctuated around there by a few degrees, exactly as you'd expect...

Cluster also works fine and reads as expected.

Anyone seen this before?


My only thaught its that its a G-box ECU, running D-box code, but the same G-box ran C-box code previously without any issues. Thinking back, the code appeared when we reshelled the car, which involved changing the ECU from C-Box to D-box code so it was compatible with the ESP on the '01 shell.

It just doesnt make any sense. If the sensor was shorted out like the fault code suggests, then it wouldnt be able to give perfect readings on the measuring blocks. They'd just default out to a standard value?

check wiring, replace the sensor and check for codes. just because the sensor is newish doesn't mean it isn't faulty. As I understand it, the sensor also could only be "half bad" CHeck VWVortex forum and search for  bad coolant sensor issues. It's in there somewhere, I just don't have the link.

 Also, could the sensor have a revision because it was faulty from the factory like on the 1.8ts?? Us 1.8t peeps had to change from the "black top" to the "green top" coolant sensors because of problems with the black one FROM THE FACTORY..maybe the same happened with the 2.7t. Good luck.

As i've said in the post though, the sensor readings are prefect.

The sensor cant be "open circuit or shorted to plus" as the fault code says, but then also give spot on accurate readings. If it was broken in the way the fault code describes, it would just be permenantly reading a default failsafe value.

For whatever its worth, I once had a crankshaft position sensor that was intermittent open/short. That was Bosch sensoor from a tdi.

If you clear the codes, does this error come back instantly?

Yep, pops back immediately, and its never tagged with "intermittent".

I need to test what the ECU does if i unplug the sensor, but i would imagine the measuring blocks would stop showing the correct temperature, and just show some default failsafe value.

When you unplug the sensor, the code should pop up as soon as you power up the ECU( turn the key).

What is the exact code number?

Just so you know, if you have a fully registered VCDS then you have free support from Ross-Tech.

Also the Ross-Tech Wiki has a list of DTC codes along with what could be causing the code and what to check to fix it.

Hope this helps.

With the sensor plugged in, the code reappears immediately after clearing codes. I didnt try unplugging the sensor.

I dont see what it has to do with rosstech? Its not a registered version though.

The point i'm struggling with, is that IF the fault code was genuine, then the ECU measuring blocks wouldnt/couldnt show the correct temperature...

If the sensor was shorted out, as the code suggests, then it cannot provide a reading, yet it is providing a perfectly accurate reading.

This isn't exactly your problem, but just something I wanted to add. If you unplug the MAF sensor from a  MK4 2.0 VW engine it will still show values in the measuring block in g/sec just like it would if it were plugged in. The same with the intake air temp (which also happens to be in the maf sensor on this particular engine). For sure on newer models if you unplug a coolant or intake air sensor, it will default to -42 c or some other completely unachievable number to show there is no sensor info available. Looking forward to see what happens when you unplug the sensor..Is it also possible it can get this info from the other coolant sensor(s) mounted on the lower coolant hose?

Coolant hose "sensor" is just an electrical switch, so nothing to be got from there.

I guess its possible that its getting the reading from the instrument cluster (which has its own seperate sensor, inside the same CTS unit as the engine ECU)

I guess the problem then is that unplugging doesnt actaully tell me anything, as unplugging it would kill both the cluster and the ECU feeds.

Anyone know if the ECU has the ability to read coolant temperature information from the cluster?

Maybe just unplug the cluster?

Yeh that would require rewiring the KLine though, as it runs thru the cluster as std.

I would change out the sensor first thing if I was sure there wasn't a short in the wiring itself. The part is only like 9 dollars, and you can just take it back if you need the money. You don't even need to to install it, just plug it in and check for DTCs! While its out in the air see if the sensor shows the same temp as the ambient air temperature by comparing the values to a thermometer.
 
So I'll ask again...what DTC code are you getting??

The point with Ross-Tech is if you have a registered version, they will help you diagnose the problem. Go on their forum and post a full scan. They won't give you the time of day if you are using the unregistered version with a cheapo "Ebay" interface. 

If none of this works, reflash the ECU to the file you had before your current one to see if it is indeed the software. I highly doubt the software is the issue though.

i had a this error code in my 1.8t. i changed sensor but problem cannot be solved. i changed thermostat, and problem has been solved.
this is strange but true :)

Yes! I remember reading about this in the VWVortex forum! Because the thermostat was malfunctioning, the coolant sensors were "out of sink" so to speak with each other. The ECU knows what both sensors should read at certain coolant temps and if the ratio of the temperature reading of the coolant sensor in the radiator compared to the coolant sensor in the coolant flange (bolted to the head) are off, things get screwy...

Ok this isnt my car, its my brothers, so i cant just nip outside and check things.

The code i have is in the first post.

I've got a spare CTS that works fine, so next time i'm over at his i'll plug it in and see if the code goes away.

I guess it could be a short in the wiring. But i dont understand why i've got a perfect readings in the measuring blocks. If the wiring was faulty, or the sensor was faulty in the way that the ECU is describing, it COULDNT provide a reading. A shorted out sensor cant provide a signal.

Which means either the fault code is erroneous, or its real but the ECU is pulling the data from the cluster?



The thermostat issue is a different one, and gives a different fault code. As i said above, there is no "Sensor" in the radiator. its just a switch and the only thing it does it power a relay for the fans. What the ECU does is have a map which says the engine should warm up in x minutes, and once warm it should stay at 90c. With a faulty thermostat, the engine either never warms up at all, or the temperature rises and falls with speed. This causes the ECU to realise something is wrong, and flags an Implausible Signal fault code. This is NOT whats going on with the S4, so please dont try and confuse matters talking about thermostats.