PDA

View Full Version : Coolant temp sensor bad?



sixfiveoh
05-27-2010, 05:34 PM
A long time ago I remember my coolant gauge dropping all the way to cold off and on, I figured it was the ect sensor, but then it normaled out and I forgot about it. Today while driving the coolant temp high warning indicator popped up and the needle started creeping to H, pulled over, topped off the resivoir with water, and kept going. Still overheating. Checked everything out, the resivoir seemed fine, no boiling water, no steam, fan seemed to be working (just one of them, is the other a secondary?) so I drove home "overheating" figuring it was the ect sensor again. Just to make sure I let the car sit for about an hour, started it again, needle went straight to H

I called audi when I got home to schedule a service appointment to replace the sensor under the recall, told them what was up, they said it wouldn't be covered unless I had a cel from the ect sensor, and that my symptoms "didn't sound like those of a bad sensor". I can't remember once getting a code from this anytime I've ever scanned my car, but I figure it has to be faulty. There's a $170 diagnosis charge so I figure I better know for sure before I take it in. I'll scan the car later tonight, but if it's not the sensor, what else could cause this?

walky_talky20
05-27-2010, 05:44 PM
If may not set a code. The ECT sensor is actually 2 sensors in 1. One is just for the ECU, the other is for the Instrument Cluster. They can (and do) fail *independently*. So the "ECU side" could be operating just peachy, but the "cluster side" is going nuts.

What you can do is use vag-com to read Measuring block 004. This will show the ECT reading for the ECU side of the sensor. Then compare that with the reading from the cluster. If they don't agree, BINGO - bad sensor. If they *do* agree, but the reading is impossible (ie: both are pegged HOT, but the car is stone cold), then you also have a bad sensor. For the best comparison, you should check it cold and hot, and when it is acting up if possible.

sixfiveoh
05-27-2010, 09:17 PM
Well I ended up pulling the implausible signal code, so that's a relief. Thanks for the info though, good reference. [up]