My ‘86 NA is really making me scratch my head. Had the typical “ac causing car to overheat problem.” Already ordered a radiator since I planned on taking care of that anyways, but the old temp sending unit for the gauge was so crusty and nasty I decided to replace it anyways and check how the car is running one last time before tackling that job.

it’s been into the 116F mark in Phoenix almost constantly over the last few days.

anyways, threw the new ending unit and all initially seemed very well. Around just above the first mark, low speed fans kicked on, lower radiator hose from water pump to rad is hot.

let the car idle for about 10 minutes with the AC off, all readings nominal, including oil pressure.

ran the AC, and after about 7 minutes of idling, though the temperature gauge reads a healthy temp, the idiot light went off. I shut the car down immediately. IR temps on the top cylinder head of about 100C. Saw 99 at the gooseneck as well. But on the gauge sending unit I saw 96. AC off turn car back on, temps on the cylinder head and the sending unit of about 92C dropping back into the upper 80s, but the idiot light is still on. I have no idea what’s right or what the next move should be.





Sending unit with idiot light on

Idiot light still on

Light on

Any Ideas?

also not sure if it matters but for what it’s worth I have the CSF spal fan kit and shroud on my car. So not stock fans.

 

Left the car off this time for about an hour, idiot light went away. Started it back up and let it idle for about 20 minutes. These are the temperatures I recorded, along with the corresponding gauge position. Idiot light never turned back on during this idle period.



Gauge reading while measuring engine temps.

Cylinder 1

Area around cylinder 2

Area around cylinder 3

Area around cylinder 4

Upper rad hose

Lower rad hose.

no boiling steaming overflow no smelling hot but it certainly feels hot. Fans kick on low speed at first then stay on high speed. They also do remain on in low speed after the car is off for a period.

the gauge never reaches half way but the temps here indicate that seems to be correct. It seems the only time any issues arise is when the air conditioning is engaged. I just want to make sure the car isn’t actually overheating.

the car reached those temps after about 10 mins of idling and then stayed there for the rest of the time I let the car run indicating to me that if the temps are not continuing to climb, the system must be at equilibrium. Is my logic right or do I need to go back to thermodynamics class?

 

I'm not well versed on the later cars, but I looked at the schematic. It appears to me (and I'm happy to be corrected) that there are two water temperature sensors. One located at A49 (NTC Water) and the other located at E46 (water temp sensor) on the schematic. Maybe you changed the wrong one.

 

Although this is a common source for confusion I changed the right one. I changed the DME sensor about a year ago. The sending unit is for the gauge. The DME has a fuel injection plug on it, this one has the two spade connectors, one for the gauge itself, the other for the warning light. They are not interchangeable.



Here’s the old one in the packaging the new one arrived in

Thanks for bringing that up for future readers though as it can be confusing.

 

A link to a video measuring all the temperatures. Ambient was about 107 in the garage.

Let me know if anything seems out of the ordinary. According to Clark’s these temp readings are what they should be, but with the odd behavior I am hesitant to trust the reading on the gauge.

 

Test 1 82c brand new sending unit

Test 1 brand new sending unit 82c gauge indicating below 80c

Test 2 brand new sending unit 85c

Test 2 brand new sending unit 85c gauge reading just below 80c mark

Test 3 brand new sending unit 92c

Test 3 brand new sending unit gauge finally reaches 80c mark.

Test 4 old sending unit 78c

Test 4 old sending unit 78c reading just above 60c on the dash

these are the results from temperature testing both sending units. They both seem to be reading way too low. I will repeat these tests and record resistance readings. Then I’ll try the Clark’s method to test the gauge. So far, not looking good.

 

I read through your posts and I am still trying to figure out what your problem/question is. It sounds like you have some concern about the mismatch between the actual coolant temp and what your gauge is showing?

Corroded wiring, age or grounds is likely why you have a mismatch between the gauge and the actual. Also check the back of the cluster and the black oval connection point. Im not sure what you mean by "idiot light" as I dont think there is a warning light to indicate high temp or overheating. 100C is within normal operating temp - usually where high speed fans kick on but it depends on what thermo temp switch you are running. Different thermo switches kick the fans on at different temps (usually stamped on the side of the thermo switch).

Slight mismatches between the gauge and actual is pretty common especially due to the age of the cars, mine is the same way. Its not something I am overly worried about personally as my gauge never surpasses the 100C mark anyways



If the car is overheating with the AC on, you need to test the fan relay, resistors, thermo temp switch, and the fans themselves. Test with AC on and AC off to see if for some reason its causing your fans to stop running (most likely reason why you are overheating). It could also be electrical related for some reason the AC on causing the gauge to read off. The procedure is on Clarks (Cooling Fan Operation and Troubleshooting)

 

The main issue that started all of this is after getting my AC fixed the car started indicating that that coolant temp was too hot. All those components you mentioned work, I have a focus 9 tech solid state relay, and the CSF spal fan kit. What I usually see when I don’t have the AC on is the needle sitting right above the second line where it should be. AC on= overheating.

So then I started taking the IR temps, and after comparing to that temperature photo I noticed something was off, because even if the car was sitting at 100c, my gauge continued to climb until the warning light came on. If I shut the ac off and continued to drive, the car would cool down within 30 or so seconds- 1 minute.

I ordered a new temp sending unit thinking that would help, but after installing it the new issue I had is that the gauge would now read “healthy” temps all the time, but with the air turned on, the warning light would illuminate even though the needle indicated around 90c.

I returned that sensor yesterday and cleaned up my old VDO one and reinstalled it. I think you’re right about corrosion maybe causing a resistance reading. I did my best to clean the contacts but the next thing I’ll do is pull the cluster and clean those ground contacts



This is what was happening after installing that new sending unit for the gauge. with the old sending unit at least the needle would rise with it.

 

I'd bet if the correct fans were installed the "problem" would go away. I find that upgrades often aint.

 

I understand now. You are bothered by the red warning light coming on when you turn on the AC and since replacing the gauge temp sensor you no longer have a mismatch between gauge and actual coolant temps as measured with an IR gun.

I think you have some sort of electrical or ground issue. I would start with a wiring diagram and trace the wires from the dash "warning light" and see if there is a short or something with the AC switch. Seems like a lot of work for something so minimal but I can tell by the condition of your car you are a perfectionist (in a good way) and want everything to be working as if the car were new.

With wiring/electrical issues its usually a case by case situation, so it will be spending some time examining wiring and wiring diagrams to chase down the potential cause. Again, grounds would be the first thing I would test and verify they are all connected, free of corrosion and grounding appropriately (with a multimeter).

Sending unit issue:
Im not sure your temp sending unit is to blame as it could be just a faulty ground which triggers the light to come on. Im not sure why you would revert to the old faulty temp sensor over the new one if the old one was clearly not measuring correct temps. I personally would run the new sensor, since it appears to be reading correctly. My guess is the warning light issue is likely not related to the sensor.

Im not sure what the comment above refers to but if you have changed out the fans, or have some funky workaround setup for fans, you should double check all the wiring. As you know fans are triggered by the AC and if any custom wiring was done, that could be a potential cause.

 

You also mention "you got your AC fixed" - how do you know this entire saga is not related to some issue with the AC wiring which was performed previously? Clearly the light is related to the AC, so thats where I would start looking, especially since you just had it worked on.

Keep the new temp sensor, start looking at the wiring, especially anything that was changed/modified during the AC service or fans. Your most recent photo, the temps look normal, and I must assume the gauge is reading correctly. There is a resistance test on Clarks you can use along with heated water to test the temp sensor resistances and whether they are in spec.

 

I spent more time with the schematic, and I have a couple of ideas for moving forward for the overheat light:

1) Ground the blue/green wire going to the sensor and make sure it lights up the overheat light. (It should.)
2) Take the sensor out and check the overheat switch by measuring resistance between the overheat terminal (small one, per Clark's) and the body. Set this up so you can drop it in a pot of boiling water. The terminal should remain open. If it closes, bad sensor. If it stays open, I suspect your car is overheating.
3) Your measurement of the temperature using an IR thermometer off the aluminum is likely low due to aluminum's low emissivity. If your IR gun allows you to adjust for emissivity, I would use the value of 0.20 based on what I found for oxidized aluminum around 200f at this link https://www.engineeringtoolbox.com/r...num-d_433.html. IR thermometers typically use a default value of 0.95.
4) If you really, really, really want to know the temperature of the coolant, I would buy a cheap meat thermometer off ebay for 7-8 bucks, get a short bolt that fits the bleeder hole, drill it for the thermometer, and epoxy the thermometer in. An aluminum or brass bolt will be easier to drill. You might test fit the bolt and mark which way is forward so you epoxy the thermometer in facing forward. Here's a thermometer in the US for 4 bucks https://www.ebay.com/itm/225673320116 If you are impatient, you can sacrifice the thermometer you have already.

I think at this point, I would really, really, really want to know the real temperature of the coolant.

 

Honestly, after testing the stock fans and directly comparing them to these spal fans, you might be dead on the money. The stock fans absolutely move more air It’s very apparent once testing them side by side. I threw both sets on the radiator I have sitting outside the car and there’s no question. I’m going to buy an anemometer to confirm. In the meantime before I replace the rad I’ll throw the stock fans back on and drive around with the AC on and see what happens…..

 

Thank you guys for the suggestions, after cleaning up the original sending unit and reinstalling it it is measuring much more accurately. I will continue cleaning grounds as suggested just ti make sure everything is in good working order.

The new sensor I installed was faulty. I did some testing in boiling water in an insulated container with the sending unit grounded. For whatever reason even at temps of 90+ C the new sending unit would only register 80. It would never display above 80c. After I realized this I decided to not spend any more time with it and I returned it.

I cleaned up all of the contacts and the probe surface of the higher quality VDO sending unit, and now the gauge is reading almost spot on. If I read 85C at the coolant neck, the needle sits just above the first line. I drove the car and let it sit outside so that the coolant neck read 95C. Key in the ignition and presto, gauge moved up to just below third white line. No warning light. Started the car and as the coolant flowed and began to cool down the engine the needle moved right back to above the first line.

I will still perform the suggested resistance reading tests and grounding of the light to make sure it’s operating correctly with numbers to back it up.

so far what I can report is that after my drive (AC off the whole time) temps settled at 79c on the lower rad hose, and about 89c on the outlet from the block.

No custom wiring besides adapters was used when installing the CSF fans. I will install the factory fans back in the car and try the AC and report back what happens.

Now as for the AC causing the overheating. I have not ruled this out as the shop that installed my compressor failed to install my under tray for the engine bay, failed to torque the mounting bolts for the compressor, and most egregiously failed to torque the belt tensioner causing the bolt to back out while driving risking my brand new compressor and crank pulley could have very well improperly filled the system both with oil or with refrigerant. After all that I wouldn’t put it beneath them. I’m still fighting to try and get my money back from that. Since that all happened I went and corrected all the work they failed to do ie properly installing everything and torquing everything to spec, but haven’t had a chance to have the pressures checked. From my research an improperly charged system can also cause a overheat with AC on condition. I have a lot of homework to do it seems like.

thank you all for taking time to help me out and for your replies it’s greatly appreciated.

 

Holy cow, I didn't realize all the things that you found wrong with the work. I'm glad you did, though, and addressed it.

Cheers