• Protip: Profile posts are public! Use Conversations to message other members privately. Everyone can see the content of a profile post.

Coolant Temp Problems

Joined
14 November 2003
Messages
6,139
Location
At the epicenter of the Zombie Apocalypse
Coolant Temp Problems with Fix

I searched but didn't find an acceptable solution so here goes...

My friend Ted's '96 NSX recently started exhibiting erratic coolant temp readings. Even after the engine should be completely warmed up, the needle barely registers 2 ticks off the bottom. If stopped at a light it will rise slightly but never near the normal indication of half way up the scale. As soon as he is moving again the temp drops almost to the bottom of the gauge.

The searches I came up with seemed to suggest the thermostat ws the culprit so it was changed at the dealer yesterday along with a complete coolant flush. He got the car back today and it's still doing the same thing.

Any ideas?
 
Last edited:
Sending unit?
 
Probably. Or possibly bad connection. Measure the resistance at known temp and compare to the chart in the FSM. If the sender's in spec measure it again through the harness.
 
There is no doubt this is a symptom of a bad thermostat. Are the symptoms EXACTLY the same now?

The gauge temp sensor, which is separate, can fail, and get erratic with age.

HTH,
LarryB
 
Larry Bastanza said:
There is no doubt this is a symptom of a bad thermostat. Are the symptoms EXACTLY the same now?

HTH,
LarryB

Hey Larry,
That's what I gathered from my earlier search. Yes the symptoms are exactly the same. Perhaps the dealer in typical dealer fashion didn't actually do any work.
 
Hugh said:
My friend Ted's '96 NSX recently started exhibiting erratic coolant temp readings.

Poor Putt-Putt! Hope all goes well with the repair Hugh. Maybe he doesn't like Florida and needs to return to AZ! :biggrin: :tongue: Say hi to Ted for us.
 
Hugh said:
Perhaps the dealer in typical dealer fashion didn't actually do any work.

Let's hope not.

But I will say this, if the car came to me I think I might be inclined to try another one if it acted like this. What I would do first is measure the output of the temp sensor used by the ECU. There is a resistence vs. voltage chart on page 11-48 in the on-line service manual for the temp sensor. That test will tell you if the temp is as the gauge states, or if it is steady. If steady, then look at the gauge sensor. The temp senor I am refering to is NOT the gauge sensor, but the temp sensor that monitors the temp in the thermostat housing.

After a look in the service manual you would have a code 6 if this semsor was not working. Sounds like it is. Just make you are "electrically inclined" to perform this:). Otherwise take it in to a competent shop.

HTH,
LarryB
 
He's sitting right here and says right back atya. Putt-putt also got a new clutch today. Last week she got her nose and ass re-painted. Still no Lambo doors but the Turkish bordello blue LED interior lights are still there!

Ted still loves the car as much as the day she arrived. For the life of me I can't understand why he likes to kneel behind it and lick the rim of the exhaust tips.
 
Larry Bastanza said:
What I would do first is measure the output of the temp sensor used by the ECU. I beleive there is a temp vs. voltage chart in the service manual for the temp sensor. That test will tell you if the temp is as the gauge states, or if it is steady. If steady, then look at the gauge sensor. The temp senor I am refering to is NOT the gauge sensor, but the temp sensor that monitors the temp in the thermostat housing.

Will try all of the above and let you know. Thanks Larry.
 
Since his car is a '96, if the ECT sensor was bad would there be an OBDII code for that?

I took a look and found the sensor but he just drove the car and it's hotter than Satan's scrotum back there. The manual says to start cold and watch the resistance until the engine reaches full temp and the fans come on so I'll do this tomorrow.
 
Larry Bastanza said:

OK, now I'm confused, do OBDII NSXs throw out the same codes as pre-OBDII cars? I thought OBDII codes were universal as described here: http://www.obdii.com/codes.html

However since the check engine light doesn't come on while driving I'm still inclined to believe the dealer somehow didn't do the the thermostat install properly. I just drove the car and the symptoms are still exactly as described in my first post.
 
I'm confused too. What does the ECT have to do with the gauge sensor? The gauge sensor is either right (bad thermostat) or it's wrong (bad gauge sensor).
 
ECT = "Electronic Coolant Temp" Sensor

The Code 6 is what you will see if you use the "blink" method of the CEL. If you use an ODBII scanner the code will be: P0117 or P0118

Sorry for the confusion, I happen to have the 1991 manual open when I wrote this last evening;).

The ECT has nothing to do with the gauge sensor, that is the point. You want to use an alternative method to check what the engine is really doing and eliminate the gauge from the equation. Since there is no CEL, the ECT is working fine and a good way to test for actual temp of the engine.

HTH,
LarryB
 
Ok, I'm no longer confused. I ran the diag as outlined on page 23-136 of the owner's manual. When shorting the yel/grn wire with the ignition on, the coolant gauge on the dash worked fine. When checking the resistance on the sender it was in the 400 to 700 ohm range. Much much higher than normal, therefore high impedance means very little gauge movement.

The reason there was no code is because there is a second water temp sensor that connects to the ECU and it would throw a code if bad. This why the car behaved normally even while indicating very low temps on the dash. The ECU was always seeing normal water temps even though the driver was not.

The sensor that connects to the dash gauge is dedicated to that gauge only and is not in the on board diag circuitry. Correct me if I'm wrong but I think the part Ted needs is 37750-PH2-014

This is all pretty much what Larry said in the previous post but perhaps a bit more "plain English" for those less techie types.
 
He drove over to my house this morning so we went to IHOP to have breakfast while his car got a chance to cool off a bit. We got back about an hour later and that's when I ran the tests. The car wasn't comlpetely cold but not too hot that I couldn't get my arm down to the sensor. (That damn Y brace on the T models SUCKS). Anyway the starting impedance was about 700 ohms and with the VOM connected, the ohms started dropping, only reaching about 400 when the fans came on. Obviously a bad thermistor.
 
Last edited:
I had my temperature gauge go bad awhile back. Readings were erratic and it was usually 2 to 3 ticks past cold. After changing out the gauge, everything was back to normal. The part was like $65.
 
I suppose that's always a possibility but according to the readings from the sensor it probably didn't need replacing. Normal readings are 142 ohms at 133 degrees F and from 49 to 32 ohms at 185 to 212 degrees F. This sucker was off by a factor of almost 10.
 
Hugh said:
What's weird? It's simply an indication that the part failed. It's a cheap little $20 thermistor.
I'm not clear on the setup. Did the fans turning on cause the resistance to drop immediately, or over time? I'm not questioning the approach (I suggested it in my first post), but the results are unexpected. I want to understand what's behind the diagnostic results.
 
Go to the online service manual to page 23-136. It's a very simple circuit. The sender unit that controls the dash gauge is a "thermistor" which means as temperature rises or falls it sends a signal of varying impedance (ohms). In this case as the temperature rises the impedance (or resistance drops). That's why the first part of the diagnostic is to ground the yel/grn wire going to the gauge. This should send the gauge flying up to a high temp reading. in the case of Ted's car, this is exactly what happened. So now knowing that the gauge and wire were good, it's time to attach an ohm-meter to the sending unit. I did this and got an initial reading of 700 ohms. We let the car warm up and I watch the ohms gradually drop to about 400 at which time the fans came on and I knew the car was fully warmed up.

Under normal conditions with a cold engine the sender should present a 142 ohm load to the wire leading to the gauge. As the water warms up the ohms drop and the needle on the dash gauge goes up. The range is supposed to be from 142 to 32 ohms. The range I got was 700 to 400 ohms explaining why the needle on the dash was staying pegged near the bottom of its range.

The second sensor, connected to the ECU was obviously working fine as there was no CEL code, the car ran fine was able to redline normally.

I worked with the coolant temperature sending unit, not the ECT as the absence of a Check Engine Light suggested it was fine.
 
Gotcha. I misinterpreted your other post to mean that it either dropped immediately or that it dropped to 400 ohms slowly after the fans came on. Neither one made sense. Thanks.
Another clue would have been to see where the rev limiter kicked in.
 
Back
Top