Temperature guage "failure mode"

Having a problem with the temp gauge on my 04. When it is cold out (below 40 F) after driving my 996 for a short while (10-15 minutes) the needle will drop to 0 and the red warning light will start to flash and the dash will display failure mode. Doesn't happen all the time and when the ambient temp is warm it never happens, only when the air temp is cold. And after continuing to drive for another 5 minutes or so the gauge will return to normal operation. Last week for the first time when this happened (and it was in the low 30's) it threw a CEL, The code was P0128 (Coolant temp below thermostat regulating temp). I cleared that code and the CEL has not occurred since; although yesterday I had another temp gauge drop and warning light episode. Coolant level looks fine. Last year I had the reservoir replaced and a complete coolant flush. Cleaned the grounds, the contacts on the cap on top of the reservoir and the contacts on the temperature sensor.. No coolant leaks from the system that my tech or I can find. Replaced the thermostat shortly after that in case that was contributing to the problem, no change. It's stock and I'll be installing a 160 next week when it is in for service. I'm thinking that the problem may be a coolant temp sensor but before starting to swap out parts i wondered if anyone else has ever experienced this or has any thought on what else to look for.

No thoughts, except that you are probably correct in assuming the temp sensor is likely culprit. Other thought is that testing would be futile unless the test can be done at the moment during the failure warning, testing at any other time would be pointless, that's why intermittent issues are the most difficult.

A flashing red light indicates low coolant level. If your coolant isn't low then it is likely that the sensor needs to be replaced. It's about $30 and not difficult to change and doesn't require draining any coolant.

I dont think its a coolant level issue, that's fine. And from what I can tell that wouldn't account for the temp gauge dropping to zero at the same time. And a few minutes later everything reads normal, including no flashing light. I'm going to have the sensor changed to see what happens.

Quotes source post: 15630267

Yes, sorry. I missed that about the gauge dropping to zero.

No problem, this situation has several strange symptoms. Didn't know if I had not fully understood the low coolest alert. So the zero reading sounds like it's not part of the low coolent alert?

Quotes source post: 15630166

Temp sensor makes sense. One addition, the flashing of the temp light has different meanings depending on the frequency of the flash. I don�t remember offhand, but it should be easily located online or in the manual.

In the owners manual is reads that a steady red light indicates excessively high temp and a flashing red indicates too low a coolant level. So my symptom would indicate too low a coolant level, except the warning light goes off after a few minutes. And the reservoir is showing coolant between min and max at rest, even checked right when the light is flashing. In the Bentley Service manual it describes a slow flashing light (1/2 sec cycles) as low coolant or engine compartment temp too high. But coolant level is good and it only happens during cold ambient temps, and not when air temp is hot. I wouldn't think the fact it is cold out would negatively impact the engine compartment fan and after the compartment temp warms up the light would go off, unless possibly a temp sensor that actives that engine compartment fan is failing cold. A rapid flash indicates the temp sensor faulty. I'm not seeing a rapid flash. more of a slower flash. And with the rapid light the needle should be pinned to the right, mine is to the left. But, a check of the fan relays and also the engine compartment temperature sensor might be in order also. I hate intermittent problems !!!!

Just to eliminate the possibility, is your engine compartment temp sensor properly secured?

what about an air pocket trapped in the system?

Picked up my 996 from having it serviced yesterday. As part of the service (transmission fluid, oil change, 160 thermostat, etc) I had them try to fix this issue based on what i had learned here and in the shop manuals. They thought it was more likely the reservoir level sensor based on the "frequency" of the flashing light than the temperature sensor, the slower flash would indicate low coolant level rather than temperature. They checked the engine compartment fans and sensor, all worked normally. The system was recently pressure bled and doesn't appear to have any air pockets. So took it for a spin this morning while it was cold. Bingo, light came on, gauge dropped to 0 (actually 100 but that's as low as the gauge goes to) and after about 5 minutes it returned to normal. So back to the shop to have the temperature sensor replaced to see if that solves it.

So it's not the temp sensor, just realized it was replaced just last year (5k miles ago), also repaired the ground associated with this sensor at the same time. Summary on repairs: New temp sensor, new coolant tank and cap (done as preventative maintenance when i got the car), new coolant tank "level sensor", system has been pressure bled, new 160 thermostat (also occurred with a 180 thermostat). My service shop has never seen this problem so everyone is at a loss for next steps. It doesn't appear to be a "real" problem, but i don't like ignoring error conditions as can mask a real problem.

Is yours a MkI? If so, there are two sensors inside the coolant temp sensor. One feeds the cluster and the other one the DMEo so 4 wires total. It may give us more clue if you can read the coolant temp sensor on an OBDII scanner, which gets info from the DME. Also, have you checked the x59 connectors. All the coolant temp sensor connections go through those connectors (or one of them). May want to clean those pins of the connectors. Do you have the wiring diagram and plan to self troubleshoot?

Thanks, Ahsai. It's a MKII, but I'm starting to think the wiring harness may have a connector issue. It's just the fact it only happens when the engine is cold is what throws me for a loss. I will pull out the wiring diagram and look for connectors. Is there a diagram that shows where they physically are in the car?

The wiring diagrams from the WSM have a diagram showing where each connector is (using 3D co-ordinates). The X59 connectors are at the far right deep in the engine bay. Since yours is a MkII, it only has a single coolant sensor. If it's not the sensor, the next logical things to check are the wiring and connectors between the sensor and the DME. Since you have an intermittent problem, you will need to have a test setup where you can monitor the signals in real-time while you drive. E.g. back probe the DME side and the sensor and see how the signals behave when the fault occurs.

These are the X59/1 and X59/2 connectors.The top one has blue dots on the connector and the bottom one has white dots. I forgot which one is X59/1 and which is X59/2. The coolant sensor signal (blue/pink wire) connects to pin 12 of X59/2, then to pin B12 of the DME. The ground of the sensor (brown/pink wire) is merged into the ground connection of other sensors.

So an update on this long-running problem. Last year the shop read the temps with their PIWIS while driving and the temps all looked fine, no drop when the gauge dropped and alarmed. There were several respected shops that suggested this might be a software glitch in the ECU. So in January i removed the ECU and immobilizer and shipped them to the ECU Doctors to see if they could diagnose a problem. The unit tested just fine. They even left it in a freezer for a while to see if the cold would cause an error, since i only see the problem during cold periods. Nope, still tested fine so they don't think there's anything wrong with the ECU. Its on its way back and I'll reinstall it and see if the situation still occurs. It doesn't seem to be an issue that creates any problems for the engine, it just thinks the engine coolant temp is too low, but only for a few minutes then the gauge reads normal. Someone suggested that i install a new factory temp sensor just in case the recently installed sensor is bad. It's been changed twice with no effect. But I'm thinking of doing that, doesn't seem like a hard task or big $. Neither of the local dealer service techs had any suggestions. But they would be happy to spend a lot of time ($$) trying to diagnose if I would leave them the car for a while.. Naw, passing on that offer. I can't test drive the car for a while as it has a bad AOS and I"m waiting on the Ultimate AOS to install before i drive it much, just in case. Sending the immobilizer to ECU Drs at least gave me a reason to buy their waterproof case, just in case i forget to raise the top on my cab before i take it thru the automatic car wash. One thought i've had is the external temp sensor in the front bumper. I removed it when i installed bumper grills and wonder if it might be damaged and contributing to the issue? Can't discuss with the PO if this was a problem earlier as he is deceased. I'm open to any and all suggestions.

After I read through this...that was my first thought....the ambient air sensor in the front bumper. You removed it? Or is it laying somewhere behind the bumper?

I removed it from the mount to install the grill and then remounted it. It's in the normal place. But I can't find any info on whether or not it effects the ECU or just influences the HVAC.

The ECU uses a variety of sensors to monitor the engine and set fuel trims (cold-richer mix; warm-leaner mix). The ambient air sensor is one of those. The ECU needs to know the ambient air temperature for its algorithms to determine how the engine is running given how warm or cold it is outside, how warm the engine is, coolant temperature and oil temperature. Then it sets fuel trims, spark duration, etc.

Maybe it's just the gauge that's causing the issue.

Quotes source posts: 16442355, 16441386

I thought about the gauge and did clean all the connectors behind the gauge. I don't think a faulty gauge would fail momentarily only under certain outside environmental conditions (low external temp and engine temp just starting to warm up) and then correct itself after a few minutes. It makes some sense to me that the external temp sensor might be at fault, but it only occurs at low temp and for a few minutes. Then the condition clears itself even if the external temp hasn't changed. So why would it do that if the ambient temp sensor was failing? Wouldn't it always be not functioning?

Anyone get this sorted yet? I have same issue in my 2003 C4S. Started happening shortly after having engine rebuilt to 4.0 liter. It only happens randomly, no real rhyme or reason. I also had the sensor replaced and it still happens. I purchased a bluetooth OBD2 adapter that I use to monitor the temp in real time just to make sure. I use the Torque Pro app and the temp is always good.... Looking for solution still.

I haven't solved it yet. I haven't had a chance to drive it after reinstalling the ECU since I'm waiting on the "Ultimate AOS" to replace the failed AOS. Hopefully in a few weeks it'll be back on the road and I can see where I am. The coolant temp seemed to be responding ok when it was monitored with PIWIS, maybe I need my own monitor just to be sure. My engine is.a rebuilt to 3.8. You don't see any consistency when it happens? External air temp? Time since starting the engine or distance driven? How about when it goes back to normal?

No consistency, completely random. My Porsche tech is also baffled. I will try to "burp" the cooling system just in case an air bubble is causing the failure indicator. If not, maybe wiring issue who knows where? It is a mystery still.

Quotes source post: 16474528

Mice love to chew on wires, did the car sit for any length of time? Maybe have it checked.

My 04 c4s cab has been doing this for yrs, 5? Needle drops to zero, no warning lights or codes. If I stop at a light it goes back up. I just leave it alone, though maybe I'll have the tech look at it next time it's in the shop.

Quotes source post: 16475113

Yeah, I've had some chew thru the speedometer wire on my Tacoma once. But, this is intermittent, occurring under very specific temperate/time conditions and is highly predicatable. That's why i ruled out wiring.

Quotes source post: 16478480

Hmm. So you don't get the warning light on the temp gauge display? I was able to monitor the engine temp with PIWIS once and the temp itself didn't change when the warning light occurred (around 150 F) when it was <45F outside air temp..

Quotes source post: 16478516

One thing that comes to mind after I read this thread, is that the DME has an "expected warm up time map" programmed into it. It monitors the engine temp during warn up and compares it to the "time map" and if the engine warms to fast or too slow it will flag it. Not a lot is known about the map other than it there and compares the "actual" warm up temps , to the "predicted" warm up temps. Maybe knowing that and thinking about what could cause the engine to warm too fast or too slow can figure out what's going on here..

There may be a clue in there since this only happens to me when it is cold outdoors, thus the engine would warm up a little more slowly. Where would such an expert be found? Complicating this is that my engine is far from stock.

Quotes source post: 16479112

The only clue I can think of is yours in a Tiptronic and has an additional heat exchanger for the trans, that has a coolant shut-off valve in-line. Perhaps if the shut-off valve is not functioning as designed, it could cause the warm up temps to not follow the "predicted" pattern.

So there is a solenoid that opens a valve to circulate coolant to the transmission cooler, That solenoid is controlled by the Tiptronic ECU, Which gets info from a sensor in the Tip "electronics" package - part #16. Maybe if this part is bad it would "trick" the ECU into thinking the coolant temp is too cold. But wouldn't this all be confined to the Tip ECU and not the engine ECU that controls the temp gauge? Or maybe the TIp ECU has to tell the engine ECU to set a temp alert? My head hurts

Quotes source post: 16480065

Not Quite, the coolant shut=off valve is just a coolant valve like was used to control heater water flow, actually its the same part with an old "928" part number92857457303. and uses a standard change-over valve to turn it on and off via signal from DME. I actually remember having had diagnosed and repaired a Carrera Tip with identical symptoms (seems like a Century ago). What happens is the shut-off valve fails, (rubber diaphragm) then coolant is allowed to flow into the Trans heat exchanger when cold, the cooler trans fluid delays the warm up time of the engine coolant and the condition is red flagged.

Yeah, i saw that diagram, it's part # 8 right? Another rubber diaphragm failure, where have i heard that story before? I was thinking it might be the sensor that controls the solenoid.But a failed diaphragm could let coolent pass by when it shouldn't. That looks like a fairly easy to replace part. $82. I ordered one.

Quotes source post: 16478516

No light, no overheating, 6 speed, I live in S Fl so it's not a cold weather thing for my car. It happens whether its warmed up or been running for a while. Seems like if I come to a stop or change gears it shoots back up.

Quotes source post: 16480207

Yep, part #8, easy to replace or test. It's coming back to me now, same symptoms as yours, was a PITA to figure out. I thought I would never forget, but I almost did, lol shows I'm getting old !!!

Replaced the transmission coolant shut-off valve today. Won't be able to see if that solves the problem until i get the UAOS installed so i can drive the car. At the rate I'm going on that project, the UAOS won't be done until summer..... Oh well, I can always wait and see what happens next winter.....

Quotes source post: 16531377

You can easily test the shut-off valve. Just push the lever up, then put your finger over the nipple to seal it off, release the lever and it should stay up. If it falls straight away, the diaphragm is busted and not holding vacuum.

I just went to the garage and tested the old one that i took out and it passed, so that wasn't the cause of the temp gauge issue. Dang, but now i have a new one installed. But, what about the little plastic line that provides the vacuum. I wonder if it might have a leak and not pull vacuum to cause the diaphragm to open for the coolant flow?

Quotes source post: 16531999

Yes, the plastic line that provides the vacuum, and also the change-over valve (vacuum solenoid) could be defective and cause same issue, the default position is full flow, so any failure in that circuit will cause the issue.

Thanks, Skip. So it fails "open", which makes sense so the transmission doesn't overheat. That could introduce cold coolant into the engine sooner than expected, although no drop in the coolant temp was observed.. I'll try to trace back the plastic hose first. Should be a lot of fun under the car.

Quotes source post: 16534299

Exactly, the system fails "open", that means coolant is flowing through the trans heat exchanger at all times. During cold temps, the engine is suppose to warm up first and fast, if the trans fluid is still very cold, it will actually cool/delay the warm up time of the engine. Then once the engine is warmed, the DME is suppose to open the shuff-off valve to allow the coolant to keep trans fluid temps down.

Am update to this thread. Last April, at Skip's suggestion, I replaced the coolant valve and solenoid for the Tip and also replaced the vacuum line to the valve. The weather has cooled off enough here to finally test the repair. Last two days got some runs in the low 50's, this morning one in the low 30's. No alarms! No temp gauge dropping back to zero, no flashing temp alarm, no error message! So it looks like something in that system was the culprit! Probably not valve but maybe the solenoid or a leak in the vacuum line. Very strange situation and really obscure fix, but, thanks to Skip's encyclopedia memory of experiences, looks like that problem is fixed!!!! Thanks, Skip! Here's to you!

Quotes source post: 17090066

Thanks for the update/conclusion of the Technical Thread ! So many get left hanging with no conclusion ,it's nice to see the ones who come back with updates/conclusions !! And thanks for the props to my old *** brain for not succumbing to the "Sometimers Disease" and still has some usefulness...lol A unique problem indeed, that can only happen to 996 TIP cars, which are a very small percentage of the Porsche cars produced, and can only be on the later model Tip with the "warm-up time map" programmed into the DME.

Quotes source post: 16474504

I find it curious that the previous two reports of this issue both had their motors rebuilt to 3.8L & 4.0L. I have had this issue for a few years now, a few months after getting the car back from being built to 4.0L. My vehicle is a manual so, the solution the OP found wouldn�t apply in my situation. After about a year, my temp gauge completely stopped working and just flashes red all the time.

Thanks much for the reported conclusion of this thread, az968gpw , Porschetech3 ! Not sure the engine replacement has to deal with it - listing my observations and relative car mods for reference: Mine is 2002 996.2 cabrio* 6-speed manual California car that rarely reproduced the issue in So-Cal but now that I relocated back to Eastern Europe, it did so during the winter. I do have the 0.5L extended oil sump capacity which I had noticed did keep the engine slightly cooler at operating temps - at near highway speeds, the arrow stays slightly left of the middle of the 0 (only if I drive in traffic does it get to the middle of the 0 of the 180). As far as I understand from this thread, the issue is with a module programmed for the Tiptronic version of the car but reproduced in the manual version only in colder outside temperatures. * my engine was replaced under warranty by Porsche in Jun-Jul 2005 with the latest stock unit at the time which for my benefit turns out was a 997-like one with the larger built-in-lower-block, non-replaceable IMS bearing (NOTE: I don't think the IMS has any relevance but just mentioning here for discoverability for other folks who wonder as the engine SKU was still 996-*-* and I had to spend money to drop the tranny to have a look at it before going to the track with a 17yr old car at the time and an engine with 110K miles )